Satellite-acquired optical remote sensing imagery is extensively applied in time-critical applications like traffic surveillance and evaluation of natural disasters. However, clouds, as a common atmospheric phenomenon, frequently obscure observation. Current approaches aim to restore visibility in cloud-obscured regions, yet they typically fall short in the presence of dense cloud cover, which are exceedingly prevalent in remote sensing imagery. Alternative approaches rely on the satellite revisit cycle, frequently surpassing ten days, a duration impractical for genuine application scenarios due to target changes and bandwidth limitations. To address these issues, this paper proposes SCo-Cloud, a novel satellite constellation collaboration framework for cloud-aware onboard-computed imaging and transmission, which consists of Center-Sat and Edge-Sats. We propose onboard thin cloud removal and re-imaging region location models to locate the impact of clouds. We further design a novel multi-satellite scheduling strategy to eliminate clouds. The models above are integrated within the Center-Sat, with the nearby Edge-Sats collaborating in tandem to execute re-imaging assignments. Furthermore, to facilitate in-depth research, we have meticulously developed a cloud-covered target detection dataset. Comprehensive experiments have conclusively demonstrated that SCo-Cloud effectively surpasses the limitations inherent in current approaches, providing accurate and timely responses within the domain of Earth observation.

Yuanhong Huang, Huili Wang, Xueying Bai, Jinrui Wang, Jiajun Liu, Ziqin Wang, Wanchun Ni, Shangguang Wang, Tao Qi. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Jinrui Wang, Zhenfeng Gao, Wendan Wang, Huili Wang, Zichen Qin, Linjie Zhu, Hongke Fu, Shangguang Wang, Tao Qi. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Peiru Yang, Haoran Zheng, Tong Ju, Shiting Wang, Wanchun Ni, Jiajun Liu, Shangguang Wang, Yongfeng Huang, Tao Qi. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Cross-datacenter distributed model training is increasingly important, but WAN packet loss and long propagation delays make retransmission-based recovery expensive. Existing RDMA FEC schemes are application-unaware and uniformly protect all packets, wasting inter-datacenter bandwidth. In this paper, we present SMART, a semantic-aware loss recovery scheme for cross-datacenter model training. SMART distinguishes data-parallel (DP) gradients from pipeline-parallel (PP) activations using lightweight message-level tags; it selectively protects high-priority DP packets, gives earlier PP traffic stronger FEC, zero-fills only small profiled residual PP loss, and compensates tolerated low-priority DP losses. Simulation experiments with Qwen3-3B-derived workloads show that SMART reduces average and P99 flow completion times compared with retransmission-only recovery and uniform FEC, while preserving convergence close to the no-loss baseline.

Satellite telecommunication networks are playing an increasingly pivotal role in modern communication infrastructures, owing to their expansive coverage, high reliability, and growing capabilities in computing, storage, and bandwidth. In response to evolving market demands, mobile network operators are progressively integrating satellite systems with edge-cloud computing platforms to deliver advanced networking functionalities within a unified architecture. This integration places strong demands on the non-functional assessment (e.g., reliability, availability, and resource efficiency) of satellite-based edge nodes, introducing unprecedented challenges due to their unique operational constraints. In this paper, we propose a lightweight certification framework tailored for satellite computing systems, designed to assess and validate the non-functional posture of satellite edge networks. Our approach explicitly addresses the distinctive characteristics of satellite environments, including intermittent connectivity and constrained resource availability. We validate the proposed scheme through a realistic testbed implementation, modeling a 5G-enabled satellite edge node based on the Tiansuan satellite constellation, an experimental platform jointly developed by Beijing University of Posts and Telecommunications, Spacety, and Peking University.

To cope with the challenges of the rapid growth of Industrial Internet of Things devices, deploying cache servers (CS) at the edge of the network becomes a promising solution. The cache node (CN) can reduce request latency and the traffic load on backhaul link by pre-caching relevant data generated by sensors. However, the storage resources of CSs are limited, and select the appropriate data item to cache is a critical issue for improving efficiency. In addition, we introduce the Age of Information (AoI) to describe the freshness of cached data. However, increasing the freshness of data through frequent caching updates could result in increased energy consumption of sensor. In this paper, we study the caching problem of multiple CNs, and further model the problem as a multi-agent Markov Decision Process to minimize long-term request latency, AoI, and energy consumption. Furthermore, to solve this problem, we propose a variant of the Soft Actor-Critic-based multi-agent caching algorithm. Finally, experiments show that the proposed algorithm can converge to a better caching strategy and has better performance in multiple objectives than existing algorithms.

Deploying Convolutional Neural Network (CNN) models on ubiquitous Internet of Things (IoT) devices in a cloud-assisted manner to provide users with a variety of high-quality services has become mainstream. Most existing studies speed up model cloud training/on-device inference by reducing the number of convolution (Conv) parameters and floating-point operations (FLOPs). However, they usually employ two or more lightweight operations (e.g., depthwise Conv,$1\times 1$cheap Conv) to replace a Conv, which can still affect the model's speedup even with fewer parameters and FLOPs. To this end, we propose the Grouped NonLinear transformation generation method (GroupNL), leveraging data-agnostic, hyperparameters-fixed, and lightweight Nonlinear Transformation Functions (NLFs) to generate diversified feature maps on demand via grouping, thereby reducing resource consumption while improving the robustness of CNNs. First, in a GroupNL Conv layer, a small set of feature maps, i.e., seed feature maps, are generated based on the seed Conv operation. Then, we split seed feature maps into several groups, each with a set of different NLFs, to generate the required number of diversified feature maps with tensor manipulation operators and nonlinear processing in a lightweight manner without additional Conv operations. We further introduce a sparse GroupNL Conv to speed up by reasonably designing the seed Conv groups between the number of input channels and seed feature maps. Experiments conducted on benchmarks and on-device resource measurements demonstrate that the GroupNL Conv is an impressive alternative to Conv layers in baseline models. Specifically, on Icons-50 dataset, the accuracy of GroupNL-ResNet-18 is 2.86% higher than ResNet-18; on ImageNet-C dataset, the accuracy of GroupNL-EfficientNet-ES achieves about 1.1% higher than EfficientNet-ES. In addition, we verified the efficiency of GroupNL-based models in terms of cloud training and on-device inference.

Federated Learning (FL) facilitates privacy-preserving fine-tuning of Large Language Models (LLMs) for mobile applications, termed FedLLM. A vital challenge of FedLLM is the tension between LLM complexity and resource constraint of mobile devices. In response to this challenge, we first introduceFwdLLM(our conference version), an innovative FL framework designed to enhance the FedLLM efficiency. The key idea ofFwdLLMis to employ backpropagation (BP)-free training methods, requiring devices only to execute memory-efficient “perturbed inference”. Enabled by mobile NPU acceleration and an expanded array of participating devices,FwdLLMdelivers substantially better wall-clock efficiency than BP-based FedLLM. However,FwdLLMbased on vanilla BP-free optimization theoretically requires more optimization steps to converge. In this work, we further enhanceFwdLLMtoFwdLLM+, which incorporates advanced zeroth-order optimization techniques and low-rank perturbation decomposition to reduce convergence steps. Finally, we conduct extensive experiments on 4 models (ranging from 110 M to 7B) and 8 more datasets, demonstrating thatFwdLLM+achieves up to 151× faster training, a 93$\%$memory reduction compared to vanilla BP-based FedLLM, and superior performance compared toFwdLLM, enabling efficient federated fine-tuning of billion-parameter LLMs on commodity mobile devices.

Satellite computing, as an emerging edge computing paradigm, extends computing and networking services into space. Due to the internal design constraints of low-Earth orbit (LEO) satellites and the challenges posed by the external environment, satellite computing faces inherent limitations, including severely constrained resources, non-rechargeable batteries, poor heat dissipation, and highly dynamic operating conditions, leading to unreliable and unsustainable quality of service. To address the above challenges and fully realize the potential of satellite computing, this paper investigates temperature- and energy-aware dynamic task scheduling and computing resource allocation, aiming to optimize service latency, reduce onboard energy consumption, and enhance operational profit. Solving this problem requires coordinating task scheduling and resource allocation, balancing communication and computation latency, and addressing the challenge of a vast search space. To solve the above challenges, we first formulate this problem as a repeated Stackelberg game by developing temperature and energy models. Through theoretical analysis, we show that this game leads to a convex optimization framework that exhibits exponential complexity. To accelerate the search for the Stackelberg equilibrium solution, we propose a dynamic task scheduling algorithm based on the interior point method, which reduces the computational complexity to polynomial order. Trace-driven simulations demonstrate that the proposed algorithm reduces task scheduling latency by 28.4% and improves utility by 13% on average.

AI is making mobile applications increasingly cooler, but also introduces serious privacy risks due to the extensive user data collection. Federated learning (FL), as a privacy-preserving machine learning paradigm, enables mobile devices to collaboratively learn a shared prediction model while keeping all training data on devices. However, a key obstacle towards practical cross-device FL training is the huge energy consumption, especially for lightweight mobile devices. Prior literature mostly optimizes the convergence speed and network communication cost. In this work, we first perform the experimental analysis of improving FL performance through low-precision training with energy-friendly Digital Signal Processor (DSP) on mobile devices. Then, we demonstrate that directly integrating the state-of-the-art INT8 (8-bit integer) training algorithm and classic FL protocols will significantly degrade the model accuracy. Finally, we propose a novel FL protocol, namelyQ-FedUpdate, incorporates two critical techniques: error-compensated aggregation and pipelined batch quantization. The former can ensure the tiny model updates be accumulated and take effects, and the latter can improve the DSP cache hit rate to reduce the context switching. Extensive experiments show that,Q-FedUpdatecan effectively reduce the on-device energy consumption by 21×, and accelerate the FL convergence by 6.1× with only 2% accuracy loss.

The rapidly growing edge computing platforms, coupled with the still imperfect edge infrastructure, present an excellent opportunity for the emergence of new edge hardware. However, it remains unclear whether alternative architectures built from energy-efficient mobile System-on-Chips (SoCs) can meet the stringent performance, cost, and energy demands of modern edge workloads. In this paper, we propose a new type of edge server composed of 60 Qualcomm Snapdragon 865 mobile SoCs in a 2U rack, referred to as SoC Cluster. We demonstrate its successful deployment on existing edge cloud platforms and its ability to natively serve mobile cloud gaming services. Despite the emergence of new hardware on edge platforms and its successful operation in serving mobile cloud gaming, our trace analysis revealed low hardware utilization and significant dynamic fluctuations in usage. To assess its broader applicability, we conducted the first measurement study of SoC Cluster to reveal its ability to run two popular and modern edge applications: deep learning inference and video transcoding. We developed a cross-platform benchmark suite to evaluate throughput, latency, power consumption, and application-specific metrics like video quality. We then directly compare SoC Cluster with a traditional edge server equipped with Intel CPUs and NVIDIA GPUs in terms of energy efficiency, space efficiency, and monetary cost. Results show that SoC Cluster exhibits up to 6.5? higher energy efficiency and 7.7? higher space efficiency. We also disclose its limitations in serving computation-intensive workloads such as large deep learning models. The outcomes provide insightful implications and offer practical direction for refining SoC Cluster toward broader deployment in edge scenarios.

Linux, developed over the past three decades, has become the cornerstone of modern computing, serving as the foundation for a wide array of applications ranging from large-scale supercomputers to resource-constrained embedded devices. However, this infrastructure has long been plagued by memory and concurrency bugs, primarily due to the permissive nature of memory operations in the C language. The Rust-for-Linux ( RFL ) project offers a promising solution to address these long-standing safety concerns by integrating Rust’s advanced features, such as static ownership and type-checking, into the kernel. This integration could potentially eliminate memory and concurrency bugs while maintaining performance. Although RFL has matured and been merged into the Linux mainline, it remains understudied, and its effectiveness in balancing safety and performance is yet to be fully assessed. In this study, we present the first empirical analysis of RFL to evaluate its current state and benefits, focusing on how Rust integrates with Linux and whether this fusion ensures driver safety without significant overhead. We analyze eight RFL drivers, examining hundreds of issues and pull requests, thousands of GitHub commits, mail exchanges from the Linux mailing list, and over 12,000 discussions on Zulip. Our findings indicate that while Rust mitigates kernel vulnerabilities, it does not completely eliminate them. If not properly managed, Rust’s safety features even introduce notable runtime overhead and increased development complexity.

Restoring degraded images has long been a significant focus in image preprocessing, aiming to recover high-quality images from their degraded counterparts. While numerous studies have addressed specific types of degradation, All-in-One methods have achieved a notable milestone by enabling a single model to handle various degradations without requiring prior knowledge of the degradation types. However, these approaches often fall short of delivering satisfactory performance in complex and multiple degradation scenarios. To address these challenges, this article proposes a Multi-Prompt Guided Diffusion Network for All-in-One image restoration, referred to as MPGNet. MPGNet consists of two key components: the Image Restoration Part (IRP) and the Prompt Generation Part (PGP). The IRP employs a diffusion model to perform the image restoration task, while the PGP generates tailored prompts to guide image restoration from degradation. By leveraging meticulously designed prompts, MPGNet empowers the diffusion network to adaptively handle complex and multiple degradation scenarios. To enhance adaptability to various degradation types, content specificity, and overall performance, we design three types of prompts: Style Prompt, Content Prompt, and Learnable Prompt. The first two prompts are carefully generated using the Image Encoder from the pre-trained vision-text model CLIP, with contrastive learning applied during respective training, while the Learnable Prompt is custom-designed separately. We also introduce a Prompt Encoder that integrates Style, Content, and Learnable Prompts into a unified representation, facilitating image restoration for IRP. The experimental results demonstrate that the proposed method significantly outperforms recent approaches in both single and multiple degradation scenarios.

Serverless computing is a popular cloud computing paradigm that enables developers to build applications at the function level, known as serverless applications. The Serverless Application Model (AWS SAM) is the most widely adopted configuration schema. However, misconfigurations pose a significant challenge due to the complexity of serverless configurations and the limitations of traditional data-driven techniques. Recent advancements in Large Language Models (LLMs), pre-trained on large-scale public data, offer promising potential for identifying and explaining misconfigurations. In this article, we present SlsDetector , the first framework that harnesses the capabilities of LLMs to perform static misconfiguration detection in serverless applications. SlsDetector utilizes effective prompt engineering with zero-shot prompting to identify configuration issues. It designs multi-dimensional constraints aligned with serverless configuration characteristics and leverages the Chain of Thought technique to enhance LLM inferences, alongside generating structured responses. We evaluate SlsDetector on a curated dataset of 110 configuration files, which includes correct configurations, real-world misconfigurations, and intentionally injected errors. Our results show that SlsDetector , based on ChatGPT-4o (one of the most representative LLMs), achieves a precision of 72.88%, recall of 88.18%, and F1-score of 79.75%, outperforming state-of-the-art data-driven methods by 53.82, 17.40, and 49.72 percentage points, respectively. We further investigate the generalization capability of SlsDetector across recent LLMs, including Llama 3.1 (405B) Instruct Turbo, Gemini 1.5 Pro, and DeepSeek V3, with consistently high effectiveness.

Using low-rank adaptation to fine-tune pre-trained neural network models has attracted widespread attention due to its advantages of low resource requirements, high precision, and no additional inference delay. However, most existing methods are designed for large language models based on the Transformer structure and lack adaptation to convolutional neural networks (CNN) widely used on Internet of Things (IoT) devices. In addition, IoT devices are usually deployed outdoors and collect large amounts of data that are affected by environmental conditions. Providing a highly robust CNN model is a prerequisite for providing high-quality services. To this end, this paper proposes a highly robust seed-filter adaptation method (SeFA) for pre-trained CNNs. SeFA introduces an adaptation branch with the same structure as the backbone network. In the adaptation branch, some filters are first designated seed filters and grouped. Then, additional filters are generated from the grouped seed filters and nonlinear transformation functions (NLFs) with different hyperparameters. The seed filters' parameters are updated during model training, and the NLF hyperparameters are randomly initialized and frozen. Both grouping seed filters and configuring NLFs with non-learnable hyperparameters can improve the model's robustness. This is because grouping seed filters can generate diverse filters on demand without increasing the model's complexity, and the NLFs' rules can regularize the model. It is worth mentioning that the number of fine-tuning parameters of the pre-trained model that can adapt to downstream tasks can be flexibly controlled by specifying the number of seed filters. The key idea of this paper is to propose SeFA with flexible, controllable, learnable parameters, high robustness, and adaptability to pretrained CNN models, thereby facilitating fine-tuning on resource-constrained IoT devices and providing highly robust visual services. Experimental results on the CIFAR-10, CIFAR-10-C, CIFAR-100, CIFAR-100-C, and Icons-50 datasets demonstrate that the proposed SeFA outperforms other state-of-the-art methods. Specifically, on the ResNet-152 model and the CIFAR-10-C dataset, the accuracy of our SeFA is about 7% higher than that of the full fine-tuning method. In addition, we verify the efficiency of SeFA-based models for fine-tuning and inference on the device.

Containerization offers great flexibility and agility for resource provisioning in edge clouds. However, this benefit is not freely available, as substantial network traffic incurred by container image pulling heavily burdens the back-haul networks. Our in-depth measurements on 516.3 GB images from Docker Hub reveal that many similar images share identical layers, with only 11.21% having no shared layers. Building on this, we investigate the resource provisioning problem by leveraging image similarity to avoid repeated layer transmissions, aiming to reduce network traffic and improve overall performance. We formulate resource provisioning as a mixed-integer non-linear programming problem, which is challenging due to the coupling of four issues and their conflicting effects on overall performance, including offloading decisions, container instance deployment, image pulling, and resource allocation. To tackle these complexities, we propose a novel Similarity-Aware Resource Provisioning approach, which decomposes the problem into independent sub-problems using counterfactual multi-agent deep reinforcement learning and then solves sub-problems individually with convex optimization and fractional programming techniques. We conduct extensive evaluations with images from Docker Hub. The results show that our approach enables up to 32.6% traffic reduction and 19.9% utility improvement, outperforming the state-of-the-art solutions.

The diverse array of services provided by satellite constellations now plays a crucial role in human life, economic development, and national security. Imagine a scenario where, regardless of a user's location on Earth, commands can be transmitted to a satellite constellation. The system intelligently selects the most suitable services across the satellites, dynamically combining them to ensure rapid and seamless global fulfillment of the user's requirements. Although ideal in concept, the utilization of services is constrained by the limitations of satellite performance. This issue becomes especially critical when multiple services are requested simultaneously within a satellite constellation. An inefficient service composition strategy can fail to ensure adequate quality of service for users while also increases system load and significantly shortens the satellites' lifespan. In this work, we consider a service composition problem under a constrained space scenario, and we introduce aCooperative-enhancedMulti-AgentRL-based model (named CMAR) to address this issue. Specifically, CMAR formalizes the service composition problem as a Multi-agent Markov Decision Process, where two agents are designed to manage the selection for satellites and services separately to shrink the searching space. Considering collaborative difficulties caused by such a decomposition strategy, where agents can only observe partial changes in the environment, a mask-guided QMIX network is further designed to strengthen their corporation for proper joint actions. Our proposed method has been evaluated on one real-world dataset released by Tiansuan Constellation and two simulated datasets, the empirical results show that our model can improve the performance significantly compared with both traditional and RL-based baselines. The code is available athttps://github.com/CMARModel/CMARModel.git.

Recently, on-device large models (e.g., 7B parameter LLMs and LVLMs) are increasingly deployed in real-world services such as mobile assistants, edge computing, and IoT systems, where low latency and resource efficiency are critical. Since large models trained on general-domain data are typically suboptimal for downstream services, In-Context Learning (ICL) technique is widely used to enhance their task-specific service performance without extra tuning. However, effective ICL methods generally necessitate a sufficient amount of supervised data to provide abundant service-related information, which usually leads to a long context (i.e., more texts or images input) for models to process. This long context issue can significantly degrade the real-time inference efficiency and also impact the information utilization effectiveness, particularly for on-device large models with limited long context modeling capacities. To tackle this challenge, we propose a Personalized Knowledge Refinement (PKR) framework to achieve efficient and effective ICL for on-device large models. Specifically, we first introduce a personalized knowledge extraction module, which analyzes the inference behaviors of the target model on small-scale supervised data and then convert these behavioral patterns into personalized service-specific knowledge as the instruction context. Furthermore, we propose an adaptive knowledge filtration mechanism to model the informativeness of the extracted knowledge and eliminate the redundant ones, further improving the knowledge encoding efficiency per unit of context length. Experiments based on 14 benchmarks spanning both textual and visual task services, and 4 large models, demonstrated that PKR consistently improves task accuracy while reducing context length and inference latency, making it a practical solution for real-world on-device services. Codes are released athttps://github.com/wanghl21/PKR.

Cloud-assisted resource-constrained and widely used outdoor IoT devices deploying highly robust convolutional neural network (CNN) models to provide high-quality visual services have attracted widespread attention from industry and academia. Most existing methods suffer from two limitations: (i) large amount of parameter transmission, and (ii) low robustness in handling data affected by the environment. To this end, this paper proposes a transmission-friendly and robust CNN design method called NonLinear transformation generation Filter with NonLearnable hyperparameters, namely NL2Filter. In NL2Filter, some filters are first designated as seed filters, whose parameters are learnable, that is, updated as the model is trained. Other filters in this layer are generated based on the seed filters and nonlinear transformation function (NLF). The hyperparameters of the NLF in NL2Filter are randomly initialized and remain unchanged, so they can be saved and reproduced using random seed. After the cloud server trains the NL2Filter-CNN model, it only needs to send a small number of learnable parameters and random seed to reproduce the CNN model trained by the cloud server. Compared with sending the complete CNN model, sending only a small number of learnable parameters and random seed can significantly reduce the number of model parameters sent by the cloud server. On the other hand, NL2Filter groups seed filters and then uses NLFs with different hyperparameters to generate diverse filters on demand for each group, thereby improving the model's ability to capture feature diversity without increasing the complexity of the model, thereby improving the model's robustness in processing data affected by the environment. Experimental results on CIFAR-10, CIFAR-10-C, and Icons-50 datasets demonstrate that the proposed NL2Filter outperforms other state-of-the-art methods. Specifically, using the ResNet-101 architecture, on the CIFAR-10-C, NL2Filter's accuracy is about 3.9% higher than that of MonoCNN; On the Icons-50, NL2Filter's accuracy is about 2.7% higher than that of MonoCNN and about 2.3% higher than that of the standard ResNet-101.

Trajectory similarity is a cornerstone of trajectory data management and analysis. Traditional similarity functions often suffer from high computational complexity and a reliance on specific distance metrics, prompting a shift towards deep representation learning in Euclidean space. However, existing Euclidean-based trajectory embeddings often face challenges due to the triangle inequality constraints that do not universally hold for trajectory data. To address this issue, this paper introduces a novel approach by incorporating non-Euclidean geometry, specifically hyperbolic space, into trajectory representation learning. We present the first-ever integration of hyperbolic space to resolve the inherent limitations of the triangle inequality in Euclidean embeddings. In particular, we achieve it by designing a Lorentz distance measure, which is proven to overcome triangle inequality constraints. Additionally, we design a model-agnostic framework LH-plugin to seamlessly integrate hyperbolic embeddings into existing representation learning pipelines. This includes a novel projection method optimized with the Cosh function to prevent the diminishment of distances, supported by a theoretical foundation. Furthermore, we propose a dynamic fusion distance that intelligently adapts to variations in triangle inequality constraints across different trajectory pairs, blending Lorentzian and Euclidean distances for more robust similarity calculations. Comprehensive experimental evaluations demonstrate that our approach effectively enhances the accuracy of trajectory similarity measures in state-of-the-art models across multiple real-world datasets. The LH-plugin not only addresses the triangle inequality issues but also significantly refines the precision of trajectory similarity computations, marking a substantial advancement in the field of trajectory representation learning.

Trajectory similarity measure is a fundamental component in trajectory databases, supporting many down-stream trajectory tasks. Existing similarity functions often exhibit unacceptable time complexities, hampering their efficiency for real-world scenarios. To address this limitation, learning-based approximation techniques utilizing trajectory embeddings have been proposed. However, creating a robust embedding model presents challenges, including the lack of direct involvement in the computational similarity process, adherence to non-metric similarity spaces, and the integration of precise similarity computation alignments. To address these challenges, we introduce DTisT, a novel embedding framework that enhances trajectory embeddings by pairwise learning from dual-trajectory input models. DTisT not only captures the dynamics of trajectory similarity computation through a dual-trajectory learning model but also integrates a learnable virtual trajectory to align the embedding space with non-metric similarity spaces effectively. Additionally, we incorporate aligned information from actual similarity computations into our embedding process using an attention mask mechanism. To ensure effective learning, we adopt a pre-train and fine-tune strategy, utilizing contrastive learning during the pre-training stage. Extensive experiments conducted on two real datasets demonstrate that DTisT surpasses state-of-the-art methods, showcasing its effectiveness in trajectory similarity embedding.

Data imputation (DI) is a common means of enhancing data quality. To adapt to the flourishing field of machine learning (ML), an innovative class of imputation methods that consider downstream models in the imputation process has been proposed, denoted as DI for ML. A critical challenge within this context is establishing the optimal order for imputing a set of incomplete samples. To address this, we propose an iterative approach that strategically determines the imputation order based on the potential impact on model performance. At first, we design the impact score in a what-if manner to evaluate the significance of each incomplete data point for downstream ML models. In addition, to tackle the challenge of insufficient reliable complete data in real-world scenarios, we ingeniously leverage meta-learning mechanisms to enhance the robustness of the impact score computation. Finally, to avoid the risk of converging to local optima and non-diverse data selection during iterative imputation, we introduce a real-time feedback strategy using the Multi-Armed Bandit mechanism. By balancing immediate rewards with long-term strategic gains, our approach effectively navigates the complex optimization landscape, leading to globally optimal imputation orders. We experimentally validated our method on eight real-world datasets and five types of ML models, with the results indicating that the imputation order optimized by our method outperforms the current state-of-the-art methods.

Mobile edge computing is envisioned as a promising paradigm to overcome the computational limitations of vehicles. However, current resource allocation optimizations for vehicular services in mobile edge computing are typically performed under quasi-static scenarios. Due to the time-varying wireless communication environments between vehicles and the base station, the communication uncertainty becomes a key factor to hinder the service reliability. To tackle this problem, the reliability-aware mobile edge resource allocation for vehicular services is explored in this paper. Firstly, the model of computation, communication, and reliability is reviewed integratively, and the problem is then formulated as a fractional programming model. Secondly, conditional value-at-risk (CVaR) is adopted to transform the original problem into a solvable semidefinite programming model. Finally, we propose an efficient two-layer reliability-aware resource allocation approach that decomposes the original problem into two sub-problems with lower complexity. Lagrange multiplier and Dinkelbach strategy are then applied for these sub-problems, which minimize average energy consumption while ensuring reliability for vehicles. Simulation results examine the performance of our proposed approach and demonstrate its superiority over other approaches.

Using pruning techniques to prune redundant weights in large language models (LLMs) for model size reduction, enabling deployment on devices to deliver high-quality services, has garnered significant attention from industry and academia. However, most existing pruning methods suffer from two major problems: (1) they rely on operations with high computational complexity, such as Hessian matrix calculation, and (2) high pruning rate leads to a significant decrease in model accuracy. To this end, this paper proposes a low resource requirement and low accuracy loss post-training pruning approach, namely the compensated post-training pruning method (CPP) for on-device LLM services. First, CPP employs singular value decomposition on weight matrices, sorting the decomposed singular values in descending order. Based on the pruning ratio, it retains the principal eigenvector corresponding to the larger singular values. To maintain consistent output feature distributions, CPP applies orthogonal transformations to input data and weight matrices, leveraging the principle of matrix orthogonality invariance. Compared with pruning approaches based on the Hessian matrix, CPP does not need to iteratively calculate the second derivative, thereby avoiding a high computational overhead. Second, the CPP incorporates bias compensation to use valuable information in pruned weights to improve the accuracy of the model. It constructs mapping relationships between input features and pruned weights through tensor decomposition techniques to generate bias compensation terms. These terms fine-tune the output of each layer, reducing pruning-induced errors from ratio-based pruning, and consequently improving model accuracy. Finally, CPP is validated through comprehensive experiments on nine benchmark datasets (e.g., WikiText-2 and PIQA) using large language models (i.e., LLaMA and OPT). Specifically, compared with the FLAP, CPP demonstrates 6.68% reduction in average pruning time. At pruning ratios of 10% and 20%, CPP achieves 2. 14% and 1.45% reductions in perplexity, respectively, along with 1.22% to 3.43% improvement in zeroshot inference accuracy.

Multi-modal relation extraction (MMRE) is a challenging task that seeks to identify relationships between entities with textual and visual attributes. However, existing methods struggle to handle the complexities posed by multiple entity pairs within a single sentence that share similar contextual information (e.g., identical text and image content). These scenarios amplify the difficulty of distinguishing relationships and hinder accurate extraction. To address these limitations, we propose the variational multi-modal hypergraph attention network (VM-HAN), a novel and robust framework for MMRE. Unlike previous approaches, VM-HAN constructs a multi-modal hypergraph for each sentence-image pair, explicitly modeling high-order intra-/inter-modal correlations among different entity pairs in the same context. This design enables a more detailed and nuanced understanding of entity relationships by capturing intricate cross-modal interactions that are often overlooked. Additionally, we introduce the variational hypergraph attention network (V-HAN). This variational attention mechanism dynamically refines the hypergraph structure, enabling the model to effectively handle the inherent ambiguity and complexity of multi-modal data. Comprehensive experiments on benchmark MMRE datasets demonstrate that VM-HAN achieves state-of-the-art performance, significantly surpassing existing methods in both accuracy and efficiency.

Satellite technology has significantly influenced our daily lives, manifested in applications such as navigation and communication. With its development, a vast amount of multimodal satellite commonsense data has been generated, thus leading to an urgent demand for conversation about satellite data. However, existing large language models suffer from prevalent hallucinations and poor comprehensibility on multimodal satellite data due to their high professional content threshold and partial information opacity. To address these issues, we propose a multimodal satellite knowledge retrieval-augmented iterative alignment framework (Sat-RIA) for satellite commonsense conversation. We first construct multi-view retrieval expert knowledge to reduce hallucinations and enhance the interpretability of responses, which incorporates the satellite expert database, satellite rule, satellite image database, and a satellite knowledge graph. We next design commonsense conversation instructions to make the answers more legible and understandable. Furthermore, the retrieval-augmented iterative alignment module refines response precision by aligning outputs with task-specific standards through multi-stage evaluations. Finally, we construct satellite multi-turn dialogue and visual question-answer datasets for a more comprehensive evaluation of satellite commonsense conversation. Experimental results demonstrate that Sat-RIA outperforms existing large language models and provides more comprehensible answers with fewer hallucinations.

Satellite computing is an emerging paradigm that empowers satellites to perform onboard processing tasks (i.e., satellite applications), thereby reducing reliance on ground-based systems and improving responsiveness. However, enabling application software updates in this context remains a fundamental challenge due to application heterogeneity, limited ground-to-satellite bandwidth, and harsh space conditions. Existing software update approaches, designed primarily for terrestrial systems, fail to address these constraints, as they assume abundant computational capacity and stable connectivity.To address this gap, we propose SateLight, a practical and effective satellite application update framework tailored for satellite computing. SateLight leverages containerization to encapsulate heterogeneous applications, enabling efficient deployment and maintenance. SateLight further integrates three capabilities: (1) a content-aware differential strategy that minimizes communication data volume, (2) a fine-grained onboard update design that reconstructs target applications, and (3) a layer-based fault-tolerant recovery mechanism to ensure reliability under failure-prone space conditions. Experimental results on a satellite simulation environment with 10 representative satellite applications demonstrate that SateLight reduces transmission latency by up to 91.18% (average 56.54%) compared to the best currently available baseline. It also consistently ensures 100% update correctness across all evaluated applications. Furthermore, a case study on a real-world in-orbit satellite demonstrates the practicality of our approach.

Network load balancing is important in datacenter network for LLM training with low entropy traffic. Through in-depth experimental investigations, we find that traditional flowlet-based switch rerouting schemes face substantial challenges in achieving an optimal balance between flowlet timeout settings and packet reordering issues under LLM traffic. Thus, we present CoLB, an application layer and network layer codesign load balancing scheme. CoLB utilizes size-based flow block instead of a timeout value-based flowlet as the rerouting unit for LLM traffic, where the size of the flow block is adaptive to the collective communication patterns generated by the application workload layer. Besides, CoLB is aware of the link utilization affected by PFC mechanism in the underlying RDMA network and guarantees a more accurate perception of link congestion conditions. We evaluate CoLB with the NS-3 simulator under the collective communication traffic and realistic workloads of LLMs. The simulation results show that CoLB is superior to previous switch-based load balancing schemes. For example, CoLB achieves up to 79.88% lower 99th Flow Completion Time (FCT) slowdown than LetFLow and up to 38.83% than Conweave.

The proliferation of satellite-related commonsense data on the internet. Traditional analytical methods are challenging to integrate and effectively uncover implicit knowledge within them. However, current deep learning and LLM-based approaches often struggle with errors and hallucinations when performing multi-hop reasoning in domain-specific contexts. To address these limitations, we propose a novel multi-hop reasoning framework for implicit commonsense mining. This framework aims to uncover the underlying meta-knowledge behind reasoning problems, thereby providing enhanced interpretability of the reasoning process. Specifically, we design an extraction-retrieval-principle multi-step reasoning method that generates different levels of meta-knowledge in stages to support the reasoning process effectively. We further design the mixture of expert knowledge graph construction to construct a satellite knowledge graph that supports multi-hop reasoning. Experimental results demonstrate that our approach outperforms baselines on satellite knowledge graph reasoning.

Infrared and visible image fusion (IVIF) aims to extract fine details from visible images and complementary information from infrared images. Most existing methods directly extract relevant and complementary features from each modality using neural networks, often overlooking the guidance process and the distinct frequency-domain characteristics of these features. To address this, we propose HRFusion-a novel frequency-domain framework that extracts complementary features from hybrid features using prior-constrained relevant features, effectively enhancing complementary information and reducing redundancy. In HRFusion, hybrid and relevant features are robustly extracted to guide the subsequent fusion stage. By leveraging frequency differences between complementary and relevant features, we introduce the Enhanced Complementary Frequency Network (ECFNet), which uses optimized Variational Mode Decomposition (VMD) to effectively separate and process these signals for fusion. The overall architecture is built with the proposed DTBlock, which captures both global and local features. Extensive experiments show that our method achieves state-of-the-art performance on the TNO, MSRS, M3FD, and Harvard Brain datasets, outperforming recent approaches. Code is available at https://github.com/liuuuuu777/HRFusion.

Monocular depth estimation stands as a fundamental pursuit in computer vision. Recently, some methods have attempted to introduce the text-to-image diffusion model into the domain of monocular depth estimation and achieved impressive results. However, these methods typically employ pre-defined templates as text prompts to guide the learning of denoising networks, resulting in limited flexibility and scalability. In this paper, we propose OGDepth, a diffusion-based monocular depth estimation network with object prompts generated by taking advantage of the object detection information from the scene. Specifically, we design an Object Prompt Module (OPM) to encode the object detection information into prompts that are more closely aligned with the image content, offering richer contextual information while circumventing the monotony and redundancy inherent in template-generated prompts. Moreover, we employ bounding box information for each object to filter and localize objects, enabling the model to grasp relative positional information within the scene. This facilitates the creation of a more precise depth map. Additionally, we design a Global-Local Interaction Decoder (GLID) to facilitate the mutual exchange of features at different scales, enabling efficient feature fusion. Our approach underwent rigorous experiments across multiple datasets, with results showcasing its state-of-the-art performance. Notably, on the KITTI dataset, our model achieves an RMSE of 1.967 and a REL of 0.047, and both metrics are the best among all compared methods. On the NYU Depth V2 dataset, our method achieves an RMSE score of 0.221, representing a notable 12.9% enhancement compared to the baseline method (VPD).

The rapid advancement in satellite technology with the adoption of commercial off-the-shelf (COTS) devices and satellite constellation networking has given rise to Space Computing Networks (SCNs). While SCN research is typically conducted on experimental platforms due to high operational costs, the unique challenges of SCNs require special consideration. In this demo, we introduce Rhone, an emulator that bridges these gaps by achieving both satellite- and constellation-level fidelity (the accurate replication of satellite and constellation states, including power, thermal, and network conditions, as well as application performance) while ensuring usability. Rhone adopts a two-phase approach: i) an offline phase builds power, thermal, orbit, network, and computation models using real satellite telemetries and hardware-in-the-loop chip mirroring, and ii) an online phase executes container-based emulation integrated with these models. Evaluation shows Rhone's power and computation model errors under 5% and thermal model errors within 1.3 – 2.5°C.

As Large Language Models (LLMs) are frequently updated, LoRA weights trained on earlier versions quickly become obsolete. The conventional practice of retraining LoRA weights from scratch on the latest model is costly, time-consuming, and environmentally detrimental, particularly as the diversity of LLMs and downstream tasks expands. This motivates a critical question: "How can we efficiently leverage existing LoRA weights to adapt to newer model versions?" To address this, we propose LoRASuite, a modular approach tailored specifically to various types of LLM updates. First, we compute a transfer matrix utilizing known parameters from both old and new LLMs. Next, we allocate corresponding layers and attention heads based on centered kernel alignment and cosine similarity metrics, respectively. A subsequent small-scale, skillful fine-tuning step ensures numerical stability. Experimental evaluations demonstrate that LoRASuite consistently surpasses small-scale vanilla LoRA methods. Notably, on backbone LLMs such as MiniCPM and Qwen, LoRASuite even exceeds the performance of full-scale LoRA retraining, with average improvements of +1.4 and +6.6 points on math tasks, respectively. Additionally, LoRASuite significantly reduces memory consumption by 5.5 GB and computational time by 78.23%.

Large vision-language models (LVLMs) derive their capabilities from extensive training on vast corpora of visual and textual data. Empowered by large-scale parameters, these models often exhibit strong memorization of their training data, rendering them susceptible to membership inference attacks (MIAs). Existing MIA methods for LVLMs typically operate under white- or gray-box assumptions, by extracting likelihood-based features for the suspected data samples based on the target LVLMs. However, mainstream LVLMs generally only expose generated outputs while concealing internal computational features during inference, limiting the applicability of these methods. In this work, we propose the first black-box MIA framework for LVLMs, based on a prior knowledge-calibrated memory probing mechanism. The core idea is to assess the model memorization of the private semantic information embedded within the suspected image data, which is unlikely to be inferred from general world knowledge alone. We conducted extensive experiments across four LVLMs and three datasets. Empirical results demonstrate that our method effectively identifies training data of LVLMs in a purely black-box setting and even achieves performance comparable to gray-box and white-box methods. Further analysis reveals the robustness of our method against potential adversarial manipulations, and the effectiveness of the methodology designs. Our code and data are available at \url{https://github.com/spmede/KCMP}.

With the rapid expansion of Internet of Things (IoT) devices and the growing demand for real-time data processing, Remote Direct Memory Access (RDMA) has become increasingly vital in edge datacenters due to its high performance and low CPU utilization. To fully harness RDMA’s potential, a lossless underlying network is essential, typically achieved through hop-by-hop Priority Flow Control (PFC). However, existing RDMA congestion control mechanisms in edge datacenters struggle to achieve rapid rate convergence and may exacerbate PFC side effects such as head-of-line blocking, unfairness, and even deadlocks. In this paper, we propose a reinforcement learning-based RDMA congestion control enhancement framework called RDI. By leveraging receiver-side information and network congestion levels, RDI provides precise rate guidance for congested flows, alleviating congestion and mitigating PFC issues. Furthermore, RDI combines both online and offline learning to assist receiver-side information, achieving finer rate adjustments to dynamically adapt to changing network workloads and optimize congestion control performance. RDI is transparent and compatible with existing RDMA network architectures, requiring no modifications to network devices. Extensive simulations under realistic traffic patterns show that the congestion control scheme enhanced by RDI significantly outperforms the original mechanisms in terms of throughput and flow completion time (FCT), while also reducing PFC side effects. RDI-enhanced congestion control reduces the 99th-percentile tail FCT by up to 92% and the average FCT by up to 60%.

Online map matching is a fundamental problem in location-based services, aiming to incrementally match trajectory data step-by-step onto a road network. However, existing methods fail to meet the needs for efficiency, robustness, and accuracy required by large-scale online applications, making this task still challenging. This paper introduces a novel framework that achieves high accuracy and efficient matching while ensuring robustness in handling diverse scenarios. To improve efficiency, we begin by modeling the online map matching problem as an Online Markov Decision Process (OMDP) based on its inherent characteristics. This approach helps efficiently merge historical and real-time data, reducing unnecessary calculations. Next, to enhance robustness, we design a reinforcement learning method, enabling robust handling of real-time data from dynamically changing environments. In particular, we propose a novel model learning process and a comprehensive reward function, allowing the model to make reasonable current matches from a future-oriented perspective, and to continuously update and optimize during the decision-making process based on feedback. Lastly, to address the heterogeneity between trajectories and roads, we design distinct graph structures, facilitating efficient representation learning through graph and recurrent neural networks. To further align trajectory and road data, we introduce contrastive learning to decrease their distance in the latent space, thereby promoting effective integration of the two. Extensive evaluations on three real-world datasets confirm that our method significantly outperforms existing state-of-the-art solutions in terms of accuracy, efficiency and robustness.

Geospatial Knowledge Graphs (KGs) are widely used data structures that integrate rich knowledge from multi-source databases and play a crucial role in applications such as data retrieval and urban management. However, existing methods for Geospatial Knowledge Graph Completion (KGC) heavily rely on extensive labeled data and lack the ability to direct inference on new unlabeled geospatial databases and thus limiting their practical deployment. To address this limitation, this paper first formalizes a novel zero-shot transfer scenario and then proposes an innovative geospatial multimodal large language model framework capable of efficient Geospatial KGC with robust zero-shot generalization capabilities. Specifically, to enable effective direct inference under significant data discrepancies inherent in zero-shot scenarios, we introduce large language models (LLMs) into the geospatial KGC problem for the first time and redefine the multimodal data processing paradigm for geospatial LLMs. Next, to overcome the challenge that LLMs cannot directly handle geospatial data, we innovatively propose a Pretrain Geospatial Encoder that performs self-supervised pretraining exclusively on spatial data. Additionally, to integrate geospatial and textual modalities, we design an adaptation component that injects geospatial features into the LLMs and introduce a multi-task fine-tuning procedure. Lastly, to ensure robustness across multi-target domain scenarios, we present an implicit data alignment strategy based on adversarial learning. Extensive evaluations conducted on four real-world datasets demonstrate that our method significantly outperforms state-of-the-art approaches in terms of accuracy and robustness.

En Route Travel Time Estimation (ER-TTE) aims to learn driving patterns from traveled routes to achieve rapid and accurate real-time predictions. However, existing methods ignore the complexity and dynamism of real-world traffic systems, resulting in significant gaps in efficiency and accuracy in real-time scenarios. Addressing this issue is a critical yet challenging task. This paper proposes a novel framework that redefines the implementation path of ER-TTE to achieve highly efficient and effective predictions. Firstly, we introduce a novel pipeline consisting of a Decision Maker and a Predictor to rectify the inefficient prediction strategies of current methods. The Decision Maker performs efficient real-time decisions to determine whether the high-complexity prediction model in the Predictor needs to be invoked, and the Predictor recalculates the travel time or infers from historical prediction results based on these decisions. Next, to tackle the dynamic and uncertain real-time scenarios, we model the online decision-making problem as a Markov decision process and design an intelligent agent based on reinforcement learning for autonomous decision-making. Moreover, to fully exploit the spatio-temporal correlation between online data and offline data, we meticulously design feature representation and encoding techniques based on the attention mechanism. Finally, to improve the flawed training and evaluation strategies of existing methods, we propose an end-to-end training and evaluation approach, incorporating curriculum learning strategies to manage spatio-temporal data for more advanced training algorithms. Extensive evaluations on three real-world datasets confirm that our method significantly outperforms state-of-the-art solutions in both accuracy and efficiency.

Traffic accidents pose a significant risk to human health and property safety. To address this issue, predicting their risks has garnered growing interest. We argue that a desired prediction solution should demonstrate resilience to the complexity of traffic accidents. In particular, it should adequately consider the streaming nature of data and key related aspects, such as regional background, accurately capture both proximity and similarity while bridging the disparities, and effectively address the sparsity. However, these factors are often overlooked or difficult to incorporate. In this paper, we propose a novel streaming multi-granularity hierarchical spatio-temporal network. Initially, we innovate by incorporating remote sensing data, facilitating the creation of hierarchical multi-granularity structure and the comprehension of regional background. We construct multiple high-level risk prediction tasks to enhance model's ability to cope with sparsity. Subsequently, to capture and bridge spatial proximity and semantic similarity, region features and multi-view graph undergo encoding processes to distill effective representations, followed by a graph-enhanced representation alignment module that reconciles their disparities. At last, an alternating experience replay with a dual-memory buffer is employed to accommodate streaming data scenarios. Extensive experiments on two real datasets verify the superiority of our model against the state-of-the-art methods.

Nanosatellite constellations equipped with sensors capturing large geographic regions provide unprecedented opportunities for Earth observation. As constellation sizes increase, network contention poses a downlink bottleneck. Orbital Edge Computing (OEC) leverages limited onboard compute resources to reduce transfer costs by processing the raw captures at the source. However, current solutions have limited practicability due to reliance on crude filtering methods or over-prioritizing particular downstream tasks. This work presents an OEC-native and task-agnostic feature compression method that preserves prediction performance and partitions high-resolution satellite imagery to maximize throughput. Further, it embeds context and leverages inter-tile dependencies to lower transfer costs with negligible overhead. While the encoding prioritizes features for downstream tasks, we can reliably recover images with competitive scores on quality measures at lower bitrates. We extensively evaluate transfer cost reduction by including the peculiarity of intermittently available network connections in low earth orbit. Finally, we test the feasibility of our system for standardized nanosatellite form factors. We demonstrate that the proposed approach permits downlinking over 100× the data volume without relying on prior information on the downstream tasks.

The recent emergence of Low Earth Orbit (LEO) mega-constellations, designed for high-speed broadband connections with low latency, has introduced new deployment challenges. Efficient launch sequence planning is crucial for rapid service rollout, performance enhancement, and service promotion. However, existing research predominantly focuses on the design and performance analysis of fully-deployed constellations and overlooks the evolving process from a partially-deployed constellation to a fully-deployed one. This paper explores the launch sequence optimization problem for mega-constellations, tailored to expedite service delivery and adapt to changing performance demands. To this end, (1) we identify critical network performance metrics for the constellation evolving process and construct a simulation toolchain capable of simulating and evaluating these metrics for any potential partially-deployed constellation. (2) Drawing upon three key observations on network availability, the number of visible satellites, and latency, we propose an algorithm that can construct a launch sequence for an arbitrary mega-constellation topology. Evaluation results show that this algorithm enables the early provision of services and maximizes network performance gains at each launch batch while catering to different user demands. For instance, our algorithm can achieve network performance nearly equivalent to that of Starlink when it initiated its service, without losing redundancy, while using 55% fewer satellites.

Prior studies on traditional centralized clouds independently optimizing static resource utilization and dynamic bandwidth cost are not applicable to edge scenarios, where edge sites are interconnected by wide area networks (WAN) rather than local area networks (LAN) as within clouds. Due to the lack of knowledge about the actual status of public edge platforms and real-world edge datasets, existing influential literature on edge scenarios demonstrates significant disparities in optimization objectives and perspectives. To bridge this gap, we collaborate with a public edge platform and perform a comprehensive measurement, which reveals limitations of the status quo VM scheduling schemes and potential opportunities for improvement. However, resolving VM scheduling considering static resource utilization, dynamic bandwidth cost, and end users’ QoE in a cost-efficient manner faces several challenges, including coupled objectives, exponentially increased complexity, and spatiotemporal dynamics. To address the above challenges, in this work, we propose a holistic online framework that integrates combinatorial bandit-based VM migration and seasonality-aware VM request allocation at two distinct time granularities. Large-scale experiments based on a real-world dataset confirm that our online framework achieves near-offline bandwidth cost and resource utilization while significantly lowering time consumption.

With the rapid development of edge computing in the post-COVID19 pandemic period, precise workload forecasting is considered the basis for making full use of the edge-limited resources, and both edge service providers (ESPs) and edge service consumers (ESCs) can benefit significantly from it. Existing paradigms of workload forecasting (i.e., edge-only or cloud-only) are improper, due to failing to consider the inter-site correlations and might suffer from significant data transmission delays. With the increasing adoption of edge platforms by web services, it is critical to balance both accuracy and efficiency in workload forecasting. In this paper, we propose XELASTIC, which offers three key improvements over the conference version. First, we redesigned the aggregation and disaggregation layers using GCNs to capture more complex relationships among workload series. Second, we introduced a supervised contrastive loss to enhance robustness against outliers, particularly for handling missing or abnormal data in real-world scenarios. Finally, we expanded the evaluation with additional baselines and larger datasets. Extensive experiments on realistic edge workload datasets collected from China’s largest edge service provider (Alibaba ENS) show that XELASTIC outperforms state-of-the-art methods, decreases time consumption, and reduces communication costs.

With the explosive growth of information and the continuous expansion of applications, Industrial Internet of Things (IIoT) is facing huge data processing and storage pressure. With mobile edge computing (MEC) technology, the computing power network connects the geographically distributed computing nodes and then coordinates the allocation and scheduling of resources, transmits data, and eventually relieves the pressure of the industrial site. However, the rigorous demands of IIoT for real-time and stability pose some daunting challenges. To this end, we propose an online computation offloading approach with dual stability guarantee, named OCODSG. Specifically, the Lyapunov function is used to optimize the stability of the virtual queue, and the system stability is optimized based on the network jitter measurement. Moreover, the Dueling Double Deep Q Network (D3QN) algorithm based on deep reinforcement learning (DRL) is used for model autonomous training, while Gaussian noise is added to the network parameter space to encourage exploration and enhance algorithm robustness. Finally, experimental results on both simulated and real datasets demonstrate that OCODSG improves service efficiency and system stability.

Executing 5G core network functions on satellites has become crucial to enhance satellite network management and service capabilities. The User Plane Function (UPF) is responsible for efficient data traffic forwarding and is envisioned as a key and pioneering core network function that will be deployed on satellites. However, managing and providing services with satellite UPFs face dual challenges. Limited satellite resources constrain the user scale that a satellite UPF can service, resulting in an unguaranteed service delay. Moreover, the extremely rapid mobility of satellites renders it difficult for satellite UPFs to provide seamless services. To address the above challenges, this paper presents the first-of-its-kind service optimization scheme for satellite UPFs in terms of switch control, state migration, and traffic routing. To provide guaranteed service delay, we provide a theoretical analysis based on the M/G/1 queue model, demonstrating the service delay-resource consumption trade-off. A satellite UPF switch control scheme is integrated into the service optimization process, which can decrease satellite UPF service delay while saving satellite resources by adjusting the switch control parameters. To provide seamless services, we propose a satellite UPF-oriented state-aware service migration and traffic routing (UPF service optimization) algorithm. A policy network-based reinforcement learning approach is employed to dynamically perceive the satellite network’s state as well as the satellite UPF switch state. Building upon the optimization of service delay through satellite UPF switch control, the processes of state-aware state migration and traffic routing are further employed to reduce delay, ensuring seamless service effectively. Experiments reveal that the proposed algorithm outperforms other benchmark algorithms under different metrics. The service delay is reduced by an average of 23.2% compared with other algorithms.

As a distributed machine learning method, federated learning (FL) can collaboratively train a global model with multiple devices without sharing the original data, thus protecting certain privacy. However, due to the strong heterogeneity of edge nodes (ENs) participating in FL, the quality of data uploaded to the parameter server (PS) varies significantly. Without an appropriate incentive mechanism, low-quality contributors may receive disproportionately high rewards, while high-quality contributors may lack sufficient motivation, leading to inefficient participation and suboptimal global model performance. Consequently, it is critical to develop an effective incentive mechanism to promote fairness for the FL process. To address the issues of existing FL incentive mechanisms lacking privacy protection performance analysis, we propose a fairness-aware incentive mechanism for multi-server FL in edge-enabled wireless differential privacy (DP) networks. Specifically, the wireless channel noise is used to provide DP protection for the local model gradients uploaded by ENs. Next, the interaction between the PSs and ENs is modeled as a Stackelberg game. Furthermore, we solve the Stackelberg game process using backward induction and theoretically propose optimal strategies for both the PSs and ENs. Finally, extensive numerical simulations using real datasets demonstrate the superior performance of our theoretical analysis of the proposed scheme.

Large language models (LLMs) such as GPTs and Mixtral-8x7B have revolutionized machine intelligence due to their exceptional abilities in generic ML tasks. Transiting LLMs from datacenters to edge devices brings benefits like better privacy and availability, but is challenged by their massive parameter size and thus unbearable runtime costs. To this end, we presentEdgeMoE, an on-device inference engine for mixture-of-expert (MoE) LLMs – a popular form of sparse LLM that scales its parameter size with almost constant computing complexity.EdgeMoEachieves both memory- and compute-efficiency by partitioning the model into the storage hierarchy: non-expert weights are held in device memory; while expert weights are held on external storage and fetched to memory only when activated. This design is motivated by a key observation that expert weights are bulky but infrequently used due to sparse activation. To further reduce the expert I/O swapping overhead,EdgeMoEincorporates two novel techniques: (1) expert-wise bitwidth adaptation that reduces the expert sizes with tolerable accuracy loss; (2) expert preloading that predicts the activated experts ahead of time and preloads it with the compute-I/O pipeline. On popular MoE LLMs and edge devices,EdgeMoEshowcase significant memory savings and speedup over competitive baselines.

As a key technology of intelligent satellite-enabled services in B5G or 6G networks, deploying Deep Neural Networks (DNN) models on satellites has been a notable trend, catering to the daily demand for extensive computing-intensive and latency-sensitive tasks. The computing resources are strategically deployed on satellites where sensor data is generated or collected, facilitating the fine-grained computational inference of DNN-based tasks. However, no prior study has comprehensively explored the crucial inference challenges – e.g., the trade-off between the number of tasks completed and accuracy and partitioning models in multi-exit models – in the resource-constrained space environment. Effective scheduling frameworks cater to various streams of inference tasks are scarce because inference performance may deviate from the ideal situation due to changes in task system status, such as task profiles and network state. To this end, we first formulate a gain-aware in-orbit computing inference problem to strike a proper trade-off between inference latency and the number of tasks completed by dynamically selecting optimal early exit points and model partitioning points. We propose an offline dynamic programming-based algorithm that provides an effective solution when comprehensive system details are to be predicted. We have developed an online learning-based method to schedule inference tasks with uncertain and dynamic system statuses in real-world situations. Our evaluation shows that, compared to baseline methods, the online learning-based algorithm can improve task gain by an average of 87.3% across various tasks.

Congestion control is a key enabler for lossless Ethernet at scale. In this paper, we revisit this classic topic from a new perspective, i.e., understanding and exploiting the intrinsic properties of the underlying lossless network. We experimentally and analytically find that the intrinsic properties of lossless networks, such as packet conservation, can indeed provide valuable implications in estimating pipe capacity and the precise number of excessive packets. Besides, we derive principles on how to treat congested flows and victim flows individually to handle HoL blocking efficiently. Then, we propose ACK-driven congestion control (ACC) for lossless Ethernet, which simply resorts to the knowledge of ACK time series (supports ACK coalescing) to exert a temporary halt to exactly drain out excessive packets of congested flows and then match its rate to pipe capacity. Testbed and large-scale simulations demonstrate that ACC ameliorates fundamental issues in lossless Ethernet (e.g., congestion spreading, HoL blocking, and deadlock) and achieves excellent low latency and high throughput performance. For instance, compared with existing schemes, ACC improves the average and 99th percentile FCT performance of small flows by$1.3\sim 3.3\times $and$1.4\sim 11.5\times $, respectively.

Serverless computing is a popular cloud computing paradigm that has found widespread adoption across various online workloads. It allows software engineers to develop cloud applications as a set of functions (called serverless functions ). However, accurately measuring the performance (i.e., end-to-end response latency) of serverless functions is challenging due to the highly dynamic nature of the environment in which they run. To tackle this problem, a potential solution is to apply checks of performance testing techniques to determine how many repetitions of a given serverless function across a range of inputs are needed to cater to the performance fluctuation. However, the available literature lacks performance testing approaches designed explicitly for serverless computing. In this article, we propose the first serverless computing-oriented performance testing (SCOPE) approach. SCOPE takes into account the unique performance characteristics of serverless functions, such as their short execution durations and on-demand triggering. As such, SCOPE is designed as a fine-grained analysis approach. SCOPE incorporates the accuracy check and the consistency check to obtain the accurate and reliable performance of serverless functions. The evaluation shows that SCOPE provides testing results with 97.25% accuracy, 33.83 percentage points higher than the best currently available technique. Moreover, the superiority of SCOPE over the state-of-the-art holds on all functions that we study.

Recent advancements in Low Earth Orbit (LEO) satellites are facilitating the provision of Deep Neural Networks (DNNs)-inherent services to achieve ubiquitous coverage via satellite computing. However, the computational demands and energy consumption of DNN models present significant challenges for satellite computing with limited power and computation resources. Based on the layered characteristics of DNN models, a satellite-ground co-inference strategy has been introduced, which executes certain layers on satellites and the remaining layers on ground servers. Determining the optimal layers for in-orbit processing, however, is non-trivial due to the under-explored energy consumption of satellite computing across different models and restricted yet varying communication conditions of satellite-ground links. In this paper, we first conduct a comprehensive measurement to uncover energy consumption of satellite computing across different layers and models. By summarizing the key observations, we develop a layer-specific energy consumption model tailored to diverse DNN architectures and kernels. We then investigate the energy-efficient satellite-ground co-inference problem and formulate it as an integer-nonlinear programming problem, which presents high computational complexity. To tackle these difficulties, we propose a satellite-ground co-inference algorithm that employs a branch-and-bound strategy, combined with the Sobol sequence and Lagrange multiplier, to reduce complexity and ensure stability across diverse DNN architectures. To evaluate the proposed algorithm, we conduct experiments based on real-world satellite parameters. The results demonstrate that our proposed algorithm can achieve an average energy savings of 96% under various data volumes compared to the existing benchmarks.

Deploying the convolutional neural network (CNN) model on Internet of Things (IoT) devices to provide diverse recognition services has received increasing attention. Due to the limited storage, computing, and other resources of IoT devices, it has become mainstream to first train the CNN model on the edge/cloud server and then send the trained CNN to the IoT device. However, most existing related methods suffer from two limitations, (i) low performance due to service interference or insufficient mutual assistance, and (ii) large memory resources and switching resource overhead. To this end, this article proposes a resource-efficient multiple recognition services framework for IoT devices. The proposed framework is based on the edge server-assisted IoT device training of the CNN model, and the framework includes a deeper weight adaptation (DeepWAdapt) algorithm to mitigate service interference. The DeepWAdapt algorithm consists of a set of learnable masks, and by inserting these masks into the appropriate layers of the CNN model, it mitigates mutual interference between services caused by training a single CNN model for multiple services. Each service has a specific set of masks. These learnable masks work like keys for each service, selecting appropriate and specific features for each service from a shared feature set. Experimental results demonstrate that the DeepWAdapt outperforms other state-of-the-art methods on image-level classification services and pixel-level dense prediction services. Specifically, when executing 40 services based on ResNet18, the proposed DeepWAdapt achieves 66.82% F1-score on the CelebA dataset, which is +2.61% F1-score than the previous state-of-the-art result. In addition, compared with the routing method, our proposed DeepWAdapt also reduces network transmission traffic by approximately 35%.

Intelligent applications rely significantly on the live machine learning pipeline, a couple of deep neural network (DNN) inference services, executed on mobile devices to meet functional requirements while ensuring user data privacy. However, executing these DNN services on resource-constrained mobile devices presents a considerable challenge: low throughput and high energy consumption of inference tasks. To address this issue, we proposeNiagara+, a novel system designed to enhance throughput by jointly scheduling DNN inference services across heterogeneous processors on mobile devices and offloading services to powerful edge servers. To achieve this,Niagara+encounters two critical challenges: unpredictable workload dynamics and high scheduling complexity. To effectively tackle these challenges,Niagara+employs a predictive model to forecast incoming workload patterns and orchestrates service allocation across device heterogeneous processors and edge servers through a combination of two-step offline scheduling optimization and online service dispatching strategies. We implementedNiagara+and conducted comprehensive experiments, demonstrating its superiority over state-of-the-art approaches, reducing DNN service latency by up to 2.6× under high-bandwidth networks and 9.1× under low-bandwidth networks, while consistently meeting stringent inference latency requirements.

Edge computing plays a pivotal role in supporting services necessitating sub-second latency, notably in domains like Industry 4.0 and autonomous driving. However, unpredictable failure occurring at edge servers can result in prolonged response time and decreased service reliability, posing significant risks to both safety and property. Traditional reliability mechanisms, namely task re-execution and task replication, are often inadequate for edge environments. The former struggles to meet the stringent end-to-end service latency requirements, while the latter imposes a high resource consumption burden on resource-limited edge clouds. To address this issue, this paper introduces a novel Flexible Shadow mechanism, where the backup instance, referred to as the Flexible Shadow, is allocated fewer computation resources compared to its primary instance to conserve computation resources, and temporally preempts a portion of resources from neighboring shadows to accelerate when necessary. To support the implementation of this mechanism, we propose the Flexible Shadow Backup Framework, a resource-efficient reliability enhancement framework for edge services through dynamic shadow coordination. This framework integrates three key components: a deployment algorithm for resource allocation, an adjustment algorithm for migration cost-latency tradeoffs, and a reconfiguration algorithm for adaptation optimization. Comprehensive experiments conducted on a Docker-based prototype demonstrate the effectiveness of the Flexible Shadow mechanism, achieving nearly 60% reduction in computing resource consumption compared to traditional approaches while maintaining sub-second latency.

The rapid growth of Low Earth Orbit (LEO) satellites, particularly with the increasing deployment of intelligent computing capabilities using commercial off-the-shelf (COTS) hardware, presents significant opportunities to enhance the quality of in-orbit services. However, the current onboard conditions remain insufficient to enhance model accuracy by increasing model size, and inadequate accuracy hampers the effectiveness of in-orbit services. The satellite-ground federated learning (FL) paradigm, leveraging collaborative fine-tuning, offers a promising solution to continuously improve onboard model performance. Prior studies have focused on optimizing fine-tuning under constraints like limited bandwidth and computational resources, they often overlook two critical challenges: the scarcity and skewness of labeled onboard data and the long revisit cycles of satellites. To address these challenges and better support in-orbit services, this paper designs a realistic simulation methodology for the onboard fine-tuning process and conducts a comprehensive measurement study. Based on insights from the measurement results, we propose an efficient satellite-ground federated fine-tuning system,FedCLR+. In this system, we design a FedCLR algorithm to enhance system accuracy through representation optimization. Additionally, we propose a hybrid bias-compensated strategy to further mitigate accuracy loss by enriching the diversity of aggregation information. Experimental results show thatFedCLR+significantly enhances accuracy by up to 21.61×, reduces transmission volume by an average of 7.29%, and maintaining acceptable additional overhead compared to baselines.

We are delighted to welcome all participants to the 2024 IEEE International Conference on Cloud Computing (CLOUD 2024), which is taking place in the beautiful city of Shenzhen, China, from July 7th to 13th.

IEEE Computer Society (IEEE-CS) Technical Community on Services Computing (TCSVC), http://tab.computer.org/tcsvc, has endorsed the affiliated technical conferences of IEEE World Congress on Services (SERVICES) since 2004. Many innovative growth initiatives have been delivered for the best interests of the worldwide services computing community, a fast-growing R&D community with more than 12,000 active participants.

IEEE Computer Society (IEEE-CS) Technical Community on Services Computing (TCSVC), http://tab.computer.org/tcsvc, has endorsed the affiliated technical conferences of IEEE World Congress on Services (SERVICES) since 2004. Many innovative growth initiatives have been delivered for the best interests of the worldwide services computing community, a fast-growing R&D community with more than 12,000 active participants

Traffic accidents pose a significant risk to human health and property safety. Therefore, to prevent traffic accidents, predicting their risks has garnered growing interest. We argue that a desired prediction solution should demonstrate resilience to the complexity of traffic accidents. In particular, it should adequately consider the regional background, accurately capture both spatial proximity and semantic similarity, and effectively address the sparsity of traffic accidents. However, these factors are often overlooked or difficult to incorporate. In this paper, we propose a novel multi-granularity hierarchical spatio-temporal network. Initially, we innovate by incorporating remote sensing data, facilitating the creation of hierarchical multi-granularity structure and the comprehension of regional background. We construct multiple high-level risk prediction tasks to enhance model's ability to cope with sparsity. Subsequently, to capture both spatial proximity and semantic similarity, region feature and multi-view graph undergo encoding processes to distill effective representations. Additionally, we propose message passing and adaptive temporal attention module that bridges different granularities and dynamically captures time correlations inherent in traffic accident patterns. At last, a multivariate hierarchical loss function is devised considering the complexity of the prediction purpose. Extensive experiments on two real datasets verify the superiority of our model against the state-of-the-art methods.

As a driving force in the advancement of intel-ligent in-orbit applications, DNN models have been gradually integrated into satellites, producing daily latency-constraint and computation-intensive tasks. However, the substantial computation capability of DNN models, coupled with the instability of the satellite-ground link, pose significant challenges, hindering the timely completion of tasks. It becomes necessary to adapt to task stream changes when dealing with tasks requiring latency guarantees, such as dynamic observation tasks on the satellites. To this end, we consider a system model for a collaborative inference system with latency constraints, leveraging the multi-exit and model partition technology. To address this, we propose an algorithm, which is tailored to effectively address the trade-off between task completion and maintaining satisfactory task accuracy by dynamically choosing early-exit and partition points. Simulation evaluations show that our proposed algorithm signif-icantly outperforms baseline algorithms across the task stream with strict latency constraints.

The next point-of-interest (POI) prediction is a significant task in location-based services, yet its complexity arises from the consolidation of spatial and semantic intent. This fusion is subject to the influences of historical preferences, prevailing location, and environmental factors, thereby posing significant challenges. In addition, the uneven POI distribution further complicates the next POI prediction procedure. To address these challenges, we enrich input features and propose an effective deep-learning method within a two-step prediction framework. Our method first incorporates remote sensing data, capturing pivotal environmental context to enhance input features regarding both location and semantics. Subsequently, we employ a region quad-tree structure to integrate urban remote sensing, road network, and POI distribution spaces, aiming to devise a more coherent graph representation method for urban spatial. Leveraging this method, we construct the QR-P graph for the user's historical trajectories to encapsulate historical travel knowledge, thereby augmenting input features with comprehensive spatial and semantic insights. We devise distinct embedding modules to encode these features and employ an attention mechanism to fuse diverse encodings. In the two-step prediction procedure, we initially identify potential spatial zones by predicting user-preferred tiles, followed by pinpointing specific POls of a designated type within the projected tiles. Empirical findings from four real-world location-based social network datasets underscore the remarkable superiority of our proposed approach over competitive baseline methods.

Federated learning (FL) has been widely adopted as a privacy-preserving model training paradigm. However, traditional FL protocol heavily relies on data transmission between clients and servers across the wide-area network (WAN), which is tightly constrained and unreliable, therefore causing expensive communication and slow convergence. To this end, we propose a LAN-aware FL (LanFL) protocol, which can efficiently leverage the network capacity of the local-area network (LAN). By frequent model aggregation among the devices within the same LAN, we can significantly reduce the global aggregation across WAN, thus accelerating the training process. However, due to the unique challenges introduced by LAN, it’s not easy to efficiently utilize LAN resources while preserving the original dignity of FL performance. Therefore, LanFL also incorporates several critical techniques: LAN-aware hierarchical aggregation, intraLAN device topology construction, and inter-LAN heterogeneous bandwidth coordination. Extensive real-world experiments are conducted and the experimental results show that LanFL can significantly accelerate FL training up to $6.0 \times$, while preserving the model accuracy.

The rapid evolution of satellite networks promises to expand global internet service. However, optimizing task allocation for the efficient and sustainable operation of satellite computing presents complex challenges. Existing approaches usually prioritize energy considerations while neglecting economic aspects, which restricts satellite networks from achieving their full economic potential. In this paper, we address this gap by investigating task allocation in satellite computing. Our approach encourages satellites to consistently provide resources and optimizes battery usage, enabling the completion of more tasks and ultimately maximizing profit. The task allocation approach involves two key components: task pricing and task scheduling. Firstly, we introduce a unique task pricing algorithm that adheres to economic properties, establishing a direct link between satellite utilization and financial income, ensuring economically viable satellite operations. Moreover, we develop two distinct task scheduling algorithms tailored for offline and online scenarios, exploiting dynamic programming and reinforcement learning respectively. Extensive simulations demonstrate that our proposed algorithms effectively enhance task completion rates and optimize total satellite profit.

Edge computing plays a pivotal role in supporting services necessitating sub-second latency, notably in domains like Industry 4.0 and autonomous driving. However, unpredictable failure occurring at edge servers can result in prolonged response time and decreased service reliability, posing significant risks to both safety and property. Traditional reliability mechanisms, namely re-execution and replication, are often inadequate for edge environments, with the former often struggle to meet latency requirements and the latter imposing a high resource consumption burden on resource-limited edge clouds. To address this issue, this paper introduces a novel "Flexible Shadow" mechanism, where the backup instance, referred to as the "Flexible Shadow", is allocated fewer computation resources compared to its primary instance to conserve computation resources, and temporally preempts a portion of resources from neighboring shadows to accelerate when necessary. To tackle the implementation challenges of this framework arising from diverse service requirements, dynamic nature of edge environments and potential deadlock in the reconfiguration process, we devised the Flexible Shadow Deployment Algorithm for accurate shadow deployment and the Flexible Shadow Reconfiguration Algorithm for dynamic strategy adjustment. We have implemented our Flexible Shadow framework on Docker and evaluated it via comprehensive experiments. The experiment results demonstrate a nearly 60% reduction in computing resource consumption while ensuring sub-second latency.

With the rapid proliferation of large Low Earth Orbit (LEO) satellite constellations, a huge amount of in-orbit data is generated and needs to be transmitted to the ground for processing. However, traditional LEO satellite constellations, which downlink raw data to the ground, are significantly restricted in transmission capability. Orbital edge computing (OEC), which exploits the computation capacities of LEO satellites and processes the raw data in orbit, is envisioned as a promising solution to relieve the downlink burden. Yet, with OEC, the bottleneck is shifted to the inelastic computation capacities. The computational bottleneck arises from two primary challenges that existing satellite systems have not adequately addressed: the inability to process all captured images and the limited energy supply available for satellite operations. In this work, we seek to fully exploit the scarce satellite computation and communication resources to achieve satellite-ground collaboration and present a satellite-ground collaborative system named TargetFuse for onboard object detection. TargetFuse incorporates a combination of techniques to minimize detection errors under energy and bandwidth constraints. Extensive experiments show that TargetFuse can reduce detection errors by 3.4× on average, compared to onboard computing. TargetFuse achieves a 9.6× improvement in bandwidth efficiency compared to the vanilla baseline under the limited bandwidth budget constraint.

Recent advancements in Low Earth Orbit (LEO) satellites are facilitating the provision of Deep Neural Networks (DNNs)-inherent services to achieve ubiquitous coverage via satellite computing. However, the computational demands and energy consumption of DNN models pose significant challenges for satellite computing with limited power and computation resources. Based on the hierarchical characteristics of DNN models, we propose a satellite-ground co-inference strategy that executing certain layers on satellites and the remaining layers on ground servers. However, identifying the optimal layers for in-orbit processing with latency constraints is challenging due to the uncertain energy consumption across diverse models. To explore the correlation between energy consumption and layer types, we conduct comprehensive measurements on a hardware device commonly found in commercial LEO satellites and develop a layer-based energy consumption prediction model. Then, we formulate an optimization problem of minimizing the energy consumption on the satellite within the latency constraint as an integer nonlinear programming problem. Solving this problem is difficult due to combinatorial explosion in the discrete solution space. To address this, we propose an improved algorithm based on genetic algorithms. Using configurations from a real satellite, we conduct simulation experiments, concluding that our algorithm significantly improves energy savings by an average of 27 ×.

In the wake of the rapid deployment of large-scale low-Earth orbit satellite constellations, exploiting the full computing potential of Commercial Off-The-Shelf (COTS) devices in these environments has become a pressing issue. However, understanding this problem is far from straightforward due to the inherent differences between the terrestrial infrastructure and the satellite platform in space. In this paper, we take an important step towards closing this knowledge gap by presenting the first measurement study on the thermal control, power management, and performance of COTS computing devices on satellites. Our measurements reveal that the satellite platform and COTS computing devices significantly interplay in terms of the temperature and energy, forming the main constraints on satellite computing. Further, we analyze the critical factors that shape the characteristics of onboard COTS computing devices. We provide guidelines for future research on optimizing the use of such devices for computing purposes. Finally, we have released the datasets to facilitate further study in satellite computing.

In the current AI era, mobile devices such as smartphones are tasked with executing a myriad of deep neural networks (DNNs) locally. It presents a complex landscape, as these models are highly fragmented in terms of architecture, operators, and implementations. Such fragmentation poses significant challenges to the co-optimization of hardware, systems, and algorithms for efficient and scalable mobile AI.

App stores allow developers to globally distribute their apps to gain more users and attention. In the highly competitive market of app stores, developers need to cater to a large number of users spanning multiple countries. We posit that the characteristics of diverse geographical, linguistic, cultural, societal, and economic environments may impact the adoption of apps. In this paper, we take the first step to characterize popular apps across over 150 countries worldwide, and explore the potential correlations to a number of underlying factors including geography, language as well as cultural, societal, and economic dimensions. Our study is based on a longitudinal (one-year) dataset of daily app popularity from the iOS app stores, covering 154 regions around the world. We reveal that app popularity shows great diversity across the world, while similarities exist among countries that share geographical proximity and linguistic convergence. The differences in app popularity across regions can be further correlated with the cultural model and socioeconomic indices we adopt. On top of the dataset and findings, we implement a prediction task that contributes to app distribution, helping developers choose the right market to distribute and promote their apps. To the best of our knowledge, we are the first to attempt to provide a global understanding of the characteristics of app popularity across the mobile app ecosystem. Our observations can benefit stakeholders in the ecosystem, striving to improve app uptake.

Speech serves as a ubiquitous input interface for embedded mobile devices. Cloud-based solutions, while offering powerful speech understanding services, raise significant concerns regarding user privacy. To address this, disentanglement-based encoders have been proposed to remove sensitive information from speech signals without compromising the speech understanding functionality. However, these encoders demand high memory usage and computation complexity, making them impractical for resource-constrained wimpy devices. Our solution is based on a key observation that speech understanding hinges on long-term dependency knowledge of the entire utterance, in contrast to privacy-sensitive elements that are short-term dependent. Exploiting this observation, we propose SILENCE, a lightweight system that selectively obscuring short-term details, without damaging the long-term dependent speech understanding performance. The crucial part of SILENCE is a differential mask generator derived from interpretable learning to automatically configure the masking process. We have implemented SILENCE on the STM32H7 microcontroller and evaluate its efficacy under different attacking scenarios. Our results demonstrate that SILENCE offers speech understanding performance and privacy protection capacity comparable to existing encoders, while achieving up to 53.3$\times$ speedup and 134.1$\times$ reduction in memory footprint.

Recently, Remote Direct Memory Access (RDMA) is gradually gaining popularity in multi-tenant datacenters due to its high performance and low CPU utilization. RDMA requires a lossless underlying network to fully realize its potential. Thus, hop-by-hop Priority Flow Control (PFC) is deployed to ensure losslessness, and congestion control is also needed to allocate per-flow bandwidth. However, we find that existing RDMA congestion control schemes in multi-tenant datacenters fail to achieve rapid rate convergence due to the heuristic nature and may exacerbate PFC side effects including head-of-line blocking, unfairness, and even deadlocks. In this paper, we propose an enhancement framework for RDMA congestion control called RDI. RDI has several noteworthy properties. First, RDI utilizes valuable receiver-side information and network congestion level to provide a precise guide rate for congested flows, thus alleviating congestion and corresponding PFC issues; Second, RDI is transparent to tenants and compatible with existing RDMA network architectures without the need of modifying in-network devices. We evaluate RDI under realistic traffic traces. The results show that the congestion control scheme enhanced by RDI significantly outperforms the original in terms of throughput and flow completion time, while also reducing the side effects of PFC. For instance, RDI-enhanced congestion control shortens by up to 44% and 60% for the 99th-percentile tail and average FCT, respectively.

Small satellites (SmallSats) are now widely used in various fields, such as real-time communication and earth observation. These increasingly complex space applications face limited support from conventional radiation-hardened processors onboard. Hence, many SmallSats are designed to utilize high performance commercial off-the-shelf (COTS) computing devices to address this problem but it remains unclear how the unique energy and thermal characteristics of SmallSats impact computing efficiency onboard. This work conducts a systematic and quantitative measurement study of COTS devices’ computing efficiency on two real orbiting SmallSats. The key findings are: 1) inadequate energy management may lead to electricity wastage in sunlit zones and shortages in eclipse zones, impacting onboard computing availability and 2) the weak heat dissipation onboard may compromise COTS computing efficiency by incurring thermal throttling. To address such challenges, we design ProScale, a lightweight application-aware power management and thermal control system to improve computing efficiency under both electrical and thermal energy constraints. Evaluation shows that ProScale can improve the average task completion latency by $2.1 \times$ for computation-intensive applications compared with baselines.

Deep neural networks now have become the de-facto standard for sequential recommendation. In the existing techniques, an embedding vector is assigned for each item, encoding all the characteristics of the latter in latent space. Then, the recommendation is transferred to devising a similarity metric to recommend user's next behavior. Here, we consider each dimension of an embedding vector as a (latent) feature. Though effective, it is unknown which feature carries what semantics toward the item. Actually, in reality, this merit is highly preferable since a specific group of features could induce a particular relation among the items while the others are in vain. Unfortunately, the previous treatment overlooks the feature semantic learning at such a fine-grained level. When each item contains multiple latent aspects, which however is prevalent in real-world, the relations between items are very complex. The existing solutions are easy to fail on better recommendation performance. It is necessary to disentangle the item embeddings and extract credible features in a context-aware manner.

AI is making the Web an even cooler place, but also introduces serious privacy risks due to the extensive user data collection. Federated learning (FL), as a privacy-preserving machine learning paradigm, enables mobile devices to collaboratively learn a shared prediction model while keeping all training data on devices. However, a key obstacle towards practical cross-device FL training is huge energy consumption, especially for lightweight mobile devices. In this work, we perform the first-of-its-kind analysis of improving FL performance through low-precision training with an energy-friendly Digital Signal Processor (DSP) on mobile devices. We first demonstrate that directly integrating the state-of-the-art INT8 (8-bit integer) training algorithm and classic FL protocols will significantly degrade the model accuracy. Moreover, we observe that there are still unavoidable frequent quantization operations on devices that cause extreme load stress on DSP-enabled INT8 training. To address the above challenges, we present Q-FedUpdate, an FL framework that efficiently preserves model accuracy with ultra-low energy consumption. It maintains a global full-precision model and allows the tiny model updates to be continuously accumulated, instead of being erased by the quantization. Furthermore, it introduces pipelining technology to parallel CPU-based quantization and DSP-enabled training, which reduces the floating-point computation overhead of frequent data quantization. Extensive experiments show that Q-FedUpdate can effectively reduce the on-device energy consumption by 21×, and accelerate the FL convergence by 6.1× with only 2% accuracy loss.

Edge computing serves as a critical solution for latency-sensitive services on mobile and IoT devices. However, the high cost and limited edge resources present significant challenges for service and infrastructure providers in establishing efficient collaborations, particularly with conflicting profit objectives. Inspired by the pseudo elbow formation of octopuses, we propose a multi-participant double auction for resource allocation and pricing between service and infrastructure providers. We introduce a neutral third-party auctioneer to eliminate direct bargaining among participants, leading to an improved amount of allocated resources and matching efficiency. The presence of heterogeneous participants, many-to-many mapping and an advisable payment strategy that satisfies economic properties exacerbate the difficulty. To address these challenges, we propose a Matching and Pricing Resource Allocation algorithm for a long-term steady market, and a Truthful Resource Allocation algorithm for a short-term market. Simulation results demonstrate that the proposed algorithms exhibit superior performance not only in maximizing social welfare and utility of both service and infrastructure providers, but also in improving resource utilization.

Computation offloading involves transmitting compute-intensive tasks from mobile users (MUs) to edge environments. This compensates for the limitations of terminal devices in computing performance and resource storage capabilities. However, with the growing demand for offloading compute-intensive tasks, the burden on edge server networks has intensified. On the other hand, the response time of application has increased, leading to a decline in user experience. Therefore, selecting a reliability metric caused by congestion becomes an important issue for analyzing the quantitative characteristics of edge networks. The current research primarily focuses on the precise validation offloading task results. Typically, methods like task redistribution and third-party assistance are employed to ensure task reliability. However, these measures lead to increased time and energy consumption. In this paper, we propose a more versatile reliability metric based on the probability distribution of average waiting time in a multi-queue model. Additionally, to incentive MUs to offload more tasks and enhance the economic utility of edge server providers (ESPs), we employ a Stackelberg game to model the dynamic interaction between ESPs and MUs. Finally, we utilize the alternating direction method of multipliers (ADMM) algorithm to derive the optimal strategies for ESPs and MUs. Simulation results demonstrate that our proposed approach surpasses the baselines in terms of reliability indicator. Moreover, it achieves faster convergence and decision-making in comparison to conventional heuristic methods.

Age of Information (AoI) has gained widespread usage and emerged as a pivotal metric for assessing timeliness performance in information-update systems. Such systems often entail service requirements for rapidly obtaining requested data in real-time. For instance, in the financial market, users rely on up-to-date and low-latency information to make appropriate trading decisions to maximize profits within their financial budge. Much of the existing research on real-time services focuses on ensuring AoI or service-level latency, but there is a growing demand for joint optimization of these two metrics to accommodate a broader range of potential applications. Therefore, this article investigates the problem of minimizing AoI within the context of statistical latency guarantees. To tackle the critical challenges posed by the joint modeling of AoI and statistical latency, system uncertainty, as well as tradeoff between performance and user’s budget, we employ a replication scheme to ensure both AoI and statistical service-level latency. To address the critical challenges posed by the unknown distribution in data updating processes and response times across providers, we formulate the AoI minimization problem with statistical latency constraints as a combinatorial multiarmed bandit problem utilizing the Lyapunov optimization theory. Subsequently, we propose an online learning-based request replication algorithm to address this problem. Our proposed algorithm achieves a cumulative regret of$O(T\sqrt {\log (T)})$compared to the genie-aided algorithm. Simulation results demonstrate the superior performance of the proposed algorithm against benchmarks.

With the significant increase in the number of motor vehicles, road-related issues such as traffic congestion and accidents have also escalated. The development of an accurate and efficient traffic flow forecasting model is essential for helping car owners plan their journeys. Despite advancements in forecasting models, there are three remaining issues: (i) failing to effectively use cyclical data; (ii) failing to adequately capture spatial dependencies; and (iii) high time complexity and memory usage. To tackle the aforementioned challenges, we present a novel Spatial-Temporal Graph Attention Gated Recurrent Transformer Network (STGAGRTN) for traffic flow forecasting. Specifically, the use of a Spatial Transformer module allows for the extraction of dynamic spatial dependencies among individual nodes, going beyond the limitation of only considering neighboring nodes. Subsequently, we propose a Temporal Transformer to extract periodic information from traffic data and capture long-term dependencies. Additionally, we utilize two additional classical techniques to complement the aforementioned modules for extracting characteristics. By incorporating comprehensive spatial-temporal characteristics into our model, we can accurately predict multiple nodes simultaneously. Finally, we have successfully optimized the computational complexity of the Transformer module from O(n2) to O(nlogn). Our model has undergone extensive testing on four authentic datasets, providing compelling evidence of its superior predictive capabilities.

As infrastructure deployment continues to expand worldwide, the development of the Internet of Vehicles has become increasingly feasible. With widespread cellular connectivity and powerful roadside computing capabilities, advanced driving assistance systems can now rely on roadside decision models in addition to vehicle-side ones, overcoming the limitations of a single vehicle’s perception range. This shift has led to improvements in manufacturing efficiency, cruising range, and battery life of intelligent vehicles. However, maintaining the ultra-low latency and high-reliability requirements of on-vehicle services is still a challenge due to air interface fluctuations and edge server computing loads, which could potentially jeopardize driving safety. To tackle this issue, we conducted real-world measurements of edge server access delay in LTE and 5G cellular networks. Our analysis identified key factors affecting delay distribution, leading to the development of an approximate fitting function for the delay probability density function. We also proposed a reliability-aware task replication algorithm that leverages delay samples and edge server status information to make real-time task replication and offloading decisions, minimizing replication while ensuring service reliability. Simulations based on real-world datasets indicate our approach reduces task completion delay by up to 42.11% and limits the maximum task replication redundancy peak value to 63.37%, effectively ensuring the reliability of on-vehicle services during driving.

Privacy, security and data governance constraints rule out a brute force process in the integration of cross-silo data, which inherits the development of the Internet of Things. Federated learning is proposed to ensure that all parties can collaboratively complete the training task while the data is not out of the local. Vertical federated learning is a specialization of federated learning for distributed features. To preserve privacy, homomorphic encryption is applied to enable encrypted operations without decryption. Nevertheless, together with a robust security guarantee, homomorphic encryption brings extra communication and computation overhead. In this paper, we analyze the current bottlenecks of vertical federated learning under homomorphic encryption comprehensively and numerically. We propose a straggler-resilient and computation-efficient accelerating system that reduces the communication overhead in heterogeneous scenarios by 65.26% at most and reduces the computation overhead caused by homomorphic encryption by 40.66% at most. Our system can improve the robustness and efficiency of the current vertical federated learning framework without loss of security.

Extracting features from image data on Internet of Things (IoT) devices to reduce the amount of data that needs to be uploaded to cloud/edge servers has received increasing attention. However, most of the existing related approaches suffer from two major limitations, (i) low performance and high network traffic, and (ii) a lot of storage resource consumption. To this end, we propose a resource-efficient feature extraction framework for image processing in IoT devices. The proposed framework consists of the edge-assisted extractor generation method and the NestE method. The extractor generated by the edge-assisted extractor generation method can extract the features required by the application, which can not only avoid the IoT device uploading useless feature data but also improve application performance. The proposed NestE generates a nonredundant subextractor by splitting the extractor into multiple subextractors, removing redundant subextractors, and nesting small-capacity subextractors in large-capacity subextractors in a parameter-sharing manner. Compared with deploying multiple independent subextractors on IoT devices, deploying the nonredundant multifunctional extractor can save considerable storage resources and switching overhead. Extensive experimental results show that the proposed framework reduces the storage footprint by approximately 90.7% and switching overhead by approximately 92.4% compared with deploying independent subextractors when using the classical principal component analysis algorithm.

Seamless cross-edge migration for real-time rendering applications is challenging. The strong interactive nature of real-time rendering applications demands a downtime lower than 15ms to achieve an imperceptible migration. Existing methods based on virtual machine migration and container migration suffer from unpleasant downtime brought by dirty page retransmission-induced repeated memory data copy and the shared storage failure-induced extensive disk data copy. In this paper, we propose Cloud-assisted Service Migration (CSM) which leverages cloud-edge collaboration to achieve seamless service migration for real-time rendering applications. CSM improves service migration user experience in three folds: First, it introduces a dual rendering mechanism to bypass the peer-to-peer data copy and compresses the freezing stage. Second, a user equipment-centric session switch mechanism is proposed to save time by well coordinating application session switches and 5G user plane session switches. Third, a smooth switching mechanism is leveraged to prevent unpleasant frame flickers during session switching. We implement CSM in edge-rendering multiplayer games and deploy it on a 5G test bed with a full-stack user plane protocol stack. The evaluation results show that CSM can reduce downtime to < 14ms and the service migration process is user imperceptible.

Large constellations of Low Earth Orbit (LEO) satellites have been launched for Earth observation and satellite-ground communication, which collect massive imagery and sensor data. These data can enhance the AI capabilities of satellites to address global challenges such as real-time disaster navigation and mitigation. Prior studies proposed leveraging federated learning (FL) across satellite-ground to collaboratively train a share machine learning (ML) model in a privacy-preserving mechanism. However, they mostly focus on single unique challenges such as limited ground-to-satellite bandwidth, short connection window, and long connection cycle, while ignoring the completeness of these challenges in deploying efficient FL frameworks in space. In this paper, we propose an efficient satellite-ground FL framework, SatelliteFL, to address these three challenges collectively. Its key idea is to ensure that each satellite must complete per-round training within each connection window. Moreover, we design a progressive block-wise quantization algorithm that determines a unique bitwidth for each block of the ML model to maximize the model utility while not exceeding the connection window. We evaluate SatelliteFL by plugging an implemented FL platform into real-world satellite networks and satellite images. The results show that SatelliteFL highly accelerates the convergence by up to 2.8× and improves the bandwidth utilization ratio by up to 9.3× compared to the state-of-the-art methods.

Deploying deep learning (DL) on mobile devices has been a notable trend in recent years. To support fast inference of on-device DL, DL libraries play a critical role as algorithms and hardware do. Unfortunately, no prior work ever dives deep into the ecosystem of modern DL libraries and provides quantitative results on their performance. In this paper, we first build a comprehensive benchmark that includes 6 representative DL libraries and 15 diversified DL models. Then we perform extensive experiments on 10 mobile devices, and the results reveal the current landscape of mobile DL libraries. For example, we find that the best-performing DL library is severely fragmented across different models and hardware, and the gap between DL libraries can be rather huge. In fact, the impacts of DL libraries can overwhelm the optimizations from algorithms or hardware, e.g., model quantization and GPU/DSP-based heterogeneous computing. Motivated by the fragmented performance of DL libraries across models and hardware, we propose an effective DL Library selection framework to obtain the optimal library on a new dataset that has been created. We evaluate the DL Library selection algorithm, and the results show that the framework at it can improve the prediction accuracy by about 10% than benchmark approaches on average.

In edge environment, many smart application instances are triggered randomly by resource-constrained Internet of Things (IoT) devices. These application instances usually consist of dependent computation components, which can be modeled as workflows in different shapes and sizes. Due to the limited computing power of IoT devices, a common approach is to schedule partial computation components of multiple workflow instances to the resource-rich edge servers to execute. However, how to schedule the stochastically arrived multiple workflow instances in edge environment with the minimum average completion time is still a challenging issue. To address such an issue, in this paper, we adopt the graph convolution neural network to transform multiple workflow instances with different shapes and sizes into embeddings, and formulate the online multiple workflow scheduling problem as a finite Markov decision process. Furthermore, we propose a policy gradient learning-based online multiple workflow scheduling scheme (PG-OMWS) to optimize the average completion time of all workflow instances. Extensive experiments are conducted on the synthetic workflows with various shapes and sizes. The experimental results demonstrate that the PG-OMWS scheme can effectively schedule the stochastically arrived multiple workflow instances, and achieve the lowest average completion time compared with four baseline algorithms in edge environments with different scales.

Internet-of-Things (IoT) applications from many industries, such as transportation (maritime, road, rail, air) and fleet management, offshore monitoring, and farming are located in remote areas without cellular connectivity. Such IoT applications continuously generate stream data with hidden values that need to unveiled in real time. Streaming analytics is emerging as a popular type of Big Data analytics to process large volume of stream data for IoT applications in remote regions. Built upon terrestrial-satellite integrated networks, Satellite Edge Computing (SEC) equipped with computing resource in satellites has been envisioning as a key enabling technology to timely analyze stream data of IoT applications in remote regions on the Earth. Considering the dynamically-moving property of satellites in an SEC network, it is vital to optimize the responsiveness of each Big Data analytical query, such that none of such queries takes much longer time to wait for available satellites with sufficient computing resource. Furthermore, the uncertain data volumes of Big Data queries in SEC networks make the resources in satellites fragmented, particularly when the collaboration among satellites is intermittent. Therefore, directly application of existing methods may not guarantee the timelineness of streaming analytics in an SEC network. To address the afore-mentioned unique challenges of streaming analytics in SEC networks, it is urgent to design new algorithms and methods for timely Big Data processing. Specifically, we use theflow timeto capture the responsiveness of streaming analytics in satellite edge computing, which is the time between the generation of the first unit of a dataset and the finish time of the data processing. We consider the flow time minimization problem for query evaluation of Big Data analytics in an SEC network with the aim of minimizing the average flow time of Big Data analytical queries, under an assumption of uncertain volumes of datasets. To this end, we first propose an approximation algorithm with a provable approximation ratio for the offline version of the flow time minimization problem. We then devise an online learning algorithm, referred to the customized Lipschitz bandit learning algorithm, with a bounded regret for the online version of the problem. We finally evaluate the performance of the proposed algorithms in a real SEC network topology. Experiment results show that the performance of the proposed algorithm outperforms its counterparts by at least 13% in terms of flow time.

Industry 5.0 facilitates the intelligent upgrade of smart factories in Industrial Internet of Things (IIoT), and also introduces a plethora of data processing challenges. Mobile edge computing offloads data to edge servers for processing, easing the data processing pressure and reducing system cost. However, smart factories contain numerous sensitive information, and offloading them to edge servers directly may pose a risk of data leakage. To address these challenges, we investigate a localedge collaborative smart factory system. Specifically, we firstly model the tasks as a directed acyclic graph, and formulate the offloading problem as a Markov decision process, considering the optimization of latency, energy consumption and the number of overtime tasks. Then, we propose a security-aware computation offloading method using federated reinforcement learning in IIoT, named SCOF. SCOF employs federated learning, keeping data for local computation and uploading model parameters to edge servers for aggregation. The data transmission passes through an artificial noise channel to protect against eavesdropping. Meanwhile, SCOF utilizes differential privacy to protect data security and employs deep reinforcement learning for selecting near-optimal offloading decisions. Finally, abundant experiments are conducted under a real dataset. The results show that SCOF has better perfomance than the state-of-the-art baseline algorithms.

Satellite edge computing can incur dramatically increased energy demand onboard, which is met by satellite batteries during eclipses. Excessive energy usage during regular operations accelerates battery wear. Therefore, it is important and timely to optimize the energy consumption onboard to extend satellite batteries life. This paper investigates battery-aware energy optimization for satellite edge computing under energy harvesting dynamics and wireless environment uncertainty. Inspired by the periodical energy harvesting and satellite-ground connection, we develop a pattern-aware online energy scheduling algorithm within an online convex optimization framework. This learning algorithm achieves theoretical guarantees of no regret and gradually zeroing constraint violations. We further exploit inter-satellites collaboration to extend the average battery life in a whole constellation where satellites have different battery capacity degradation. Trace-driven simulations show that our algorithm can significantly extend the battery life by 1.32× and effectively adapt to the energy harvesting dynamics and wireless environment uncertainty.

The widespread adoption of Industrial Internet of Things (IIoT) has significantly transformed various aspects of industrial manufacturing. However, the massive volume and complexity of IIoT data highlight the need for innovative solutions to enhance the overall performance of IIoT systems. In this regard, mobile edge computing, assisted by deep reinforcement learning, can alleviate the burden on IIoT systems through computation offloading. Nevertheless, in increasingly digitized industrial environments, how to make real-time, efficient task offloading decisions remains a subject of deep exploration. To address this issue, we propose a two-stage resource allocation and task offloading method, named MCORM. In the first stage, we use the Combinatorial Upper Confidence Bound algorithm, based on the combinatorial multi-armed bandit problem, to solve the resource allocation problem. In the second stage, the Multi-Agent Proximal Policy Optimization algorithm is employed to determine the approximate optimal offloading strategy. Specifically, the Convolutional Block Attention Module is utilized to process observation information, focusing on important features. Finally, extensive experiments are conducted using both simulated and real datasets. The results demonstrate that our proposed algorithm MCORM can reduce latency and energy consumption effectively, promoting efficient industrial production.

As a critical and prevalent service in future mobile networks, virtual reality (VR) is latency-sensitive and power-hungry, bringing out the optimization problem of trade-off among power saving, delay, and resource utilization. Content caching and render offloading are deemed as promising solutions to meet the stringent requirements of VR on data transmission speed and end-to-end latency. In this article, we propose a novel distributed computing framework based on multi-agent deep deterministic policy gradient (MADDPG) for joint optimizing terminal-cooperative caching and offloading for VR tasks. Since the individual VR user can hardly reach the optimal actions based on its limited local observed states and samples, MADDPG with centralized training and distributed execution is exploited to solve the above challenge. In addition, the generative adversarial network (GAN) is introduced to obtain experience-enhanced agents in the offline training phase and to achieve an optimal allocation to minimize energy consumption in the online inferring phase. The Nash equilibrium is proven in the case that the distribution of finite real VR data samples is well imitated and complemented by GAN. Numerical results demonstrate that our algorithm has significant superiorities in terms of convergence performance and energy consumption over other benchmarks.

The rapid development of Low Earth Orbit (LEO) satellite constellations offers significant potential for in-orbit services, particularly in mitigating the impact of sudden natural disasters. However, the massive data collected by these satellites are often large and severely constrained by limited transmission capabilities when sending data to the ground. Satellite computing, which utilizes onboard computational capacity to process data before transmission, presents a promising solution to alleviate the downlink burden. Nonetheless, this paradigm introduces another bottleneck: limited onboard computing capacity, resulting in slow in-orbit processing and poor results. Current satellite computing systems struggle to efficiently address both data transmission and computing bottlenecks, particularly for urgent disaster services that demand accurate and timely results. Thus, we introduce an efficient satellite computing system designed to jointly mitigate these bottlenecks, thereby providing better service. The core idea is to utilize onboard computing capacity for swift in-orbit annotation of image regions, enabling adaptive compression and download based on annotation confidence and perceived downlink availability. Once the data is downloaded, image restoration and re-inference are performed on the ground to enhance accuracy. Compared to satellite-only inference, our system demonstrates an average improvement in inference accuracy of 3.8%. Furthermore, compared to ground-only inference, with only a 2.8% accuracy loss, our system achieves a 38.4% reduction in response time and saves 71.6% of downlink volume on average.

The space-air-ground-ocean integrated network (SAGOI-Net) has become the focus of research in recent years, which has the characteristics of wide coverage and strong adaptability. However, due to the influence of multiple heterogeneous network segments, this network is unable to provide excellent quality of service (QoS). Based on the software-defined network and virtual network architecture, we abstract SAGOI-Net as a three-layer heterogeneous physical network resource, and propose a multi-domain virtual network embedding solution to optimize QoS. Specifically, before virtual network embedding, we collected SAGOI-Net's resource information through software-defined network and modeled it. In the virtual network embedding process, we first classify the virtual network request through K-means, and dynamically adjust the reward function to use reinforcement learning to solve the optimal virtual network embedding strategy. Finally, simulation experiments verify the effectiveness of the scheme.

IEEE Computer Society (IEEE-CS) Technical Community on Services Computing (TCSVC), http://tab.computer.org/tcsvc, has endorsed the affiliated technical conferences of IEEE World Congress on Services (SERVICES) since 2004. Many innovative growth initiatives have been delivered for the best interests of the worldwide services computing community, a fast-growing R&D community with more than 12,000 active participants.

Recommender systems play a pivotal role in navigating users through vast reservoirs of information. However, data sparseness can compromise recommendation accuracy, making it challenging to improve recommendation performance. To address this issue, researchers have explored incorporating multiple data types. Yet, this approach can introduce noise that impairs the recommendations' accuracy. Therefore, it is crucial to denoise the data to enhance recommendation quality. This tutorial highlights the importance of data denoising metrics for improving the accuracy and quality of recommendations. Four groups of data denoising metrics are introduced: feature, item, pattern, and modality level. For each group, various denoising methods are presented. The tutorial emphasizes the significance of selecting the right data denoising methods to enhance recommendation quality. It provides valuable guidance for practitioners and researchers implementing reliable data denoising metrics in recommender systems. Finally, the tutorial proposes open research questions for future studies, making it a valuable resource for the research community.

Multi-task learning (MTL) seeks to learn a single model to accomplish multiple tasks by leveraging shared information among the tasks. Existing MTL models, however, have been known to suffer from negative interference among tasks. Efforts to mitigate task interference have focused on either loss/gradient balancing or implicit parameter partitioning with partial overlaps among the tasks. In this paper, we propose ETR-NLP to mitigate task interference through a synergistic combination of non-learnable primitives (NLPs) and explicit task routing (ETR). Our key idea is to employ non-learnable primitives to extract a diverse set of task-agnostic features and recombine them into a shared branch common to all tasks and explicit task-specific branches reserved for each task. The non-learnable primitives and the explicit decoupling of learnable parameters into shared and task-specific ones afford the flexibility needed for minimizing task interference. We evaluate the efficacy of ETR-NLP networks for both image-level classification and pixel-level dense prediction MTL problems. Experimental results indicate that ETR-NLP significantly outperforms state-of-the-art baselines with fewer learnable parameters and similar FLOPs across all datasets. Code is available at this URL.

Content distribution networks play a crucial role in the services provided by streaming applications, but current terrestrial CDNs have uneven coverage and poor performance in some remote and rural areas. Satellite communications have emerged as an effective way to address these issues. However, the satellite environment is highly dynamic, and how to reasonably deploy CDNs in the satellite constellation to realize efficient content caching and distribution is an urgent problem to be solved. To this end, this paper proposes a content caching and content distribution approach for satellite CDNs. The approach uses Lyapunov optimization, Gibbs sampling, and matching theory to control the traffic cost of satellite-cached content copies while reducing the content distribution latency. A cost-effective content cache deployment and distribution strategy is proposed for large low-orbiting satellite constellations operating at high speeds. Extensive simulation results show that the algorithm can quickly converge to favorable values compared to the reference algorithm, effectively reducing latency for end users while maintaining low deployment traffic costs.

Satellite networking has been a very active research area in both academia and industry recently. Starlink provides high-speed, low-latency Internet access. More companies have claimed to provide similar services soon. It is expected that different industries will embrace satellite network business in the future. This paper reviews the achievements in satellite-related works in standard organizations (IETF and 3GPP) from the networking perspective (Layer 2/3 and above). The issues and challenges in different areas are analyzed, such as IP networking for massive satellite constellations, wireless network architecture for Non-Terrestrial Network (NTN) integration, etc. Based on the latest developments, the paper envisions the evolution direction and outlines the study items for satellite networks in 6G.

Edge computing is extending the frontier of computation beyond terrestrial boundaries, with Low Earth Orbit (LEO) constellations emerging as the cutting-edge paradigm, aspiring to deliver ubiquitous computing capabilities globally. However, the unresolved service reliability issues within LEO constellations continue to barricade the realization of the vision for space computing. Terrestrial-native classical methods struggle to meet the stringent environmental conditions of space computing and often fail to satisfy the rigorous requirements of LEO constellations for failure response times and computational resource overhead. To address these challenges, we introduce the LEO Service Reliability Enhancement Framework (LSREF). LSREF employs the Variable Speed Replication technology to balance failure response times and computational resource overhead and adopts a decentralized design, more congruent with the characteristics of LEO constellations. Specifically, to counter the immense scale and high dynamism of LEO constellations, LSREF proposes orbital plane-based autonomous domains and leverages the Satellite Autonomous Backup Deployment Algorithm in conjunction with the Registry Polling Mechanism to enable autonomous decision-making for service backup strategies on each satellite. Our simulation experiments demonstrate that, compared to classical methods, LSREF reduces average failure response times by 10.76% and diminishes computational resource consumption by nearly 15.53%.

Deploying high-performance convolutional neural network (CNN) models on low-earth orbit (LEO) satellites for rapid remote sensing image processing has attracted significant interest from industry and academia. However, the limited resources available on LEO satellites contrast with the demands of resource-intensive CNN models, necessitating the adoption of ground-station server assistance for training and updating these models. Existing approaches often require large floating-point operations (FLOPs) and substantial model parameter transmissions, presenting considerable challenges. To address these issues, this paper introduces a ground-station server-assisted framework. With the proposed framework, each layer of the CNN model contains only one learnable feature map (called the seed feature map) from which other feature maps are generated based on specific rules. The hyperparameters of these rules are randomly generated instead of being trained, thus enabling the generation of multiple feature maps from the seed feature map and significantly reducing FLOPs. Furthermore, since the random hyperparameters can be saved using a few random seeds, the ground station server assistance can be facilitated in updating the CNN model deployed on the LEO satellite. Experimental results on the ISPRS Vaihingen, ISPRS Potsdam, UAVid, and LoveDA datasets for semantic segmentation services demonstrate that the proposed framework outperforms existing state-of-the-art approaches. In particular, the SineFM-based model achieves a higher mIoU than the UNetFormer on the UAVid dataset, with 3.3 × fewer parameters and 2.2 × fewer FLOPs.

Existing satellite constellations typically lack on-orbit processing capabilities, and on-orbit data backhaul face to reliability and effiency challenges. Tiansuan constellation is with the capability of providing satellite edge computing services, which can support complex data coding transmission methods. To effectively support effective data backhaul for existing satellite constellations, we propose a reliable data transmission service in Tiansuan constellation. The service adopts the adaptive coding transmission strategy to adjust the optimal size of the encoded data packet according to real-time satellite network conditions. An adaptive encoding transmission approach is proposed to reduce end-to-end delay. Experimental results based on the ground simulation platform demonstrate that the proposed method can effectively reduce end-to-end delay and improve effective data transmission rate.

It has become mainstream for mobile devices to offload latency-sensitive applications to edge servers for execution to meet low-latency requirements. However, the existing related studies lack the consideration of application hit ratio, which makes them unable to meet the increasingly complex offloading of multi-applications including multi-tasks. To this end, this paper proposes a Multi-task offloading and Service placement optimization (MSO) method with the goal of maximizing the application hit ratio to provide high-quality service. The proposed MSO is constructed with Improved Multi-Agent Q-Learning (IMAQL) and load-balancing algorithms. IMAQL aims to learn an optimal service placement policy by using Q-learning techniques. Next, the load-balancing algorithm is designed to offload tasks according to the service placement policy. To verify the effectiveness of the MSO method, we conduct extensive experiments on a publicly available dataset. The experimental results show that the proposed method can improve the application hit ratio by appropriately 2.6% to 9% compared with other methods.

Container-enabled serverless computing has become a widely adopted approach for resource provisioning in the edge cloud. However, traffic incurred by container image pulling heavily burdens the already congested back-haul network. To relieve the problem, we do an analysis on Docker Hub, and find that instance deployment strategy has a significant impact on the back-haul traffic due to the varying similarity levels of different images. We incorporate this feature into task offloading decision and resource provisioning, and formulate the problem with a mixed integer non-linear programming (MINLP) problem. To address the challenges arising from the coupling and contradiction of instance deployment, image pulling, offloading decision, and resource allocation, we employ multi-agent deep reinforcement learning to decompose the problem into several simpler sub-problems, and design an algorithm for each sub-problem individually by exploiting convex optimization and fractional programming techniques. Simulations are conducted to validate the effectiveness of the proposed algorithm. The experiment results illustrate that our algorithm outperforms current notable solutions and improves the global utility by 13%–74%.

Differential privacy (DP) enables model training with a guaranteed bound on privacy leakage, therefore is widely adopted in federated learning (FL) to protect the model update. However, each DP-enhanced FL job accumulates privacy leakage, which necessitates a unified platform to enforce a global privacy budget for each dataset owned by users. In this work, we present a novel DP-enhanced FL platform that treats privacy as a resource and schedules multiple FL jobs across sensitive data. It first introduces a novel notion of device-time blocks for distributed data streams. Such data abstraction enables fine-grained privacy consumption composition across multiple FL jobs. Regarding the non-replenishable nature of the privacy resource (that differs it from traditional hardware resources like CPU and memory), it further employs an allocation-then-recycle scheduling algorithm. Its key idea is to first allocate an estimated upper-bound privacy budget for each arrived FL job, and then progressively recycle the unused budget as training goes on to serve further FL jobs. Extensive experiments show that our platform is able to deliver up to 2.1× as many completed jobs while reducing the violation rate by up to 55.2% under limited privacy budget constraint.

Federated learning (FL) has emerged as a prominent paradigm enabling collaborative model training without transmitting local data, thereby safeguarding data privacy. However, the practical implementation of FL systems on these devices faces a significant challenge: the heterogeneous corruption of data on individual clients, leading to unanticipated accuracy degradation during real-world deployment. In this work, we first introduce a realistic data corruption simulation framework to test the robustness of FL systems. In this framework, an in-depth analysis of potential data corruption patterns occurring on devices is conducted, followed by the construction of individual datasets with varying corruption types and degrees. Such data corruption results in the robustness degradation of conventional FL protocol (FedAVG) significantly higher than centralized learning (CL). Atop this key observation, we propose an adaptive FL protocol that emulates the CL training process. The protocol leverages imbalanced client data sampling to mitigate the negative impact of data corruption. Furthermore, a hybrid aggregation strategy is designed to accelerate model convergence and reduce additional communication overhead. Extensive experiments validate the effectiveness of our approach in enhancing the robustness of FL systems against client data corruption, which achieves up to 12% higher converge accuracy than FedAVG-based systems with acceptable overhead.

Software system complexity and security vulnerability diversity are plausible sources of the persistent challenges in software vulnerability research. Applying deep learning methods for automatic vulnerability detection has been proven an effective means to complement traditional detection approaches. Unfortunately, lacking well-qualified benchmark datasets could critically restrict the effectiveness of deep learning-based vulnerability detection techniques. Specifically, the long-term existence of erroneous labels in the existing vulnerability datasets may lead to inaccurate, biased, and even flawed results.

Transformer-based pre-trained models have revolutionized NLP for superior performance and generality. Fine-tuning pre-trained models for downstream tasks often requires private data, for which federated learning is the de-facto approach (i.e., FedNLP). However, our measurements show that FedNLP is prohibitively slow due to the large model sizes and the resultant high network/computation cost. Towards practical FedNLP, we identify as the key building blocks adapters, small bottleneck modules inserted at a variety of model layers. A key challenge is to properly configure the depth and width of adapters, to which the training speed and efficiency is highly sensitive. No silver-bullet configuration exists: the optimal choice varies across downstream NLP tasks, desired model accuracy, and mobile resources. To automate adapter configuration, we propose AdaFL1, a framework that enhances the existing FedNLP with two novel designs. First, AdaFL progressively upgrades the adapter configuration throughout a training session; the principle is to quickly learn shallow knowledge by only training fewer and smaller adapters at the model's top layers, and incrementally learn deep knowledge by incorporating deeper and larger adapters. Second, AdaFL continuously profiles future adapter configurations by allocating participant devices to trial groups. Extensive experiments show that AdaFL can reduce FedNLP's model convergence delay to no more than several hours, which is up to 155.5× faster compared to vanilla FedNLP and 48× faster compared to strong baselines.

Natural language processing (NLP) sees rich mobile applications. To support various language understanding tasks, a foundation NLP model is often fine-tuned in a federated, privacy-preserving setting (FL). This process currently relies on at least hundreds of thousands of labeled training samples from mobile clients; yet mobile users often lack willingness or knowledge to label their data. Such an inadequacy of data labels is known as a few-shot scenario; it becomes the key blocker for mobile NLP applications.

DNNs are ubiquitous on edge devices nowadays. With its increasing importance and use cases, it's not likely to pack all DNNs into device memory and expect that each inference has been warmed up. Therefore, cold inference, the process to read, initialize, and execute a DNN model, is becoming commonplace and its performance is urgently demanded to be optimized. To this end, we present NNV12, the first on-device inference engine optimizing cold inference. NNV12 is built atop three novel optimization knobs: selecting a proper kernel (i.e., operator implementation) for each DNN operator, bypassing the weights transformation process by caching the post-transformed weights on disk, and pipelined execution of many kernels on asymmetric processors. To tackle with the huge search space, NNV12 employs a heuristic-based scheme to obtain a near-optimal kernel scheduling plan. We fully implement a prototype of NNV12 and evaluate its performance across extensive experiments. It shows that NNV12 achieves up to 15.2× speedup compared to the state-of-the-art DNN engines on edge CPUs and 401.5× speedup on edge GPUs, respectively.

The repeated user-item interaction now is becoming a common phenomenon in the e-commerce scenario. Due to its potential economic profit, various models are emerging to predict which item will be re-interacted based on the user-item interactions. In this specific scenario, item relevance is a critical factor that needs to be concerned, which tends to have different effects on the succeeding re-interacted one (i.e., stimulating or delaying its emergence). It is necessary to make a detailed discernment of item relevance for a better repetition-aware recommendation. Unfortunately, existing works usually mixed all these types, which may disturb the learning process and result in poor performance.

With the rapid development of edge computing in the post-COVID19 pandemic period, precise workload forecasting is considered the basis for making full use of the edge limited resources, and both edge service providers (ESPs) and edge service consumers (ESCs) can benefit significantly from it. Existing paradigms of workload forecasting (i.e., edge-only or cloud-only) are improper, due to failing to consider the inter-site correlations and might suffer from significant data transmission delays. With the increasing adoption of edge platforms by web services, it is critical to balance both accuracy and efficiency in workload forecasting. In this paper, we propose ELASTIC, which is the first study that leverages a cloud-edge collaborative paradigm for edge workload forecasting with multi-view graphs. Specifically, at the global stage, we design a learnable aggregation layer on each edge site to reduce the time consumption while capturing the inter-site correlation. Additionally, at the local stage, we design a disaggregation layer combining both the intra-site correlation and inter-site correlation to improve the prediction accuracy. Extensive experiments on realistic edge workload datasets collected from China’s largest edge service provider show that ELASTIC outperforms state-of-the-art methods, decreases time consumption, and reduces communication cost.

The space industry experiences a rise in low-Earth-orbit satellite mega-constellations to achieve universal connectivity. At the same time, cloud firms (such as Google, Microsoft, and AWS) also have ambitions for computing in space to offer public cloud services in orbit. Satellite computing, as a new emerging concept, is promising to enable new paradigms in on-orbit autonomy, remote sensing, edge computing, and other areas by empowering satellites with computing resources. However, LEO mega-constellations bring inherent challenges for satellite computing in satellite networking, computing, and others due to the moving core infrastructure, reduced system power budget, and harsh space environment. This article presents vision and challenges for satellite computing based on a brief survey of the very recent literature in the “NewSpace” era and gives a case study of an open research platform on real satellites named Tiansuan constellation. This article aims to call researchers to collaboratively undertake the research of satellite computing and provide some insights for the research community.

Although microservice envisioned as a promising approach for edge applications which can improves the development efficiency and the deployment productivity, it also leads to the operational complexity during service runtime. This leads to the emergence of service mesh, a decided infrastructure layer over microservices for reliable service-to-service communication. However, when migrating the service mesh to the multi-tier edge computing, the corresponding extensions on the top of service mesh is needed to overcome the challenges brought by the multi-tier edge servers, shared microservices and diverse QoS requirements. Therefore, in this paper, we investigate the problem of service routing to achieve the efficient microservice-based service provision in multi-tier edge computing. The objective is to route user requests to the optimal microservice instances with low service delay and resource cost. To this end, we first formulate the service routing problem as an integer non-linear optimization problem, which is NP-hard. Then we map it to a many-to-one matching problem by employing the matching game theory and propose a dependency-aware deferred acceptance algorithm with dynamic quota. The experimental results based on a real-world dataset demonstrate that our proposed algorithm can significantly outperform existing representative algorithms in terms of service delay and resource cost.

Video streaming applications (VSAs) are increasingly being deployed on large-scale edge platforms, which have the potential to significantly improve the quality of service (QoS) and end-user experience (QoE), ultimately maximizing business outcomes. However, there is currently very little understanding of how QoS, QoE, and the impact of QoS on QoE for VSAs on edge platforms in the wild and at scale. To close the knowledge gap, we collect SNESet, an active measurement dataset comprising QoS and QoE telemetry metrics of 8 VSAs over four months, covering end-users from 798 edge sites,30 cities, and 3 ISPs in one country.We characterize and compare the QoS and QoE metrics in SNESet with existing publicly available datasets, highlighting that SNESet includes a significantly greater number of metrics (horizontal diversity and vertical hierarchy) and provides more comprehensive coverage of specific metrics.Moreover, we qualitatively and quantitatively analyze the impact of QoS on QoE in both domain-general and domain-specific scenarios. Our findings can inform the system design decisions that different entities in the video ecosystem (content providers, video player designers, third-party optimizers, edge vendors) make to maximize end-users experience and ultimately maximize the business outcomes. We hope SNESet can attract more research efforts in the data management community, computer network community, and beyond.

Extracting roads from multi-source data, such as aerial images and vehicle trajectories, is an important way to maintain road networks in the filed of urban computing. In this paper, we revisit the problem of road extraction and aim to boost its accuracy by solving three significant issues: the insufficient complementarity among multiple sources, rough edges of extracted roads, and many false positives caused by confusing pixels. In particular, we design an end-to-end neural network model to achieve this goal. At first, this model leverages two encoding networks to extract relative information from the inputs of two sources respectively, and then applies the attention mechanism to fuse them for sufficiently capturing the complementary correlation. Next, we introduce an auxiliary task, predicting road edges based on fused representations, to make the extracted roads smooth and continuous. At last, to reduce false positives relative to confusing pixels, we propose a pixel-aware contrastive-learning module to distinguish positive (roads) and negative (objects similar to roads) pixels. In addition, to improve the model's learning effectiveness, we propose a model-agnostic transfer learning method, which first builds auxiliary tasks to pre-train the whole model, and then fine-tunes the model's parameters for the main task. Extensive experiments on real datasets verify the superiority of our method as well as the importance of solving the three issues outlined above.

With the rapid development of the Internet of Things (IoT) and emerging Mobile Edge Computing (MEC) technologies, various IoT image recognition services are revolutionizing our lives by providing diverse cognitive assistance. However, most existing related approaches are difficult to meet the diversified needs of users because they believe that the MEC platform is a single layer. In addition, due to the mutual interference between the data, it is not easy for them to extract the discriminative features (DFs) necessary to analyze the input data. To this end, this article proposes an IoT image recognition services framework for different needs in the MEC environment, which consists of Hierarchical Discriminative Feature Extraction (HDFE) and Sub-extractor Deployment (Sub-ED) algorithms. We first propose HDFE, which can avoid mutual interference between data by separately optimizing the data structure, thereby generating an extractor that extracts effective DFs. Then there is Sub-ED, which divides the extractor into a series of sub-extractors and deploys them on appropriate MEC platforms. By doing so, the IoT device can connect to the corresponding MEC platform according to its service types, and use the sub-extract to extract DFs. Then, the MEC platform uploads the extracted feature data to the cloud server for further processing, e.g., feature matching. Finally, the cloud server sends the processed result back to the IoT device. Experimental results show that compared with the state-of-the-art approaches, the proposed framework improves recognition accuracy by about 6% and reduces network traffic by up to 94%.

With the rapid growth of the Industrial Internet of Things, a large amount of industrial data that needs to be processed promptly. Edge computing-based computation offloading can well assist industrial devices to process these data and reduce the overall time overhead. However, there are dependencies among tasks and some tasks have high latency requirements, so completing computation offloading while considering the above factors faces important challenges. In this paper, we design a computation offloading method based on a directed acyclic graph task model by modeling task dependencies. In addition to considering traditional optimization objectives in previous computation offloading problems (e.g., latency, energy consumption, etc.), we also propose an age of information (AoI) model to reflect the freshness of information and transform the task offloading problem into an optimization problem for latency, energy consumption, and AoI. To address this issue, we propose a method based on an improved dueling double deep Q-network computation offloading algorithm, named ID3CO. Specifically, it combines the advantages of deep Q-network, double deep Q-network, and dueling deep Q-network algorithms while further utilizing deep residual neural networks to improve convergence. Extensive simulations are conducted to demonstrate that ID3CO outperforms the existing baselines in terms of performance.

Task offloading offloads latency-sensitive and computation-intensive applications from resource-constrained terminal devices to relatively resource-rich edge servers to meet users’ demands for latency and energy consumption, which has attracted extensive attention from academia and industry. However, most of the existing researches only considers offloading dependent tasks within a single application or multiple independent applications, while ignoring the dependencies between applications. To this end, this paper proposes an offloading strategy for distributed dependent applications under the condition of limited computing and cache resources. The goal of the proposed strategy is to minimize the weighted sum of latency and energy to complete all applications while solving the offloading and resource allocation problems of dependent applications. However, the dual dependencies between applications and tasks within the application complicate offloading tasks. To accommodate this issue, we represent the dual dependencies as a directed acyclic graph. Then, we design the offloading strategy as follows: First, we transform the formulated non-convex problem into convex optimization subproblems. Second, we iteratively calculate the task priority and obtain the optimal offloading decision of the task according to the priority. Finally, we perform validation on real datasets. Compared with several state-of-the-art methods, our proposed strategy can significantly reduce the weighted sum of latency and energy.

Vehicles on roads have increasingly powerful computing capabilities and edge nodes are being widely deployed. They can work together to provide computing services for onboard driving systems, passengers, and pedestrians. Typical applications in vehicular systems have service requirements such as low latency and high reliability. Most studies in vehicular networks concerning latency and reliability focus on vehicular communication at the network level. Based on these fundamental works, an increasing proportion of vehicles boast complex applications that require service-level end-to-end performance guarantees. Several works guarantee service-level latency or reliability while new and innovative applications are demanding a joint optimization of the above two metrics. To address the critical challenges induced by the joint modeling of latency and reliability, system uncertainty, and performance and cost trade-off, we employ service request duplication to ensure both latency and reliability performance at the service level. We propose an online learning-based service request duplication algorithm based on a multi-armed bandit framework and Lyapunov optimization theory. The proposed algorithm achieves an upper-bounded regret compared to the oracle algorithm. Simulations are based on real-world datasets and the results demonstrate that the proposed algorithm outperforms the benchmarks.

In mobile edge computing networks, densely deployed access points are empowered with computation and storage capacities. This brings benefits of enlarged edge capacity, ultra-low latency, and reduced backhaul congestion. This paper concerns edge node grouping in mobile edge computing, where multiple edge nodes serve one end user cooperatively to enhance user experience. Most existing studies focus on centralized schemes that have to collect global information and thus induce high overhead. Although some recent studies propose efficient decentralized schemes, most of them did not consider the system uncertainty from both the wireless environment and other users. To tackle the aforementioned problems, we first formulate the edge node grouping problem as a game that is proved to be an exact potential game with a unique Nash equilibrium. Then, we propose a novel decentralized learning-based edge node grouping algorithm, which guides users to make decisions by learning from historical feedback. Furthermore, we investigate two extended scenarios by generalizing our computation model and communication model, respectively. We further prove that our algorithms converge to the Nash equilibrium with upper-bounded learning loss. Simulation results show that our mechanisms can achieve up to 96.99% of the oracle benchmark.

The cloud-assisted mobile edge computing system is a critical architecture to process computation-intensive and delay-sensitive mobile applications in close proximity to mobile users with high resource efficiency. Due to the heterogenous dynamics of task arrivals at edge nodes and the distributed nature of the system, the workloads of edge nodes are prone to be unbalanced, which can cause high task response time and resource cost. This paper solves the dynamic task scheduling problem in cloud-assisted mobile edge computing (including both peer task scheduling among edge nodes and cross-layer task scheduling from edge nodes to the cloud), aiming at minimizing average task response time within resource budget limit. To overcome the challenges of task arrival dynamics, edge node heterogeneity, and computation-communication delay tradeoff, we propose aWater-filling BasedDynamic TaskScheduling (WiDaS) algorithm. WiDaS dynamically tunes the usage of cloud resources based on the Lyapunov optimization method and efficiently schedules mobile tasks among edge nodes (and the cloud) by exploiting the idea of water filling. Extensive simulations are conducted to evaluate WiDaS under a trace-driven traffic pattern and two mathematic traffic patterns. The results demonstrate that WiDaS shows two-fold benefits of efficiency and effectiveness. In terms of efficiency, WiDaS can achieve the approximate results with the KKT-based algorithm while reducing the computation complexity from exponential order to polynomial order. In terms of effectiveness, WiDaS can reduce the average task response time by up to 64.4% and 47.2% over the Fair-ratio and the Edge-first algorithm.

Deploying high-performance vision transformer (ViT) models on ubiquitous Internet of Things (IoT) devices to provide high-quality vision services will revolutionize the way we live, work, and interact with the world. Due to the contradiction between the limited resources of IoT devices and resource-intensive ViT models, the use of cloud servers to assist ViT model training has become mainstream. However, due to the larger number of parameters and floating-point operations (FLOPs) of the existing ViT models, the model parameters transmitted by cloud servers are large and difficult to run on resource-constrained IoT devices. To this end, this article proposes a transmission-friendly ViT model, TFormer, for deployment on resource-constrained IoT devices with the assistance of a cloud server. The high performance and small number of model parameters and FLOPs of TFormer are attributed to the proposed hybrid layer and the proposed partially connected feed-forward network (PCS-FFN). The hybrid layer consists of nonlearnable modules and a pointwise convolution, which can obtain multitype and multiscale features with only a few parameters and FLOPs to improve the TFormer performance. The PCS-FFN adopts group convolution to reduce the number of parameters. The key idea of this article is to propose TFormer with few model parameters and FLOPs to facilitate applications running on resource-constrained IoT devices to benefit from the high performance of the ViT models. Experimental results on the ImageNet-1K, MS COCO, and ADE20K datasets for image classification, object detection, and semantic segmentation tasks demonstrate that the proposed model outperforms other state-of-the-art models. Specifically, TFormer-S achieves 5% higher accuracy on ImageNet-1K than ResNet18 with 1.4× fewer parameters and FLOPs.

5G and beyond network will support vertical industry applications, and the resource requirements of each service vary widely. The introduction of network slices provides great flexibility to the network, which can realize the differentiated customization requirements of service. However, while determining how to intelligently orchestrate the network slices is an important challenge, current solutions rarely treat multiple customized requirements of delay, bandwidth, load balancing, and slice isolation. In this article, network slice orchestration is considered from the perspective of slice isolation and cloud-edge collaboration. First, differentiated isolation level requirements are restricted to constraints, the customized isolation is realized. Second, bandwidth is saved and network latency is reduced via the collaboration of cloud and edge data centers. In addition, exclusive orchestration optimization objectives that match various service needs are proposed to distinguish the specific requirements of different slices. Finally, two deep reinforcement learning-based algorithms are proposed. The experimental results demonstrate that the proposed algorithms can optimize the objectives while ensuring differentiated isolation levels. For typical slices, the proposed algorithms respectively reduce bandwidth consumption by about 29% and 64%, reduce slice delay by about 14% and 70%, and optimize load balancing by about 17% and 23%.

Edge servers are key to the success of edge computing. Compared to cloud servers, edge servers suffer from a more constrained and costly electricity supply due to their dense, near-population deployment. Towards higher energy efficiency, we propose an extreme design of edge servers - SoC-Cluster that consists of massive, inter-connected ARM SoCs. Indeed, such SoC-Clusters have already been adopted to serve the cloud gaming application in the wild. In this paper, we present a concrete implementation of a COTS SoC-Cluster and its hardware specifications. We then discuss the potential killer applications that such SoC-Cluster can well serve and the major challenges to be solved. We also dive deep into two of such applications (live video transcoding and deep learning serving) and carry out a measurement study to demystify the application performance of SoC-Cluster. The results reveal that, compared to traditional servers, SoC-Cluster not only can reduce energy consumption but even deliver higher workload throughput in certain scenarios. Finally, we conclude the paper and discuss the primary research directions that can be explored by our community from applications, software, and hardware aspects.

This paper proposes Mandheling, the first system that enables highly resource-efficient on-device training by orchestrating mixed-precision training with on-chip Digital Signal Processor (DSP) offloading. Mandheling fully explores the advantages of DSP in integer-based numerical calculations using four novel techniques: (1) a CPU-DSP co-scheduling scheme to situationally mitigate the overhead from DSP-unfriendly operators; (2) a self-adaptive rescaling algorithm to reduce the overhead of dynamic rescaling in backward propagation; (3) a batch-splitting algorithm to improve DSP cache efficiency; (4) a DSP compute subgraph-reusing mechanism to eliminate the preparation overhead on DSP. We have fully implemented Mandheling and demonstrated its effectiveness through extensive experiments. The results show that, compared to the state-of-the-art DNN engines from TFLite and MNN, Mandheling reduces per-batch training time by 5.5X and energy consumption by 8.9X on average. In end-to-end training tasks, Mandheling reduces convergence time by up to 10.7X and energy consumption by 13.1X, with only 1.9%--2.7% accuracy loss compared to the FP32 precision setting.

Researchers and practitioners are exploiting Remote Direct Memory Access (RDMA) technology to improve the efficiency of distributed machine learning and meet the demands of data-center applications. RDMA requires lossless network link to fully unleash its power. RDMA Over Converged Ethernet (RoCE) v2 focuses on congestion control, but fails to achieve efficient packet loss recovery; Improved RoCE NIC (IRN) addresses this issue based on RoCEv2, but does not use the Priority-based Flow Control (PFC) to maintain the advantage of RoCEv2 in detecting congestion. This paper proposes a method of congestion detection and link control via feedback in RDMA transmission, namely Feedback Data Flow Control (FDFC), that does not rely on PFC. FDFC detects and controls the link condition in real time to achieve the goals of precise detection, congestion control, and efficient packet loss recovering.

[J1C2 Presentation Abstract at IEEE SERVICES 2022 for IEEE Transactions on Services Computing DOI 10.1109/TSC.2019.2955486]

The app markets enable users to submit feedback for downloaded apps in the form of star ratings and text reviews, which are meant to be helpful and trustworthy for decision making to both developers and other users. App markets have released strict guidelines/policies for user review submissions. However, there has been growing evidence showing the untrustworthy and poor-quality of app reviews, making the app store review environment a shambles. Therefore, review removal is a common practice, and market maintainers have to remove undesired reviews from the market periodically in a reactive manner. Although some reports and news outlets have mentioned removed reviews, our research community still lacks the comprehensive understanding of the landscape of this kind of reviews. To fill the void, in this paper, we present a large-scale and longitudinal study of removed reviews in iOS App Store. We first collaborate with our industry partner to collect over 30 million removed reviews for 33,665 popular apps over the course of a full year in 2020. This comprehensive dataset enables us to characterize the overall landscape of removed reviews. We next investigate the practical reasons leading to the removal of policy-violating reviews, and summarize several interesting reasons, including fake reviews, offensive reviews, etc. More importantly, most of these mis-behaviors can be reflected on reviews’ basic information including the posters, narrative content, and posting time. It motivates us to design an automated approach to flag the policy-violation reviews, and our experiment result on the labelled benchmark can achieve a good performance (F1=97%). We further make an attempt to apply our approach to the large-scale industry setting, and the result suggests the promising industry usage scenario of our approach. Our approach can act as a gatekeeper to pinpoint policy-violation reviews beforehand, which will be quite effective in improving the maintenance process of app reviews in the industrial setting.

Owing to the benefit of light weight, containers have become a promising enabler for cloud native computing. Container images composed of applications and dependencies support flexible service deployment and migration. Rapid adoption and integration of containers generate millions of images to be stored. Additionally, non-local images have to be frequently downloaded from the registry, resulting in huge amounts of traffic. Content Addressable Storage (CAS) has been adopted for saving storage and networking by enabling identical layers sharing across images. However, according to our measurements, the implication of CAS is significantly limited as layers are rarely fully identical in practice. In this paper, we propose to reconstruct the docker images to raise the number of identical layers and thereby reduce storage and network consumption. We explore the layered structure of images and define the commutativity of files to assure image validity. The image reconstruction is formulated as an integer nonlinear programming problem. Inspired by the observed similarity of layers, we design a similarity-aware online image reconstruction algorithm. Extensive evaluations are conducted to verify the performance of the proposed approach.

Deploying deep learning (DL) on mobile devices has been a notable trend in recent years. To support fast inference of on-device DL, DL libraries play a critical role as algorithms and hardware do. Unfortunately, no prior work ever dives deep into the ecosystem of modern DL libs and provides quantitative results on their performance. In this paper, we first build a comprehensive benchmark that includes 6 representative DL libs and 15 diversified DL models. We then perform extensive experiments on 10 mobile devices, which help reveal a complete landscape of the current mobile DL libs ecosystem. For example, we find that the best-performing DL lib is severely fragmented across different models and hardware, and the gap between those DL libs can be rather huge. In fact, the impacts of DL libs can overwhelm the optimizations from algorithms or hardware, e.g., model quantization and GPU/DSP-based heterogeneous computing. Finally, atop the observations, we summarize practical implications to different roles in the DL lib ecosystem.

Multiaccess edge computing (MEC) is considered as a backbone for the 5G network. The successive MEC network combines the networking and computation at the edge of the network to achieve the Quality of Services (QoS) with ultralow latency. The devices with mobility feature, whether hand-held devices or vehicles move from one edge server (ES) location to another ES, creates a nonoptimal environment in the long run. To maintain QoS and avoid service disruptions, existing network topologies do not fulfill the requirement. Hence, a unique traffic steering with dynamic path selection is required for live service migration of time-sensitive applications. In this article, we are the first to introduce a distributed traffic steering through the differentiation of two different types of network elements (i.e., ESs and routers), in a large MEC system. Using this concept, we, for the first time, resolve the scalability problem of a large MEC network into a partitioned MEC network. The proposed framework bounds the path-finding procedure with a filter strategy based on the network distance to eliminate the excess of nonrelated network elements. With a decentralized framework for MEC, we propose matrix-based dynamic shortest path selection and matrix-based dynamic multipath searching algorithms for dynamic path selection under the proposed autonomous network boundary discovery and must connect node block benchmarks. Our proposed dynamic traffic steering system works under two most important metric measurements (time delay and available bandwidth). Experimental results validate the effectiveness of dynamic and adaptive path searching in a partitioned controlled MEC network that significantly outperforms the centralized approaches with 35%–70% efficiency in QoS.

The explosion of Internet of Things (IoT) devices will generate massive amounts of data. Due to the limited resources of IoT devices, they usually upload the collected data to edge/cloud servers for processing. To reduce the amount of data uploaded to the edge/cloud server, extracting the features of data on IoT devices has attracted increasing attention. However, most existing related works suffer from two major limitations: 1) difficulty meeting user needs: the extractors they generate are difficult to extract effective features from data on IoT devices and 2) consume a lot of storage resources: they deploy multiple extractors to adapt to the dynamically changing resources of IoT devices. To this end, we propose a nonredundant discriminative feature extraction (DFE) framework, which consists of similarity-based DFE (SDFE) and 2RNestE algorithms. SDFE is first proposed to generate an extractor E that can extract effective discriminative features by rationally exploring the structural information of the data set on the edge/cloud server. Then, 2RNestE is proposed, which takes E as input, and outputs a nonredundant multifunctional extractor by removing redundant subextractors and nesting the remaining nonredundant subextractors together. Finally, the edge/cloud server sends the generated extractor to the IoT device. Experimental results show that the proposed framework reduces memory footprint by about 82.6% and switching overhead by about 84.6% compared with state-of-the-art works.

Mobile-edge computing (MEC) is considered as a promising solution to release pressure on the core network and reduce service response time. Edge nodes with storage and computation resources are able to cache various services and process tasks rather than offloading to remote clouds. However, it is difficult to make service deployment decisions appropriately since resources are limited in edge nodes and requirements of services are diverse. The hierarchical MEC structure in the 5G network causes extra complication, and the cost of service deployment aggravates the hardness, especially for service providers. In this article, we focus on the service deployment problem considering service caching, resource allocation, and task scheduling in the hierarchical MEC network, aiming at minimizing monetary cost. To address the heterogeneous limitations and requirements, we formulate the problem as a mixed-integer nonlinear programming problem and develop an iterative service deployment algorithm by exploiting Gibbs sampling. We make the service caching strategies of edge nodes iteratively. Furthermore, we transform the resource allocation and task scheduling optimization into the linear programming problem and employ a typical optimization function. Simulation results show that our algorithm always obtains minimal monetary cost for various number of services and task arrival rates, compared with benchmarks.

Recently, increasing investments in satellite-related technologies make the low earth orbit (LEO) satellite constellation a strong complement to terrestrial networks. To mitigate the limitations of the traditional satellite constellation “bent-pipe” architecture, satellite edge computing (SEC) has been proposed by placing computing resources at the LEO satellite constellation. Most existing works focus on space-air-ground integrated network architecture and SEC computing framework. Beyond these works, we are the first to investigate how to efficiently deploy services on the SEC nodes to realize robustness aware service coverage with constrained resources. Facing the challenges of spatial-temporal system dynamics and service coverage-robustness conflict, we propose a novel online service placement algorithm with a theoretical performance guarantee by leveraging Lyapunov optimization and Gibbs sampling. Extensive simulation results show that our algorithm can improve the service coverage by$4.3\times $compared with the baseline.

In a 5G network, mobile-edge computing (MEC) plays a key role in providing low access delay services. The placement of edge servers not only determines the quality of services on the user side but also affects the profit of running a MEC system. In this article, we study how to properly place edge servers so as to guarantee the access delay and maximize the profit of edge providers. We first propose a profit model which involves both access delay and energy consumption. In this model, we take the 5G user plane function (UPF) into consideration to calculate access delay for the first time. Then, we devise a particle swarm optimization-based algorithm to optimize the profit. In the algorithm, we introduce a weight value$q$to guarantee the access delay and assign base stations properly. Moreover, a service-level agreement is adopted to balance the tradeoff between access delay and energy consumption. We take advantage of our 5G network emulator called mini5Gedge and data set from Shanghai Telecom to conduct massive experiments. The results show that our algorithm stands out in terms of achieving the highest profit.

Edge servers can cache some delay-sensitive and resource-intensive applications to reduce the delay of user tasks. However, due to the limited resources of edge servers, they cannot cache all services and process all user requests like remote clouds. In order to take advantage of the low latency, we need to make full use of the limited resources of the edge servers and reasonably allocate user requests to edge servers for processing. Meanwhile, in order to maintain the efficient and long-term operation of the server cluster, we should also consider the load balancing of the cluster. How to balance the server load while making users have the best experience is an urgent problem to be solved. Solving the problem faces the challenge of the interaction between service placement and request scheduling, the tradeoff between communication and computation, and the consideration of response time and edge server load. We propose the service placement algorithm based on user visits and our request scheduling algorithm based on the simulated annealing algorithm. We verify the superiority of our algorithms in response time and server load balancing. The experimental results based on the real data sets prove our algorithm can cope with the actual situation and quickly converge to a favorable value.

With the rapid upgrading and explosive growth of Internet of Things (IoT) devices in mobile-edge computing, more and more IoT applications with high resource requirements are developed and utilized. Meanwhile, there are large quantities of edge nodes (e.g., switches and edge servers) with limited resources, higher operating costs, and certain failure probabilities in the mobile-edge computing environment. Therefore, when an IoT application is split into multiple collaborative tasks and offloaded into multiple edge clouds, there is an urgent need to increase the availability level of the task allocation scheme and the resource utilization of edge servers under the condition of certain communication delay. In this article, we first present a joint optimization objective to evaluate the unavailability level, communication delay, and resource wastage while allocating the same batch of IoT applications to multiple edge clouds. We then propose an approach to minimize the joint optimization objective under the condition of certain communication delay. Finally, we performed a comprehensive simulation experiment analysis to demonstrate that our proposed approach is superior to other related approaches.

Internet of Things (IoT) devices have become an integral part of our lives and are increasingly used in almost every field. Subsequently, there are a large number of latency-sensitive IoT applications (e.g., face recognition and autonomous driving) targeted for mobile edge computing environments. These IoT applications are often split into multiple collaborative tasks and offloaded onto containers or virtual machines (VMs) with certain failure rates and recovery rates. If these containers or VMs are not deployed in the same edge servers, the bandwidth resources of edge clouds must be consumed to transfer data. These factors increase the completion time of IoT applications to different degrees, and then affect their reliability level. Therefore, there exists equilibrium between the reliability level and bandwidth consumption. In this article, we investigate the equilibrium of minimizing the bandwidth consumption of IoT applications while maximizing the reliability level of these IoT applications during task offloading. We propose a multiobjective optimization problem, and transform it to a single-objective optimization problem. Furthermore, we introduce two efficient approaches to acquire two near-optimal solutions. The results of simulation experiments demonstrate that our proposed approaches can observably enhance the reliability level and reduce the bandwidth consumption of IoT applications compared with other related approaches. Meanwhile, we also make a comparative analysis of our proposed approaches.

Vehicular edge computing (VEC) is an innovative computing paradigm with an exceptional ability to improve the vehicles’ capacity to manage computation-intensive applications with both low latency and energy consumption. Vehicles require to make task offloading decisions in dynamic network conditions to obtain maximum computation efficiency. In this article, we analyze computation efficiency in a VEC scenario, where a vehicle offloads its tasks to maximize computation efficiency as a tradeoff between computation time and energy consumption. Although, it is quite a challenge to ensure the quality of experience of the vehicle due to diverse task requirements and the dynamic wireless conditions caused by vehicle mobility. To tackle this problem, a computation efficiency problem is formulated by jointly optimizing task offloading decision and computation resource allocation. We propose a mobility-aware computational efficiency-based task offloading and resource allocation (MACTER) scheme and develop a distributed MACTER algorithm that provides the near-optimal solution. We further consider the fifth-generation new-radio vehicle-to-everything communication model, i.e., cellular link and millimeter wave, to enhance the system performance. The simulation outcomes demonstrate that the proposed algorithm can efficiently enhance computation efficiency while satisfying computing time and energy consumption constraints.

Viewport prediction and prefetch have an important influence on VR video streaming performance. This work proposes a novel federated learning-based viewport prediction model training algorithm, ComPer-FedAvg. The proposed algorithm leverages a VR video’s common viewing pattern and users’ personal viewing patterns to train the prediction model in a distributed and privacy-preserving manner. Further, considering the VR video viewport prediction accuracy, a stochastic game is formulated to solve the VR streaming network’s communication resource allocation problem, where limited communication resource blocks are auctioned to users to achieve the optimal overall VR viewing experience. For each user, the auction is decomposed into two disjoint subproblems, namely, the optimal number of data rate requesting and true value claiming (bidding). The optimal true value claiming has been analytically proved to be equal to the VR viewing reward with given data rate. Due to the lack of global information when users request data rate, we reformulate users’ data rate requesting problem as a POMDP problem. A novel deep reinforcement learning algorithm is adopted to solve the problem. Evaluation and simulation results show the proposed viewport prediction and VR streaming schemes outperform conventional solutions in terms of prediction accuracy and VR viewing experience.

Integrating dense small cell (DSC) networks with mobile edge computing is employed by 5G to tackle the contradiction between the computation limitations of user equipment (UE) and the stringent latency requirement of services. This paper investigates the service-oriented edge resource allocation problem in DSC networks, determining where to deploy the service entity, how many service entities should be deployed at each edge cloud, and how to assign the UEs to service entities. The problem is challenging for the following three aspects: 1) Service entity deployment and UE assignment are highly coupled. 2) Due to the overlap of coverage regions of densely deployed small cells, the allocation mechanism of different base stations has mutual effects on the overall service performance. 3) Considering the limited resources of edge clouds, it is a thorny problem to encourage edge clouds to cache and share service startup images. We devote the following efforts to tackle the problem under these challenges. First, we explore blockchain’s decentralized, traceable, and secure characteristics, and propose a scheme to encourage image sharing in mobile edge computing. Second, we formulate the service-oriented edge resource allocation as mixed integer non-linear programming. Third, towards the target of reducing the computational complexity, we decouple UE assignment from service entity deployment and solve it through Gibbs sampling. Moreover, the power of Lyapunov optimization and convex optimization is incorporated to reduce the long-term power consumption and budget. Experiment results demonstrate the superiority of our approach over current notable solutions.

Mobile applications can leverage high-quality deep learning models such as convolutional neural networks and deep neural networks to provide high-performance cognitive services. Prior work on deep learning models-based mobile applications in a cloud-edge computing environment focuses on performing lightweight data pre-processing tasks on edge servers for cloud-hosted cognitive servers. These approaches have two major limitations. First, it is uneasy for the mobile applications to assure satisfactory user experience in terms of network communication delay, because the intermediary edge servers are used only to pre-process data (e.g., images and videos) and the cloud servers are used to complete the tasks. Second, these approaches assume the pre-trained deep learning models deployed on cloud servers are static, and will not attempt to automatically upgrade in a context-aware manner. In this article, we propose a cloud-edge collaboration framework that facilitates delivering cognitive services with long-lasting, fast response, and high accuracy properties. We fist deploy a shallow model (i.e., EdgeCNN) on the edge server and a deep model (i.e., CloudCNN) on the cloud server. EdgeCNN can provide durable and rapid response cognitive services, because edge servers not only provide computing resources for mobile applications, but also close to users. Then, we enable CloudCNN to assist in training EdgeCNN to improve the performance of the latter. Thus, EdgeCNN also provides high-accuracy cognitive services. Furthermore, because users may continue to upload data to edge servers in real-world scenarios, we propose to use the ongoing assistance of CloudCNN to further improve the accuracy of the shallow model. Experimental results show that EdgeCNN can reduce the average response time of cognitive services by up to 55.08 percent and improve accuracy by up to 26.70 percent.

Next-generation intelligent transportation systems (ITS) are envisioned to greatly improve transportation safety and efficiency by incorporating wireless communications and informatics technologies into transportation systems. As the cornerstone for ITS, vehicular communication networks enable vehicles on the go to exchange information with other vehicles and the external environments, which expect to play a significant role in supporting a variety of services such as road safety, traffic management, and infotainment. However, the existing terrestrial networks including dedicated shortrange communications (DSRC)-based networks and cellular networks alone cannot serve the vehicular applications very well in different scenarios, due to the inherent issues of deployment, coverage, and capacity. It is imperative to exploit other communication infrastructures, such as low-earth orbit (LEO) satellites, unmanned aerial vehicles (UAVs), and high-altitude platforms, to support vehicular applications better, resulting in space-air-ground integrated networks (SAGIN). SAGIN can provide more comprehensive and three-dimensional network connectivity for moving vehicles, anywhere and anytime, by exploiting their respective advantages in terms of coverage, flexibility, reliability, and availability.

Mobile image recognition services, which provide people with image recognition services through the cameras of mobile devices, are revolutionizing our lives. However, most existing cloud/edge-based approaches suffer from two major limitations, (i) Low recognition accuracy and high network bandwidth pressure, and (ii) Not easy to extract features based on currently available resources of mobile devices. In this paper, we propose a resource-aware feature extraction framework for mobile image recognition services. The proposed framework consists of discriminative feature extraction (DFE) and NestDFE algorithms. The DFE algorithm can generate an extractor${{\mathbf E}}$to extract discriminative features from the image data set on the edge server and images on mobile devices. Thus, the proposed framework can achieve higher recognition accuracy and require mobile devices to upload less feature data to the edge server. The NestDFE algorithm generates a single multi-capacity extractor that acts as a series of sub-extractors and enables mobile devices to dynamically select sub-extractors. Experimental results show that the proposed framework improves recognition accuracy by about 23 percent and reduces network traffic by about 76 percent compared with existing approaches.

In mobile edge computing, popular mobile applications, such as augmented reality, usually offload their tasks to resource-rich edge servers. The user experience can be considerably affected when many mobile users compete for the limited communication and computation resources. The key technical challenge in task offloading is to guarantee the Quality of Service (QoS) for such applications. Existing work on task offloading focus on deterministic QoS (delay) guarantee, which means that tasks have to complete before the given deadline with 100 percent. However, it is impractical to impose a deterministic QoS guarantee for tasks due to the high dynamics of the wireless environment when offloading to edge servers. In this paper, we focus on task offloading with statistical QoS guarantee (tasks are allowed to complete before a given deadline with a probability above the given threshold), which can further save more energy by loosing the QoS requirement. Specially, we first propose a statistical computation model and a statistical transmission model to quantify the correlation between the statistical QoS guarantee and task offloading strategy. Then, we formulate the task offloading problem as a mixed integer non-Linear programming problem with the statistical delay constraint. We transform the statistical delay constraint into the constraints on CPU cycle numbers and the delay exponent respectively. We propose an algorithm to provide the statistical QoS guarantee for tasks using convex optimization theory and Gibbs sampling method. Experiment results show that the proposed algorithm outperforms the three baselines.

Nowadays, a growing number of computation-intensive applications appear in our daily life. Those applications make the loads of both the core network and the mobile devices, in terms of energy and bandwidth, hugely increase. Offloading computation-intensive tasks to edge cloud is proposed to address this issue. Since edge clouds have limited computation resources compared with the remote cloud, they would get over-loaded because of the heavy computation burden. Parked-vehicle-assisted mobile edge computing becomes one of the promising solutions for this problem. However, several critical issues in parked-vehicle-assisted mobile edge computing would result in low reliable edge service. The open environment would bring about uncertainty, and the data privacy is hard to ensure. In addition, different from edge cloud, each parked vehicle only has limited parking duration and can leave unexpectedly for personal reasons. Moreover, edge cloud and vehicle adopt different execution models of computation and communication. The heterogeneous environment may result in negative effect on cooperativeness. Ignoring those issues can result in substantial performance degradation. To tackle this challenge and explore the benefits of parked-vehicle-assisted offloading, we study the task offloading and resource-allocation problem by fully considering the above issues. First, we propose a resource-management scheme to address the privacy issue. Second, we review the execution model of computation and communication in parked-vehicle-assisted computation offloading. Then, we formulate the problem into a mixed-integer nonlinear programming. The problem is hard to tackle due to its non-convex nature, which means that the time complexity of finding global optimal solution is unaffordable. Finally, we decompose the original problem into two sub-problems with lower complexity, and related algorithms are given to deal with the sub-problems. Simulation results demonstrate the effectiveness of the proposed solution.

In the last decade there has been a strong move towards mobile computing and the proliferation of the IoT (Internet of Things). A huge number of devices have been connected to the Internet and created zettabytes of data items. To extract value from such massive data volumes, processing power offered by cloud computing is often utilized. However, streaming data to the cloud exposes some limitations related to increased communication and data transfer, which introduces delays and consumes network bandwidth. Another limitation that cloud-based computing for IoT poses is a limited or no network connectivity. Other problems with cloud-based processing of IoT generated data regard the sensitivity of the information, because sending and storing so much information in the cloud involves privacy and security challenges, related to the protection of personally identifiable information, storing it in compliance with privacy laws, securing stored information, and preventing from being stolen, or accessed and shared illegally. The use of AI in edge processing resulted in a new interdisciplinary field that enables distributed intelligence with edge devices and is known as edge AI or edge intelligence. However, research on edge AI is still relatively new, and thus models, techniques, and protocols supporting intelligent management, querying and mining of large-scale amounts of data produced at the edge are required. A lot of challenges related to providing edge intelligence include training edge devices, so they can become more and more smart. There is also a need for the presentation of the most recent outcome of research of distributed intelligence. The papers in this special issue address many of the challenges we have outlined, and are briefly summarized.

With the development of complex networks and with increasing service demands, service use is becoming more complex and the composition of services is becoming more complicated. In the XaaS (X as a Service) environment, users only care about the QoE of a service and do not care about the composition process of the service. Therefore, it is important to evaluate the reliability of the entire service. In this article, we use Petri Net as a basis for modeling the composition of services. In addition, we consider the problems of shared resources and common cause faults. Both of these problems can cause network congestion and regional failures. We use distance to assess the effects of regional faults and queuing theory to simulate the network congestion process. Moreover, in the simulation, we verify the impacts of regional failures and network congestion on service reliability. We choose the Tree-Based Search algorithm and the Semi-Markov Model as comparison algorithms. The results of our algorithm are related to service time. Our algorithm can timely reflect the impact of regional failure or network congestion, and it can feedback different evaluation results according to environmental changes. Therefore, our algorithm is more comprehensive and has better performance.

With the envision of sixth generation (6G) network technology, varied artificial intelligence (AI) services gradually develop from the network center to the edge, which makes unmanned aerial vehicle (UAV) a hot spot to provide auxiliary services of machine learning (ML) to empower terrestrial users intelligence. However, due to the sensitive privacy and limited resources, traditional centralized ML may not be used directly in such networks. As a promising distributed collaborative ML, fed-erated learning (FL) could be more suitable. Meanwhile, unlike conventional FL working on terrestrial networks, applying FL in UAV-assisted networks should strictly consider the impact of air-ground wireless channel caused by the maneuverability of UAVs, as well as the allocation of various network resources, including frequency and latency. To address these challenges, we propose to jointly optimize the UAV location and resource allocation, subject to the constraints of learning accuracy and training latency to minimize the energy consumption of terrestrial users. The for-mulated complicated non-convex problem is efficiently solved by an alternating optimization algorithm based on successive convex approximation (SCA) approaches after problem decomposition. Simulations results show that our proposed algorithm can reduce more overall users' energy consumption than three benchmarks while guaranteeing the learning accuracy within the maximum training latency.

Geosynchronous Earth Orbit (GEO) satellites, which can relay image data for Low Earth Orbit (LEO) satellites, play an important role in remote sensing. With the development of satellite technologies, the significantly improved computation capabilities of GEO satellites have enabled space service computing, through which GEO satellites can provide data processing services before forwarding to reduce the quantity of transmitted data. In the presence of multiple LEO satellites, how to make effective use of limited communication and computation resources in GEO satellites has become crucial. At present, the research on satellite resource management typically focuses on either communication or computation resources. Existing resource management algorithms are usually of slow convergence speed, which limits their applicability in real-time remote sensing scenarios. Therefore, we propose an aggregated resource management method for remote sensing applications. We first propose models for transmission tasks and processing tasks of remote sensing images. Then we formulate the aggregated resource management for satellite edge computing as a hybrid Stackelberg game and simplify the problem to speed up its convergence speed. Then we propose a distributed resource management algorithm to determine the optimal strategies. Simulation results show that the proposed method can quickly obtain the optimal resource allocation strategy and outperforms typical dynamic iterative algorithms in terms of service quantity and throughput.

Public edge platforms have drawn increasing attention from both academia and industry. In this study, we perform a first-of-its-kind measurement study on a leading public edge platform that has been densely deployed in China. Based on this measurement, we quantitatively answer two critical yet unexplored questions. First, from end users' perspective, what is the performance of commodity edge platforms compared to cloud, in terms of the end-to-end network delay, throughput, and the application QoE. Second, from the edge service provider's perspective, how are the edge workloads different from cloud, in terms of their VM subscription, monetary cost, and resource usage. Our study quantitatively reveals the status quo of today's public edge platforms, and provides crucial insights towards developing and operating future edge services.

With the popularization of the cloud, an increasing number of users and developers need to combine multiple different services to satisfy their requirements in the cloud. Consequently, the question of how to combine these services that are published by different service providers in the cloud has become an active focus of research in service-oriented cloud systems. Service composition is the core technology of service-oriented cloud systems.

With the increasing video demand in daily network traffic, it is an urgent task to develop effective algorithms to facilitate high-quality content delivery service. Recently, numerous adaptive streaming algorithms have been proposed to improve the user perceived experience. However, these algorithms were mainly developed from the perspective of single user. There is not yet systematical evaluation and comparison of the bitrate adaptation methods for multi-user video streaming. Besides, the Quality of Experience (QoE) metrics were not unified.In this work, we propose a new mininet-based testbed framework which is able to conduct real-time video streaming emulation in various multi-user scenarios. Seven state-of-the-art adaptation methods are incorporated into the testbed. Meanwhile, ITU-T P.1203 model, the world's first standard for measuring QoE of HTTP adaptive streaming, is implemented to calculate the mean opinion scores of different methods. Using the developed testbed, the performance of current adaptation methods in multi-user network is analyzed and compared. A variety of experiments are carried out by changing the user number and network conditions, in which the QoE of different users are investigated. It is found that current algorithms perform inconsistently in various network scenarios. In the excessive user and limited bandwidth cases, machine learning and scheduling techniques show superiority in providing high and equal QoE for all users. While in the high-delay case, the buffer-based approaches show robust performance. Overall, the findings of this work give an insight for designing and choosing adaptive streaming strategies in different multi-user network conditions.

Abstract In cloud computing, service composition provides an appealing paradigm to compose existing single services into a powerful workflow. The critical issues in service composition is that service providers may lie about the service's cost and QoS for more profits. However, existing studies only address the issue of services' untruthful cost, none of the studies consider both the issue of untruthful cost and untruthful QoS. Thus, it is necessary to design an incentive mechanism to motivate service providers to offer services with the truthful cost and truthful QoS. To address the above challenging issues, we propose the strategy‐proof auction mechanism for service composition (SPASC), which can determine the winners and dynamically calculate the price of services. To the best of our knowledge, this is the first study to design an incentive mechanism to motivate service providers to offer services with the truthful cost and truthful QoS in this field. Furthermore, we propose an auction algorithm to implement the SPASC. Both theoretical proof and the extensively experimental results validate that SPASC achieves the truthfulness and individual rationality. At the same time, the experimental results also prove that this mechanism can motivate service providers to offer services with the truthful cost and truthful QoS.

The emerging 6G network will make it possible for cybertwin, which relies deeply on the low latency and powerful computation provided by the edge network. To this end, the convergence of computing and network has been attached great importance. Most existing work study either placing edge servers or deploying user plane functions (UPFs), seldom considers the two processes jointly. In this article, we study how to minimize the latency with cost limitation by means of jointly deploying edge servers and UPFs in 6G scenario. We have shown that the problem is NP-hard. Then, we simplify the problem by analyzing the placement relationship between edge servers and UPFs and prune the solution space of the problem. To solve the problem effectively, a UPF and edge server placement algorithm is proposed. Massive experiments are conducted based on real-world data set and an edge core network emulator. The evaluation results show that our algorithm outperforms the benchmark algorithms.

Mobile-edge computing is a promising computing paradigm with the advantages of reduced delay and relieved outsourcing traffic to the core network. In mobile-edge computing, reducing the computation offloading cost of mobile users and maintaining fresh information at edge nodes are two critical while conflicted objectives, as both consume the limited wireless bandwidth of edge nodes. Although extensive efforts have been devoted to optimizing computation offloading decisions and some works have investigated freshness-aware channel allocation issues recently, no prior works have considered the above conflict. This article is the first work to jointly optimize the channel allocation and computation offloading decisions, aiming at reducing the computation offloading cost within freshness requirements of sensors. We analyze the recursiveness of Age of Information (AoI) in analogy to the evolvement of a queue and formulate the problem as a nonlinear integer dynamic optimization problem. To overcome the challenges of AoI-computation cost tradeoff, AoI time dependency and high complexity caused by the heterogeneity of users, we propose an algorithm to solve the problem with reduced computation complexity. Specifically, we first transform the original problem into a static optimization problem in each time slot (which is NP-hard) based on Lyapunov optimization techniques. To reduce the computation complexity, we exploit the finite improvement property of potential games and further enforce centralized control to reduce the number of improvement iterations. Simulations have been conducted and the results demonstrate that the proposed algorithm shows good effectiveness and scalability.

With the rapid development of Internet of Things (IoT), more and more applications focus on the detection of unmanned areas. With the assistance of unmanned aerial vehicle (UAV), IoT devices are able to access the network via aerial base stations. These UAV-assisted IoT applications still face security and energy challenges. The open environment of IoT applications makes the application easy to encounter external invasion. Limited energy of UAV results in the limited lifetime of network access. To address these challenges, researches on IoT security and energy efficiency are becoming hotspots. Nevertheless, in the UAV continuous coverage scenario, there is still an enormous potential to improve the security and efficiency of data collection in IoT applications. In this article, blockchain is introduced into the scene of UAV-assisted IoT, and a data collection system considering security and energy efficiency is proposed. In this system, UAV, as an edge data collection node, provides a long-term network access for IoT devices through regular cruises with recharging. By forwarding data and recording transactions, UAVs get charging coins as rewards. UAVs use charging coins to exchange charging time. UAV swarm builds distributed ledgers based on blockchain to resist the invasion of malicious UAV. In order to reduce energy consumption, this article designs an adaptive linear prediction algorithm. Through this algorithm, IoT devices upload prediction model instead of original data to greatly reduce in-network transmissions. Simulation results show that the proposed system can effectively improve the security and efficiency of data collection.

Mobile edge computing as a novel computing paradigm brings remote cloud resource to the edge servers nearby mobile users. Within one-hop communication range of mobile users, a number of edge servers equipped with enormous computation and storage resources are deployed. Mobile users can offload their partial or all computation tasks of a workflow application to the edge servers, thereby significantly reducing the completion time of the workflow application. However, due to the open nature of mobile edge computing environment, these tasks, offloaded to the edge servers, are susceptible to be intentionally overheard or tampered by malicious attackers. In addition, the edge computing environment is dynamical and time-variant, which results in the fact that the existing quasistatic workflow application scheduling scheme cannot be applied to the workflow scheduling problem in dynamical mobile edge computing with malicious attacks. To address these two problems, this paper formulates the workflow scheduling problem with risk probability constraint in the dynamic edge computing environment with malicious attacks to be a Markov Decision Process (MDP). To solve this problem, this paper designs a reinforcement learning-based security-aware workflow scheduling (SAWS) scheme. To demonstrate the effectiveness of our proposed SAWS scheme, this paper compares SAWS with MSAWS, AWM, Greedy, and HEFT baseline algorithms in terms of different performance parameters including risk probability, security service, and risk coefficient. The extensive experiments results show that, compared with the four baseline algorithms in workflows of different scales, the SAWS strategy can achieve better execution efficiency while satisfying the risk probability constraints.

Mobile edge computing is emerging as a new computing paradigm that provides enhanced experience to mobile users via low latency connections and augmented computation capacity. As the amount of user requests is time-varying, while the computation capacity of edge hosts is limited, Cloud Assisted Mobile Edge (CAME) computing framework is introduced to improve the scalability of the edge platform. By outsourcing mobile requests to clouds with various types of instances, the CAME framework can accommodate dynamic mobile requests with diverse quality of service requirements. In order to provide guaranteed services at minimal system cost, the edge resource provisioning and cloud outsourcing of the CAME framework should be carefully designed in a cost-efficient manner. Specifically, two fundamental issues should be answered: (1) what is the optimal edge computation capacity configuration? and (2) what types of cloud instances should be tenanted and what is the amount of each type? To solve these issues, we formulate the resource provisioning in CAME framework as an optimization problem. By exploiting the piecewise convex property of this problem, the Optimal Resource Provisioning (ORP) algorithms with different instances are proposed, so as to optimize the computation capacity of edge hosts and meanwhile dynamically adjust the cloud tenancy strategy. The proposed algorithms are proved to be with polynomial computational complexity. To evaluate the performance of the ORP algorithms, extensive simulations and experiments are conducted based on both the widely-used traffic models and the Google cluster usage tracelogs, respectively. It is shown that the proposed ORP algorithms outperform the local-first and cloud-first benchmark algorithms in system flexibility and cost-efficiency.

5G communication is making much progress in achieving the Internet of Things and improving the quality of user experience in large bandwidth scenarios. By introducing a variety of new technologies, the performance of 5G has been greatly improved. However, emerging applications put forward more stringent requirements in terms of latency, reliability, peak data rate, service continuity, etc. Communication technology still needs to be further developed. In this article, the next generation of core networks is conceptualized. Inspired by the nervous system of the octopus, we propose a new cognitive service architecture. Cognitive service architecture is a new architecture designed for the 6G core network. It is proposed to enhance the core network so that it is qualified for the increasingly high requirement for quality of service and complicated scenarios. We first give a short vision of the 6G core network. Then cognitive service architecture is demonstrated in detail. A case study is demonstrated to show how cognitive service architecture enhances the performance of the system. Enabling technologies for 6G cognitive service architecture are discussed at last.

Mobile Edge Computing (MEC) can facilitate various important image retrieval applications for mobile users by offloading partial computation tasks from resource-limited mobile devices to edge servers. However, existing related works suffer from two major limitations. (i) High network bandwidth cost: they need to extract numerous features from the image and upload these feature data to the cloud server. (ii) Lowretrieval accuracy: they separate the feature extraction processes from the image data set in the cloud server, thus unable to provide effective features for accurate image retrieval. In this paper, we propose a cloud-guided feature extraction approach for mobile image retrieval. In the proposed approach, the cloud server first leverages the relationships among labeled images in the data set to learn a projection matrix P. Then, it uses the matrix P to extract discriminative features from the image data set and form a low-dimensional feature data set. Following that, the cloud server sends the matrix P to the edge server and uses it to multiply the image χ. The result PTχ, i.e., image features, is uploaded to the cloud server to find the label of the image with the most similar multiplying result. The label is regarded as the retrieval result and returned to the mobile user. In the cloud-guided feature extraction approach, the matrix P can extract a small number of effective image features, which not only reduces network traffic but also improves retrieval accuracy. We have implemented a prototype system to validate the proposed approach and evaluate its performance by conducting extensive experiments using a real MEC environment and data set. The experimental results show that the proposed approach reduces the network traffic by nearly 93 percent and improves the retrieval accuracy by nearly 6.9 percent compared with the state-of-the-art image retrieval approaches in MEC.

As an emerging service architecture, microservice enables decomposition of a monolithic web service into a set of independent lightweight services which can be executed independently. With mobile edge computing, microservices can be further deployed in edge clouds dynamically, launched quickly, and migrated across edge clouds easily, providing better services for users in proximity. However, the user mobility can result in frequent switch of nearby edge clouds, which increases the service delay when users move away from their serving edge clouds. To address this issue, this article investigates microservice coordination among edge clouds to enable seamless and real-time responses to service requests from mobile users. The objective of this work is to devise the optimal microservice coordination scheme which can reduce the overall service delay with low costs. To this end, we first propose a dynamic programming-based offline microservice coordination algorithm, that can achieve the globally optimal performance. However, the offline algorithm heavily relies on the availability of the prior information such as computation request arrivals, time-varying channel conditions and edge cloud's computation capabilities required, which is hard to be obtained. Therefore, we reformulate the microservice coordination problem using Markov decision process framework and then propose a reinforcement learning-based online microservice coordination algorithm to learn the optimal strategy. Theoretical analysis proves that the offline algorithm can find the optimal solution while the online algorithm can achieve near-optimal performance. Furthermore, based on two real-world datasets, i.e., the Telecom's base station dataset and Taxi Track dataset from Shanghai, experiments are conducted. The experimental results demonstrate that the proposed online algorithm outperforms existing algorithms in terms of service delay and migration costs, and the achieved performance is close to the optimal performance obtained by the offline algorithm.

Network Function Virtualization (NFV) brings flexible provisioning and great convenience for enterprises outsource their network functions to the Data Center Networks (DCNs). Network service in NFV is deployed as a Service Function Chain (SFC), which includes an ordered set of Virtual Network Functions (VNFs). However, in one SFC, the SFC delay increases linearly as the length of SFC increases. SFC parallelism can achieve high performance of SFC. In this article, we focus on the parallelized SFC placement problem in DCN considering availability guarantee and resource optimization. Firstly, we define the parallelized SFC and propose a multi-flow backup model. The parallelized SFC consists of multiple parallelized sub-SFCs. We split large data flow into multiple small sub-flows, each of them can be transmitted in one sub-SFC. The backup model provides backup sub-SFCs for working sub-SFCs to improve availability. Finally, we design three placement strategies and a Hybrid Placement Algorithm (HPA) aimed at mapping SFCs to DCN. Evaluation results show that our proposed solutions outperform the related work. We can reduce SFC delay (30%) and optimize link consumption (reduce 40%) while guaranteeing availability (99.999%).

The past decades witnessed the fast and wide deployment of Internet. The Internet has bred the ubiquitous computing environment that is spanning the cloud, edge, mobile devices, and IoT. Software running over such a ubiquitous computing environment environment is eating the world. A recently emerging trend of Internet-based software systems is “ resource adaptive ,” i.e., software systems should be robust and intelligent enough to the changes of heterogeneous resources, both physical and logical, provided by their running environment. To keep pace of such a trend, we argue that some considerations should be taken into account for the future operating system design and implementation. From the structural perspective, rather than the “monolithic OS” that manages the aggregated resources on the single machine, the OS should be dynamically composed over the distributed resources and flexibly adapt to the resource and environment changes. Meanwhile, the OS should leverage advanced machine/deep learning techniques to derive configurations and policies and automatically learn to tune itself and schedule resources. This article envisions our recent thinking of the new OS abstraction, namely, ServiceOS , for future resource-adaptive intelligent software systems. The idea of ServiceOS is inspired by the delivery model of “ Software-as-a-Service ” that is supported by the Service-Oriented Architecture (SOA). The key principle of ServiceOS is based on resource disaggregation, resource provisioning as a service, and learning-based resource scheduling and allocation. The major goal of this article is not providing an immediately deployable OS. Instead, we aim to summarize the challenges and potentially promising opportunities and try to provide some practical implications for researchers and practitioners.

With the increasing popularity of cloud computing, many notable quality of service (QoS)-aware service composition approaches have been incorporated in service-oriented cloud computing systems. However, these approaches are implemented without considering the energy and network resource consumption of the composite services. The increases in energy and network resource consumption resulting from these compositions can incur a high cost in data centers. In this paper, the trade-off among QoS performance, energy consumption, and network resource consumption in a service composition process is first analyzed. Then, a green service composition approach is proposed. It gives priority to those composite services that are hosted on the same virtual machine, physical server, or edge switch with end-to-end QoS guarantee. It fulfills the green service composition optimization by minimizing the energy and network resource consumption on physical servers and switches in cloud data centers. Experimental results indicate that, with comparisons to other approaches, our approach saves 20-50 percent of energy consumption and 10-50 percent of network resource consumption.

Countries across the globe have been pushing strict regulations on the protection of personal or private data collected. The traditional centralized machine learning method, where data is collected from end-users or IoT devices, so that it can discover insights behind real-world data, may not be feasible for many data-driven industry applications in light of such regulations. A new machine learning method, coined by Google as Federated Learning (FL) enables multiple participants to train a machine learning model collectively without directly exchanging data. However, recent studies have shown that there is still a possibility to exploit the shared models to extract personal or confidential data. In this paper, we propose to adopt Multi-Party Computation (MPC) to achieve privacy-preserving model aggregation for FL. The MPC-enabled model aggregation in a peer-to-peer manner incurs high communication overhead with low scalability. To address this problem, the authors proposed to develop a two-phase mechanism by 1) electing a small committee and 2) providing MPC-enabled model aggregation service to a larger number of participants through the committee. The MPC-enabled FL framework has been integrated in an IoT platform for smart manufacturing. It enables a set of companies to train high quality models collectively by leveraging their complementary data-sets on their own premises, without compromising privacy, model accuracy vis-a`-vis traditional machine learning methods and execution efficiency in terms of communication cost and execution time.

Cognitive services have revolutionized the way we live, work and interact with the world. In recent years, deep neural networks have become the mainstream approach in cognitive service, and mobile edge computing facilitates a variety of cognitive services for users by offloading computation tasks from resource-limited mobile devices to relatively wealthy edge servers. Combining the two to provide users with a higher quality of cognitive service is an issue worth researching. However, many related studies are not easy to provide fast responses because in these systems, edge servers are only used to pre-process data, and the cloud server is used to perform tasks. In this paper, we aim to study deploying deep neural network models on edge servers to provide fast services. However, a single edge server collects only a small amount of data, which results in low inference accuracy. To address this problem, we propose a cloud and edge collaboration framework. The key idea of the proposed framework is to use a cloud model to assist in training an edge model to improve the latter's inference accuracy and enable the latter to provide fast response and high-performance cognitive service. Experimental results demonstrate the effectiveness of our proposed framework.

Mobile edge computing is beneficial for reducing service response time and core network traffic by pushing cloud functionalities to network edge. Equipped with storage and computation capacities, edge nodes can cache services of resource-intensive and delay-sensitive mobile applications and process the corresponding computation tasks without outsourcing to central clouds. However, the heterogeneity of edge resource capacities and mismatch of edge storage and computation capacities make it difficult to fully utilize both the storage and computation capacities in the absence of edge cooperation. To address this issue, we consider cooperation among edge nodes and investigate cooperative service caching and workload scheduling in mobile edge computing. This problem can be formulated as a mixed integer nonlinear programming problem, which has non-polynomial computation complexity. Addressing this problem faces challenges of sub-problem coupling, computation-communication tradeoff, and edge node heterogeneity. We develop an iterative algorithm named ICE to solve this problem. It is designed based on Gibbs sampling, which has provably near-optimal performance, and the idea of water filling, which has polynomial computation complexity. Simulation results demonstrate that our algorithm can jointly reduce the service response time and the outsourcing traffic, compared with the benchmark algorithms.

Edge computing as a highly distributed multi-tenant computing system relies on its effective interactions with the underlying 5G network to achieve its advantages in service provisioning. On the other hand, edge computing is also transforming 5G network via deploying more computing capabilities at network edge. In this paper, we advocate a holistic vision of service orchestration to facilitate integration of edge computing and 5G network. We attempt to sketch a big picture of the state of the art research enabling the integrated network-edge service architecture. We also discuss technical challenges and identify opportunities for future research in this area.

Vehicular edge computing (VEC) offers a new paradigm to improve vehicular services and augment the capabilities of vehicles. In this article, we study the problem of task scheduling in VEC, where multiple computation-intensive vehicular applications can be offloaded to roadside units (RSUs) and each application can be further divided into multiple tasks with task dependency. The tasks can be scheduled to different mobile-edge computing servers on RSUs for execution to minimize the average completion time of multiple applications. Considering the completion time constraint of each application and the processing dependency of multiple tasks belonging to the same application, we formulate the multiple tasks scheduling problem as an optimization problem that is NP-hard. To solve the optimization problem, we develop an efficient task scheduling algorithm. The basic idea is to prioritize multiple applications and prioritize multiple tasks so as to guarantee the completion time constraints of applications and the processing dependency requirements of tasks. The numerical results demonstrate that our proposed algorithm can significantly reduce the average completion time of multiple applications compared with benchmark algorithms.

Mobile-edge computing provisions computing and storage resources by deploying edge servers (ESs) at the edge of the network to support ultralow delay and high bandwidth services. To ensure QoS of latency-sensitive services in vehicular networks, service migration is required to migrate data of the ongoing services to the closest ES seamlessly when users move across different ESs. To achieve seamless service migration, path selection is proposed to obtain one or more paths (consisting of several switches and ESs) to transfer service data. We focus on the following problems about path selection: 1) where to implement path selection? 2) how to coordinate interests of mobile users (i.e., vehicles) and network providers since they have conflicting interests during path selection? and 3) how to ensure seamless service migration during the migration of vehicles? To address the above problems, this article investigates path selection for seamless service migration. We propose a path-selection algorithm to jointly optimize both interests of the network plane (i.e., the cost for network providers) and service plane (i.e., QoE of users). We first formulate it as a multiobjective optimization problem and further prove theoretically that the proposed algorithm can give aweakly Pareto-optimal solution. Moreover, to improve the scalability of the proposed algorithm, a distance-based filter strategy is designed to eliminate undesired switches in advance. We conduct experiments on two synthesized data sets and the results validate the effectiveness of the proposed algorithm.

Summary Mobile edge computing is emerging as a novel ubiquitous computing platform to overcome the limit resources of mobile devices and bandwidth bottleneck of the core network in mobile cloud computing. In mobile edge computing, it is a significant issue for cost reduction and QoS improvement to place edge clouds at the edge network as a small data center to serve users. In this paper, we study the edge cloud placement problem, which is to place the edge clouds at the candidate locations and allocate the mobile users to the edge clouds. Specifically, we formulate it as a multiobjective optimization problem with objective to balance the workload between edge clouds and minimize the service communication delay of mobile users. To this end, we propose an approximate approach that adopted the K‐means and mixed‐integer quadratic programming. Furthermore, we conduct experiments based on Shanghai Telecom's base station data set and compare our approach with other representative approaches. The results show that our approach performs better to some extent in terms of workload balance and communication delay and validate the proposed approach.

Composing several API-defined services into one composite service per user requirements has become an important service creation approach in the cloud-enabled API economy. Various service selection approaches in support of service composition on demand have been proposed. They usually assume that networking resources are over-provisioned and their usage needs not be considered when making quality-aware service composition decisions. In practice, these approaches often lead to wasteful network resource consumption and impractical end-to-end QoS optimality for cloud-based services. This paper proposes a network-aware cloud service composition approach, named NetMIP, with comparative experimental evaluations for the clouds that adopt the widely deployed fat-tree network topology. By formalizing the service composition goal as a multi-objective constraint optimization problem, we have validated the proposed approach can be used to effectively reduce network resource consumption and deliver QoS optimality while satisfying the end-to-end QoS constraints for the candidate composite services in the cloud. The comparative experimental evaluations are done via a credible cloud infrastructure simulation system, named WebCloudSim. Extensive evaluation results show that NetMIP outperforms several representative cloud service composition approaches in terms of network resource consumption, QoS optimality, and computation time under various service selection workloads and fat-tree network topology settings.

With greater numbers of cloud applications being deployed in virtual clusters (VCs) and running in datacenters, the energy consumption of datacenters is experiencing rapidly cumulative growth. Thus, it is a critical issue that how to allocate virtual machines (VMs) of the VC in physical machines (PMs) for energy savings as much as possible. Considering each PM/switch has a certain failure rate, VMs may not be executed when they meet with any PM/switch fault. So, a compact allocation scheme may benefit in low energy consumption but increase the risk of violating the availability of the VC. In this paper, we consider four optimization objectives about the VC and the datacenter, i.e., availability, energy consumption, average resource utilization, and resource load balance. Then we propose a multi-objective optimization model and raise an evolution algorithm to trade-off among these four optimization objectives. Finally, experimental results show the effectiveness and efficiency of our algorithm.

As more and more services are deployed in the cloud datacenter, network traffic is growing exponentially. Virtual machines (VMs) of a virtual cluster (VC) must be allocated on physical machines (PMs) in the datacenter with a certain topology. Each VM needs some resources to run various services. Apparently, allocating VMs in the datacenter as compactly as possible can reduce traffic consumption and avoid bandwidth-related bottlenecks. However, loose allocation scheme can reduce the loss expectation of VMs due to failure possibilities of PMs and switches, and the availability of the VC is increased thereby. To enhance availability and reduce network bandwidth usage, it is significant to determine the scheme of allocating VMs of the VC. In this paper, we first introduce four typical datacenter architectures with network topologies and corresponding cost matrices, and extend to generality. Then we propose a joint optimization function to measure the risk of VC and the core bandwidth usage with a global availability constraint. Subsequently, an evolutionary algorithm is raised to minimize the value of the constrained optimization function. Finally, the evaluation results show the effectiveness of the proposed approach and performance advancement over the existing approaches.

As greater numbers of data-intensive applications are required to process big data in bandwidth-constrained datacenters with heterogeneous physical machines (PMs) and virtual machines (VMs), network core traffic is experiencing rapid growth. The VMs of a virtual cluster (VC) must be allocated as compactly as possible to avoid bandwidth-related bottlenecks. Since each PM/switch has a certain failure probability, a VC may not be executed when it meets with any PM/switch fault. Although the VMs of a VC can be spread out across different fault domains to minimize the risk of violating the availability requirement of the VC, this increases the network core traffic. Therefore, avoiding the decrease in availability caused by the heterogeneous PM/switch failure probabilities and bandwidth-related bottlenecks has been a constant challenge. In this paper, we first introduce a joint optimization function to measure the overall risk cost and overall bandwidth usage in the network core to allocate the same set of data-intensive applications. We then introduce an approach to maximize the value of the joint optimization function. Finally, we performed a side-by-side comparison with prior algorithms, and the experimental results show that our approach outperforms the other existing algorithms.

There is a growing trend for service providers to migrate their services from local clusters to the cloud data center. When there is no single service can satisfy the functionality requirement of the end user, existing services are combined together to fulfill the requirements. The data communication between component service hosting servers imposes a heavy burden on the data center network. In this article, we seek to reduce the data center network resource consumption by designing a novel service composition aware virtual machine migration approach. First, we formulate the problem as a multi-object integer non-linear(INLP) programming problem. The problem, which can be reduced into a well-known multi-object quadratic assignment problem, is proved to be NP-hard. Second, we simplify the multiple-objects INLP formulation into an equivalent, but much simplified single object ILP formulation. Then, we prove that the simplified formulation can also lead to the optimal solutions. Finally, optimization problem solvers, such as LPSolver, are employed to solve the problem. Experimental results in a large scale cloud data center demonstrate that our method significantly reduce the network resource consumption than other approaches.

Mobile edge computing (MEC) enables to provide relatively rich computing resources in close proximity to mobile users, which enables resource-limited mobile devices to offload workloads to nearby edge servers, and thereby greatly reducing the processing delay of various mobile applications and the energy consumption of mobile devices. Despite its advantages, when a large number of mobile users simultaneously offloads their computation tasks to an edge server, due to the limited computation and communication resources of edge server, inefficiency resource allocation will not make full use of the limited resource and cause waste of resource, resulting in low system performance (the weighted sum of the number of processed tasks, the number of punished tasks, and the number of dropped tasks). Therefore, it is a challenging problem to effectively allocate the computing and communication resources to multiple mobile users. To cope with this problem, we propose a performance-aware resource allocation (PARA) scheme, the goal of which is to maximize the long-term system performance. More specifically, we first build the multiuser resource allocation architecture for computing workloads and transmitting result data to mobile devices. Then, we formulate the multiuser resource allocation problem as a Markova Decision Process (MDP). To achieve this problem, a performance-aware resource allocation (PARA) scheme based on a deep deterministic policy gradient (DDPG) is adopted to derive optimal resource allocation policy. Finally, extensive simulation experiments demonstrate the effectiveness of the PARA scheme.

The recent developments in networking research leverage the principles of virtualization and service-orientation to enable fundamental changes in network architecture, which forms a trend of network cloudification that enables network systems to be realized using cloud technologies and network services to be provisioned following the cloud service model. On the other hand, the latest progress in cloud and fog computing has made networking an indispensable ingredient for cloud/fog service delivery. Convergence of networking and cloud/fog computing technologies enables unification of network and cloud/fog service provisioning, which has become an active research area that attracts interest from both academia and industry. In this paper, we introduce the notion of network-cloud/fog service unification and propose an architectural framework for unified network-cloud/fog service provisioning. Then we present a survey that reflects the state of the art of research on enabling network-cloud/fog service unification. We also discuss challenges to realizing such unification and identify some opportunities for future research, with a hope to arouse the research community's interest in this exciting interdisciplinary field.

Crowd intelligence tries to gather, process, infer, and ascertain massive useful information by utilizing the intelligence of crowds or distributed computers, which has great potential in Industrial Internet of Things. A crowd-intelligence ecosystem involves three stakeholders, namely the platform, workers (e.g., individuals, sensors, or processors), and task publisher. The stakeholders have no mutual trust but interest conflict, which means bad cooperation of them. Due to lack of trust, transferring raw data (e.g., pictures or video clips) between publisher and workers requires the remote platform center to serve as a relay node, which implies network congestion. First, we use a reward-penalty model to align the incentives of stakeholders. Then the predefined rules are implemented using blockchain smart contract on many edge servers (ES) of the mobile edge computing network, which together function as a trustless hybrid human-machine crowd-intelligence platform. As ES are near to workers and publisher, network congestion can be effectively improved. Further, we proved the existence of the only one strong Nash equilibrium, which can maximize the interests of involved ES and make the ecosystem bigger. Theoretical analysis and experiments validate the proposed method, respectively.

Crowdsourcing is a promising platform, whereby massive tasks are broadcasted to a crowd of semi-skilled workers by the requester for reliable solutions. In this paper, we consider four key evaluation indices of a crowdsourcing community (i.e., quality, cost, latency, and platform improvement), and demonstrate that these indices involve the interests of the three stakeholders, namely the requester, worker, and crowdsourcing platform. Since the incentives among these three stakeholders always conflict with each other, to elevate the long-term development of the crowdsourcing community, we take the perspective of the whole crowdsourcing community, and design a crowdsourcing mechanism to align incentives of stakeholders together. Specifically, we give workers reward or penalty according to their reporting solutions instead of only nonnegative payment. Furthermore, we find a series of proper reward-penalty function pairs and compute workers personal order values, which can provide different amounts of reward and penalty according to both the workers reporting beliefs and their individual history performances, and keep the incentive of workers at the same time. The proposed mechanism can help latency control, promote quality and platform evolution of crowdsourcing community, and improve the aforementioned four key evaluation indices. Theoretical analysis and experimental results are provided to validate and evaluate the proposed mechanism, respectively.

This paper studies the practically important problem of range query for sensor-augmented RFID systems, which is to classify the target tags according to the ranges specified by the user. The existing RFID protocols that seem to address this problem suffer from either low time-efficiency or the information corruption issue. To overcome their limitations, we first propose a basic classification protocol called Range Query (RQ), in which each tag pseudo-randomly chooses a slot from the time frame and uses the ON-OFF Keying modulation to reply its range identifier. Then, RQ employs a collaborative decoding method to extract the tag range information from singleton and even collision slots. The numerical results reveal that the number of queried ranges significantly affects the performance of RQ. To optimize the number of queried ranges, we further propose the Partition&Mergence (PM) approach that consists of two steps, i.e., top-down partitioning and bottom-up merging. Sufficient theoretical analyses are proposed to optimize the involved parameters, thereby minimizing the time cost of RQ+PM or minimizing its energy cost. We can trade off between time cost and energy cost by adjusting the related parameters. The prominent advantages of the RQ+PM protocol over previous protocols are two-fold: (i) it is able to make use of the collision slots, which are treated as useless in previous protocols. Thus, frame utilization can be significantly improved; (ii) it is immune to the interference from unexpected tags, and does not suffer information corruption issue. We use USRP and WISP tags to conduct a set of experiments, which demonstrate the feasibility of RQ+PM. Extensive simulation results reveal that RQ+PM can ensure 100% query accuracy, and reduce the time cost as much as 40% when comparing with the state-of-the-art protocols.

Advances in mobile Internet technology have enabled the clients of Web services to be able to keep their service sessions alive while they are on the move. Since the services consumed by a mobile client may be different over time due to client location changes, a multi-dimensional spatiotemporal model is necessary for analyzing the service consumption relations. Moreover, competitive Web service recommenders for the mobile clients must be able to predict unknown quality-of-service (QoS) values well by taking into account the target client's service requesting time and location, e.g., performing the prediction via a set of multi-dimensional QoS measures. Most contemporary QoS prediction methods exploit the QoS characteristics for one specific dimension, e.g., time or location, and do not exploit the structural relationships among the multi-dimensional QoS data. This paper proposes an integrated QoS prediction approach which unifies the modeling of multi-dimensional QoS data via multi-linear-algebra based concepts of tensor and enables efficient Web service recommendation for mobile clients via tensor decomposition and reconstruction optimization algorithms. In light of the unavailability of measured multi-dimensional QoS datasets in the public domain, this paper also presents a transformational approach to creating a credible multi-dimensional QoS dataset from a measured taxi usage dataset which contains high dimensional time and space information. Comparative experimental evaluation results show that the proposed QoS prediction approach can result in much better accuracy in recommending Web services than several other representative ones.

With the explosive growth of mobile applications, mobile devices need to be equipped with abundant resources to process massive and complex mobile applications. However, mobile devices are usually resource-constrained due to their physical size. Fortunately, mobile edge computing, which enables mobile devices to offload computation tasks to edge servers with abundant computing resources, can significantly meet the ever-increasing computation demands from mobile applications. Nevertheless, offloading tasks to the edge servers are liable to suffer from external security threats (e.g., snooping and alteration). Aiming at this problem, we propose a security and cost-aware computation offloading (SCACO) strategy for mobile users in mobile edge computing environment, the goal of which is to minimize the overall cost (including mobile device’s energy consumption, processing delay, and task loss probability) under the risk probability constraints. Specifically, we first formulate the computation offloading problem as a Markov decision process (MDP). Then, based on the popular deep reinforcement learning approach, deep Q-network (DQN), the optimal offloading policy for the proposed problem is derived. Finally, extensive experimental results demonstrate that SCACO can achieve the security and cost efficiency for the mobile user in the mobile edge computing environment.

A new networking paradigm, Vehicular Edge Computing (VEC), has been introduced in recent years to the vehicular network to augment its computing capacity. The ultimate challenge to fulfill the requirements of both communication and computation is increasingly prominent, with the advent of ever-growing modern vehicular applications. With the breakthrough of VEC, service providers directly host services in close proximity to smart vehicles for reducing latency and improving quality of service (QoS). This paper illustrates the VEC architecture, coupled with the concept of the smart vehicle, its services, communication, and applications. Moreover, we categorized all the technical issues in the VEC architecture and reviewed all the relevant and latest solutions. We also shed some light and pinpoint future research challenges. This article not only enables naive readers to get a better understanding of this latest research field but also gives new directions in the field of VEC to the other researchers.

Effective virtual machine placement algorithms can improve network resource utilization in cloud data centers. In order to design a more effective VM placement algorithm, we usually rely on a large-scale experimental platform to evaluate and verify its performance. However, large-scale experiment may have a huge impact on data center network, which makes it very unlikely to happen. Therefore, researchers need an experiment platform that can support large-scale environment to meet the above requirements. In order to evaluate the performance of virtual machine deployment algorithm, this paper first designs and implements a cloud data center network experiment system, WebCloudSim, which can support the joint experimental verification of real environment and simulation environment. Then, based on WebCloudSim, the paper implements three classic algorithms for virtual machine deployment, and verifies and analyzes the results. The experimental results show that WebCloudSim can effectively support the performance evaluation of virtual machine placement algorithm.

Throughput prediction is essential for ensuring high quality of service for video streaming transmissions. However, current methods are incapable of accurately predicting throughput in mobile networks, especially for moving user scenarios. Therefore, we propose a TCP throughput prediction method TRUST using machine learning for mobile networks. TRUST has two stages: user movement pattern identification and throughput prediction. In the prediction stage, the long short-term memory (LSTM) model is employed for TCP throughput prediction. TRUST takes all the communication quality factors, sensor data and scenario information into consideration. Field experiments are conducted to evaluate TRUST in various scenarios. The results indicate that TRUST can predict future throughput with higher accuracy than the conventional methods, which decreases the throughput prediction error by maximum 44% under the moving bus scenario.

Mobile Edge Computing (MEC) offers a new paradigm to improve vehicular services and augment the capabilities of vehicles. In this paper, to reduce the latency of the computation offloading of vehicles, we study multiple vehicles computation offloading problem in vehicular edge networks. We formulate the problem as a multi-user computation offloading game problem, prove the existence of Nash equilibrium (NE) of the game and propose a distributed computation offloading algorithm to compute the equilibrium. We analyze the price of anarchy of the game algorithm and evaluate the performance of the game algorithm using extensive simulations. Numerical results show that the proposed algorithm can greatly reduce the computation overhead of vehicles.

Mobile Edge Computing (MEC) is designed to extend the edge of the cloud network to decrease latency and network congestion, which would significantly improve the quality of experience (QoE) of adaptive streaming service for mobile users. This paper proposes QCSS, a QoE-aware control plane for adaptive streaming service over MEC infrastructures. QCSS aims to assure high QoE delivery of online streaming service to mobile users. The design of QCSS features: 1) a timeslot system with a look-ahead window for calculating cost of edge node switch and video quality adaption (to balance network load and reduce latency); 2) conducting service adaption via a set of cooperative action components running on client devices, edge nodes, and center nodes (to ensure a smooth viewing experience); 3) constructing a flexible QoE model and extending the scope and meaning of user-perceived experience. The effectiveness of QCSS has been validated via three real datasets. The validation results show that the proposed QCSS can improve QoE performance and network load performance for adaptive streaming service over MEC Infrastructures.

With the development of mobile Internet, mobile service is emerging one after another, and the problem of information overload is becoming ever more serious. As an important tool to alleviate information overload, mobile service recommendation has attracted more and more attention. However, traditional recommendation algorithms always recommend popular services to users, which result into a rich-get-richer problem and become a barrier for the unpopular services to startup and growth. In order to promote the healthy development of the service ecosystem, it is necessary to guarantee the fairness of unpopular services. To address this problem, this paper proposes a fairness-aware mobile service recommendation method (FMSR), which gives a relatively fair recommendation opportunity for unpopular services. FMSR makes a tradeoff between recommendation accuracy and fairness, and can recommend popular services and unpopular services respectively. For unpopular services, we design a fair efficiency function and use combinatorial optimization techniques to achieve recommendations. For popular services, bias matrix factorization is utilized to implement recommendations. Experimental results based on real-world demonstrate that FMSR significantly improve the fairness of mobile service recommendation in the evolving mobile service ecosystem.

Cloud data center traffic is experiencing a rapid growth as more and more data-intensive applications are required to process big data in a cloud data center. Although fat-tree networks own rich path multiplicity, and have been widely adopted as network topologies in cloud data center networks to transmit vast bisection bandwidth, they lead to bandwidth-related bottlenecks. To address this issue, in this paper, we propose a traffic-aware virtual cluster reallocation approach via biogeography-based optimization to allocate some reallocated virtual machines (VMs) as compact as possible with those VMs in the same virtual clusters. In order to validate our approach, we build a system model to perform a thorough evaluation of its performance. Experimental results show that our proposed approach outperforms six existed approaches in term of total transmission cost, total processing time, and total network resource consumption.

The large-scale utilization of cloud computing services for hosting industrial/enterprise applications has led to the emergence of cloud service reliability as an important issue for both cloud service providers and users. To enhance cloud service reliability, two types of fault tolerance schemes, reactive and proactive, have been proposed. Existing schemes rarely consider the problem of coordination among multiple virtual machines (VMs) that jointly complete a parallel application. Without VM coordination, the parallel application execution results will be incorrect. To overcome this problem, we first propose an initial virtual cluster allocation algorithm according to the VM characteristics to reduce the total network resource consumption and total energy consumption in the data center. Then, we model CPU temperature to anticipate a deteriorating physical machine (PM). We migrate VMs from a detected deteriorating PM to some optimal PMs. Finally, the selection of the optimal target PMs is modeled as an optimization problem that is solved using an improved particle swarm optimization algorithm. We evaluate our approach against five related approaches in terms of the overall transmission overhead, overall network resource consumption, and total execution time while executing a set of parallel applications. Experimental results demonstrate the efficiency and effectiveness of our approach.

A huge amount of data, generated by Internet of Things (IoT), is growing up exponentially based on nonstop operational states. Those IoT devices are generating an avalanche of information that is disruptive for predictable data processing and analytics functionality, which is perfectly handled by the cloud before explosion growth of IoT. Fog computing structure confronts those disruptions, with powerful complement functionality of cloud framework, based on deployment of micro clouds (fog nodes) at proximity edge of data sources. Particularly big IoT data analytics by fog computing structure is on emerging phase and requires extensive research to produce more proficient knowledge and smart decisions. This survey summarizes the fog challenges and opportunities in the context of big IoT data analytics on fog networking. In addition, it emphasizes that the key characteristics in some proposed research works make the fog computing a suitable platform for new proliferating IoT devices, services, and applications. Most significant fog applications (e.g., health care monitoring, smart cities, connected vehicles, and smart grid) will be discussed here to create a well‐organized green computing paradigm to support the next generation of IoT applications.

Mobile edge computing is a promising paradigm that brings computing resources to mobile users at the network edge, allowing computing-intensive and delay-sensitive applications to be quickly processed by edge servers to satisfy the requirements of mobile users. In this chapter, we first introduce a hierarchical architecture of mobile edge computing that consists of a cloud plane, an edge plane, and a user plane. We then introduce three typical computation offloading decisions. Finally, we review state-of-the-art works on computation offloading and present the use case of joint computation offloading.

Fog computing, as the supplement of cloud computing, can provide low‐latency services between mobile users and the cloud. However, fog devices may encounter security challenges as a result of the fog nodes being close to the end users and having limited computing ability. Traditional network attacks may destroy the system of fog nodes. Intrusion detection system (IDS) is a proactive security protection technology and can be used in the fog environment. Although IDS in tradition network has been well investigated, unfortunately directly using them in the fog environment may be inappropriate. Fog nodes produce massive amounts of data at all times, and, thus, enabling an IDS system over big data in the fog environment is of paramount importance. In this study, we propose an IDS system based on decision tree. Firstly, we propose a preprocessing algorithm to digitize the strings in the given dataset and then normalize the whole data, to ensure the quality of the input data so as to improve the efficiency of detection. Secondly, we use decision tree method for our IDS system, and then we compare this method with Naïve Bayesian method as well as KNN method. Both the 10% dataset and the full dataset are tested. Our proposed method not only completely detects four kinds of attacks but also enables the detection of twenty‐two kinds of attacks. The experimental results show that our IDS system is effective and precise. Above all, our IDS system can be used in fog computing environment over big data.

In order to build highly reliable composite service via Service Oriented Architecture (SOA) in the Mobile Fog Computing environment, various fault tolerance strategies have been widely studied and got notable achievements. In this paper, we provide a comprehensive overview of key fault tolerance strategies. Firstly, fault tolerance strategies are categorized into static and dynamic fault tolerance according to the phase of their adoption. Secondly, we review various static fault tolerance strategies. Then, dynamic fault tolerance implementation mechanisms are analyzed. Finally, main challenges confronted by fault tolerance for composite service are reviewed.

The proliferation of real-time applications, such as online gaming, video chatting, brings unprecedented pressure for multi-cloud application providers to secure the good quality of experience. These applications are designed for cooperative group users/members/clients and are expected to have low data transmission latency due to frequent interactions. We use total data transmission latency (TDTL) to represent the group latency in this paper. However, existing approaches perform poorly since they often ignore the feature of the cooperative group scenario and cannot provide a flexible rental cost model for application providers. Minimizing TDTL is challenging due to the difficulties in 1) Finding an effective method to transmit data through multi-cloud datacenters, 2) Taking the cooperative group scenario into consideration, 3) Providing a flexible transmission rental cost model. To this end, we propose CMIP, an innovative approach that aims to minimize the TDTL by renting virtual machines of well-connected datacenters as the proxy for the cooperative group scenario in multi-cloud providers. First, a weighted graph is constructed to illustrate the multi-cloud network. Second, we define a path latency model and a rental cost model for the cooperative group. After that, Yen's algorithm and mixed integer programming are adopted to optimize the TDTL with rental cost constraint. To study the performance of CMIP, we conduct extensive evaluations using multiple real data traces and compare it with related approaches. The comprehensive evaluation analysis shows that the CMIP achieves much lower total latency and is more flexible than other approaches.

It is challenging to conduct quality of experience (QoE) evaluations of web-based streaming video services effectively and efficiently. Aiming to overcome this challenge, we have created QoECenter, a web-based visual platform that innovatively facilitates comprehensive QoE evaluations of the streaming video services. QoECenter offers a holistic approach to conducting the QoE evaluations via an integrated set of technologies for source video classification, QoS realization of video encoding and network transmission, and context-aware user experience data gathering and analysis. From a QoECenter consumer's viewpoint, three kinds of data are required for an end-to-end streaming video QoE evaluation: video source level data, system process level data, and end user level data. QoECenter provides visual interfaces for parameter setting and data acquisition for each data level, and supports both objective and subjective datadriven QoE analyses. A QoECenter consumer can easily conduct comparative QoE evaluations like running easy-to-use visual applications. The effectiveness and efficiency design objectives of QoECenter have been validated by various real experiments.

Summary Cloud has become a very popular infrastructure for many smart city applications. A growing number of smart city applications from all over the world are deployed on the clouds. However, node failure events from the cloud data center have negative impact on the performance of smart city applications. Survivable virtual machine placement has been proposed by the researchers to enhance the service reliability. Because of the ignorance of switch failure, current survivable virtual machine placement approaches cannot achieve the best effect. In this paper, we study to enhance the service reliability by designing a novel network failure–aware redundant virtual machine placement approach in a cloud data center. Firstly, we formulate the network failure–aware redundant virtual machine placement problem as an integer nonlinear programming problem and prove that the problem is NP‐hard. Secondly, we propose a heuristic algorithm to solve the problem. Finally, extensive simulation results show the effectiveness of our algorithm.

Recently, the frequent coal mine safety accidents have caused serious casualties and huge economic losses. It is urgent for the global mining industry to increase operational efficiency and improve overall mining safety. This paper proposes a lightweight mashup middleware to achieve remote monitoring and control automation of underground physical sensor devices. First, the cluster tree based on ZigBee Wireless Sensor Network (WSN) is deployed in an underground coal mine, and propose an Open Service Gateway initiative (OSGi)-based uniform devices access framework. Then, propose a uniform message space and data distribution model, and also, a lightweight services mashup approach is implemented. With the help of visualization technology, the graphical user interface of different underground physical sensor devices could be created, which allows the sensors to combine with other resources easily. Besides, four types of coal mine safety monitoring and control automation scenarios are illustrated, and the performance has also been measured and analyzed. It has been proved that our lightweight mashup middleware can reduce the costs efficiently to create coal mine safety monitoring and control automation applications.

Most state-of-the-art service selection approaches assume user preferences can be provided by the target user with sufficient precision and ignore historical service usage data for all users. It is desirable for ordinary users to possess a new service selection approach that can recommend satisfactory services to them even when their service selection preferences are specified imprecisely in terms of vagueness, inaccuracy, and incompleteness. This paper proposes a novel service selection approach that resolves the imprecise characteristics of user preferences and can recommend satisfactory services for users with varying cognitive levels in terms of service experience. The proposed service selection approach is comprised of four major tasks: 1) employ user-friendly linguistic variables to collect apparent user preferences (AUP) and convert the linguistic variables to standardized fuzzy weights as AUP weights; 2) evaluate all users' respective cognitive levels for the target service type and obtain the cognitive level threshold for that type of services; 3) adjust the AUP weights based on the calculated cognitive levels and the threshold, and supplement the potential user preferences weights; and 4) prioritize candidate services per a user satisfaction maximization objective. In-depth comparative experimental evaluations were performed using two real-world datasets. The results show that our service selection model outperforms three other representative ones and could provide a stable and reliable selection of services for the users with low service cognitive levels.

Cloud computing refers to both the applications delivered as services over the Internet and hardware and system software in the cloud server farm that provides those services. The research on server backup of cloud server farm appears to be an important issue of cloud computing economics. Optimal logistic policy should be considered to be a profit-oriented framework simultaneously for providing qualified service to cloud users while the whole cloud center is under construction. The kernel point of the proposed approach is that a novel design pattern is developed for approaching optimal profit on logistics using the finite-source queuing theory. To model the proposed approach for qualitative analysis, a Petri Net model was developed to configure all relevant system aspects in a concise fashion. On quantitative work, a comprehensive mathematical analysis on profit pattern has been made in detail. Relevant simulations have also been conducted tovalidate the proposed optimization model. The design illustration is presented to demonstrate engineering application scenario in cloud environment, hence the proposed approach indeed provides a feasibly profit-oriented framework to meet logistic economy.

With rapid adoption of the cloud computing model, many enterprises have begun deploying cloud-based services. Failures of virtual machines (VMs) in clouds have caused serious quality assurance issues for those services. VM replication is a commonly used technique for enhancing the reliability of cloud services. However, when determining the VM redundancy strategy for a specific service, many state-of-the-art methods ignore the huge network resource consumption issue that could be experienced when the service is in failure recovery mode. This paper proposes a redundant VM placement optimization approach to enhancing the reliability of cloud services. The approach employs three algorithms. The first algorithm selects an appropriate set of VM-hosting servers from a potentially large set of candidate host servers based upon the network topology. The second algorithm determines an optimal strategy to place the primary and backup VMs on the selected host servers with k-fault-tolerance assurance. Lastly, a heuristic is used to address the task-to-VM reassignment optimization problem, which is formulated as finding a maximum weight matching in bipartite graphs. The evaluation results show that the proposed approach outperforms four other representative methods in network resource consumption in the service recovery stage.

In this paper, we propose a Skyline service selection approach based on QoS prediction. We first consider the QoS history records as time series and predict the QoS values by using Autoregressive Integrated Moving Average (ARIMA) model to provide more accurate QoS attributes values. And then we calculate the uncertainty of the prediction result by adopting an improved Coefficient of Variation. In order to downsize the search space, we employ Skyline computing to prune redundant services and then perform Skyline service selection by using Mixed Integer Programming. Extensive experimental results show that our approach has a better performance than other approaches.

The widespread utilization of cloud computing services has brought in the emergence of cloud service reliability as an important issue for both cloud providers and users. To enhance cloud service reliability and reduce the subsequent losses, the future status of virtual machines should be monitored in real time and predicted before they crash. However, most existing methods ignore the following two characteristics of actual cloud environment, and will result in bad performance of status prediction: 1. cloud environment is dynamically changing, 2. cloud environment consists of many heterogeneous physical and virtual machines. In this paper, we investigate the predictive power of collected data from cloud environment, and propose a simple yet general machine learning model StaP to predict multiple machine status. We introduce the motivation, the model development and optimization of the proposed StaP. The experimental results validated the effectiveness of the proposed StaP.

Given a service composition, it is challenging but important to have a runtime adaptation, due to the complicated execution environment and evolving feature of Web service. In this paper, we present a runtime adaptive service composition approach, taking execution time minimization and revenue maximization into consideration. Based on dynamic programming, we deduce the optimal policy. Through this policy, orchestrator selects one concrete service for per task on runtime. The experimental results show that the proposed approach outperforms previous approach.

Abstract Radio Frequency Identification (RFID) technology provides a seamless link between physical world and the information system in cyber space. However, the emerging Internet of Things and cloud systems are full of security holes, which introduce new challenging security problems in both tag identification and data privacy. In this paper, we present efficient methods for identity authentication and data encryption to enhance RFID privacy. The proposed techniques aim to ensure no adversary interacting with the tags and the reader, to infer any information on a tag's identity from the communication. The main idea, a symmetric cryptography technique, termed as Advanced Encryption Standard, was applied to implement both mutual authentication and data encryption between the front and back ends of RFID systems. The Advanced Encryption Standard cryptographic algorithm was adopted because of its low hardware complexity and high security strength. In addition, the proposed key management protocol is proved resistant to several known RFID attacks such as Man‐in‐the‐Middle, Denial‐of‐Service, replay, clone attack, and backward/forward traceability. The computational time complexity of the proposed scheme through the entire identity authentication and data encryption phases is 3Tencrypt + 1Tnonce + 4Txor, which is superior to most existing approaches. The proposed scheme was also proved to be able to provide higher data encryption performance than other symmetric cryptographic algorithms with regard to the same level of security strength. Copyright © 2016 John Wiley & Sons, Ltd.

An increasing number of companies are beginning to deploy services/applications in the cloud computing environment. Enhancing the reliability of cloud service has become a critical and challenging research problem. In the cloud computing environment, all resources are commercialized. Therefore, a reliability enhancement approach should not consume too much resource. However, existing approaches cannot achieve the optimal effect because of checkpoint image-sharing neglect, and checkpoint image inaccessibility caused by node crashing. To address this problem, we propose a cloud service reliability enhancement approach for minimizing network and storage resource usage in a cloud data center. In our proposed approach, the identical parts of all virtual machines that provide the same service are checkpointed once as the service checkpoint image, which can be shared by those virtual machines to reduce the storage resource consumption. Then, the remaining checkpoint images only save the modified page. To persistently store the checkpoint image, the checkpoint image storage problem is modeled as an optimization problem. Finally, we present an efficient heuristic algorithm to solve the problem. The algorithm exploits the data center network architecture characteristics and the node failure predicator to minimize network resource usage. To verify the effectiveness of the proposed approach, we extend the renowned cloud simulator Cloudsim and conduct experiments on it. Experimental results based on the extended Cloudsim show that the proposed approach not only guarantees cloud service reliability, but also consumes fewer network and storage resources than other approaches.

Reputation plays an important role for users in choosing or paying for multimedia applications or services. Some efficient multimedia reputation-measurement approaches have been proposed to achieve accurate reputation measurement based on feedback ratings that users give to a multimedia service after invoking. However, the implementation of these approaches suffers from the problems of wide abuse and low utilization of user context. In this article, we study the relationship between user context and feedback ratings according to which one user often gives different feedback ratings to the same multimedia service in different user contexts. We further propose an enhanced user context-aware reputation-measurement approach for multimedia services that is accurate in two senses: (1) Each multimedia service has three reputation values with three different user context levels when its feedback ratings are sufficient and (2) the reputation of a multimedia service with different user context levels is found using user context sensitivity and user similarity when its feedback ratings are limited or not available. Experimental results based on a real-world dataset show that our approach outperforms other approaches in terms of accuracy.

Quality of service (QoS) guarantee is an important component of service recommendation. Generally, some QoS values of a service are unknown to its users who has never invoked it before, and therefore the accurate prediction of unknown QoS values is significant for the successful deployment of web service-based applications. Collaborative filtering is an important method for predicting missing values, and has thus been widely adopted in the prediction of unknown QoS values. However, collaborative filtering originated from the processing of subjective data, such as movie scores. The QoS data of web services are usually objective, meaning that existing collaborative filtering-based approaches are not always applicable for unknown QoS values. Based on real world web service QoS data and a number of experiments, in this paper, we determine some important characteristics of objective QoS datasets that have never been found before. We propose a prediction algorithm to realize these characteristics, allowing the unknown QoS values to be predicted accurately. Experimental results show that the proposed algorithm predicts unknown web service QoS values more accurately than other existing approaches.

Many factors affect the time cost of Cloud computing tasks. One of the most serious factors is data transmission latency, which reduces the efficiency of Cloud computing. Existing notable schemes ignore the communication cost among virtual machines (VMs) in the MapReduce environment. In this paper, we propose a VM placement approach to reduce data transmission latency with the communication cost among VMs. We first construct bipartite graph and classify VMs as two groups according to their transmission latency with data nodes. Then we propose two VM placement optimization algorithms to minimize the total data transmission latency (TDTL) and the maximum data transmission latency (MDTL) in the MapReduce environment. Finally, we place VMs for Reduce phase. The evaluation results show that our approach reduces the average data transmission latency by 26.3% compared with other approaches.

Fast deployment of mobile Internet makes Web services often consumed under a multi-dimensional spatiotemporal model, wherein a specific service client could keep active while its location is changing. Recommending Web services for such clients must be able to predict unknown QoS values with the target client's service requesting time and location taken into account, e.g., Performing the prediction via a set of measured multi-dimensional QoS data. Most QoS prediction methods focus on the QoS characteristics for one specific dimension, e.g., Time or location, and do not exploit the structural relationships among the multi-dimensional QoS data. This paper proposes an integrated QoS prediction approach which unifies the modeling of multi-dimensional QoS data via multi-linear-algebra based concepts of tensor and enables efficient service recommendation for Web service based mobile clients via tensor decomposition and reconstruction optimization algorithms. Comparative experimental evaluation results show that the proposed QoS prediction approach could result in much better accuracy in recommending Web services than several other representative ones.

Currently, many researchers play more attention on cloud data center network and hope to verify their achievements. However, existing cloud data center simulation tools are difficult to provide an online cloud data center simulation. In this paper, we design an online cloud data center simulation system based on CloudSim and some extensions. This system support Browser/Server mode and some simple configuration can simplify the simulation process.

Exiting schemes rarely take into account coordinated problem among multiple virtual machines (VMs) which collectively complete a task, since if these VMs are uncoordinated, the execution results of a task are not correct. To solve the problem, we first exploit a proactive prediction scheme to predict the VM's status. Then, if VM's status is deteriorating, coordinated checkpoint is adopted to suspend the current task to search an optimal target host. Finally, we introduce an efficient heuristic algorithm to solve the optimal target host selection problem. The experimental results demonstrate the efficiency and effectiveness of our proposed approach.

Video may pass through various types of heterogeneous networks during the process of transmission, which has adverse impacts on the real-time video quality. Traditional methods focus on how to compress videos based on the video flow without considering the real-time network information. This paper presents an adaptive method that combines video encoding and the video transmission control system over heterogeneous networks. This method includes the following steps: first, to collect and standardize the real-time information describing the network and the video, then to assess the video quality and calculate the video coding rate based on the standardized information, and then to process the encoded compression of the video according to the calculated coding rate and transfer the compressed video. The experiments show that there is a significant improvement for the quality of real-time videos transmission without changing the existing network, particularly the core equipment. Our solution is easy to deploy and implement quickly and may help to extensively ensure video quality for normal users.

Web service recommendation systems can help service users to locate the right service from the large number of available web services. Avoiding recommending dishonest or unsatisfactory services is a fundamental research problem in the design of web service recommendation systems. Reputation of web services is a widely-employed metric that determines whether the service should be recommended to a user. The service reputation score is usually calculated using feedback ratings provided by users. Although the reputation measurement of web service has been studied in the recent literature, existing malicious and subjective user feedback ratings often lead to a bias that degrades the performance of the service recommendation system. In this paper, we propose a novel reputation measurement approach for web service recommendations. We first detect malicious feedback ratings by adopting the cumulative sum control chart, and then we reduce the effect of subjective user feedback preferences employing the Pearson Correlation Coefficient. Moreover, in order to defend malicious feedback ratings, we propose a malicious feedback rating prevention scheme employing Bloom filtering to enhance the recommendation performance. Extensive experiments are conducted by employing a real feedback rating data set with 1.5 million web service invocation records. The experimental results show that our proposed measurement approach can reduce the deviation of the reputation measurement and enhance the success ratio of the web service recommendation.

How to select the optimal composited service from a set of functionally equivalent services but different QoS attributes has become a hot research in service computing. However existing approaches are inefficient as they search all solution spaces. More importantly, they neglect the QoS inherently uncertainty due to the dynamic network environment. In this paper, we propose a fast and reliable Web service selection approach that attempts to select the best reliable composited service on the basis of filtering low reliable Web services according to the uncertainty of QoS. The approach first employs information theory and variance theory to abandon high QoS uncertainty services and downsize the solution spaces. A reliability fitness function is then designed to select the best reliable service for composited services. We experimented with real-world and synthetic datasets and compared our approach with other approaches. Our results show that our approach is not only fast, but also find more reliable composited services.

As cloud computing becomes widely deployed, more and more cloud services are offered to end users in a pay-as-you-go manner. Today’s increasing number of end user-oriented cloud services are generally operated by Software as a Service (SaaS) providers using rental virtual resources from third-party infrastructure vendors. As far as SaaS providers are concerned, how to process the dynamic user service requests more cost-effectively without any SLA violation is an intractable problem. To deal with this challenge, we first establish a cloud service request model with SLA constraints, and then present a cost-aware service request scheduling approach based on genetic algorithm. According to the personalized features of user requests and the current system load, our approach can not only lease and reuse virtual resources on demand to achieve optimal scheduling of dynamic cloud service requests in reasonable time, but can also minimize the rental cost of the overall infrastructure for maximizing SaaS providers ’ profits while meeting SLA constraints. The comparison of simulation experiments indicates that our proposed approach outperforms other revenue-aware algorithms in terms of virtual resource utilization, rate of return on investment and operation profit and provides a cost-effective solution for service request scheduling in cloud computing environments.

To maximize the profit of cloud service provider, we present a dynamic virtual resource renting method under SLA (service level agreement) constraints. According to the price distribution and current task emergency, the method attempts to adjust acceptable price of each virtual resource type at different price interval. If there is price acceptable resource, we choose to rent the most profitable one. Otherwise, if SLA permits, we even suspend the task and restart it when price falls. Partial experimental result in simulation environment is also presented.

A critical research issue is to lower the energy consumption of a virtualized data center by means of virtual machine placement optimization while satisfying the resource requirements of the cloud services. In this paper, we focus on different existing schemes and on the energy-aware virtual machine placement optimization problem of a heterogeneous virtualized data center. We attempt to explore a better alternative approach to minimizing the energy consumption, and we observe that particle swarm optimization (PSO) has considerable potential. However, the PSO must be improved to solve an optimization problem. The improvement includes redefining the parameters and operators of the PSO, adopting an energy-aware local fitness first strategy and designing a novel coding scheme. Using the improved PSO, an optimal virtual machine replacement scheme with the lowest energy consumption can be found. Experimental results indicate that our approach significantly outperforms other approaches, and can lessen 13%-23% energy consumption in the context of this paper.

With the rapid development of Cloud computing, more and more service providers could provide cloud services (applications) to users. Faced with mass Cloud services, trust and reputation mechanisms offer a promising way to solve the trust evaluation of Cloud services. Hence, trust and reputation play an important role in evaluating of Cloud services. In this paper, we propose a lightweight reputation measurement approach for Cloud services based on (user) feedback ratings. The proposed approach first adopts cloud model to obtain the trust vector of each cloud service by exploiting feedback ratings. The trust vector consists of Expected value, Entropy value and Hyper-Entropy value. Then we use fuzzy set theory to calculate the reputation scores of Cloud services. Simulation results show that the proposed approach is significantly effective for unstable feedback ratings.

With an increasing number of cloud service providers (CSP) delivering services to customers from the cloud, maximizing the profits of CSPs becomes a critical problem. Existing methods are difficult to solve the problem because they do not make full use of temporal price differences. This paper introduces a dynamic virtual resource renting method that attempts to dynamically adjust the virtual resource rental strategy according to price distribution and task urgency. We first pretreat the historical price series and adopt the outlier detection technique to filter the extreme price. Then, considering task urgency and price distribution, we design a weak equilibrium operator to calculate the acceptable price for each type of virtual resource. All types of virtual resources that are at an acceptable price are inserted into a set. Finally, we design a novel rental decision-making algorithm to select the most profitable resource from the set. We provide an extensive evaluation of our method using Amazon EC2 spot price dataset and normally distributed price dataset. The results demonstrate the effectiveness of our method.

ABSTRACT SYN flooding attacks are a common type of distributed denial‐of‐service attacks. Up to now, many defense schemes have been proposed against SYN flooding attacks. Traditional defense schemes rely on passively sniffing an attacking signature and are inaccurate in the early stages of an attack. These schemes are effective only at the later stages when attacking signatures are obvious. In this paper, we propose a detection approach that makes use of SYN traffic prediction to determine whether SYN flooding attacks happen at the early stage. We firstly adopt grey prediction model to predict SYN traffic, and then, we employ cumulative sum algorithm to detect SYN flooding attack traffic among forecasted SYN traffic. Trace‐driven simulation results demonstrate that our proposed detection approach can detect SYN flooding attacks effectively. Copyright © 2012 John Wiley & Sons, Ltd.

ABSTRACT Trusted quality of Web service (QoWS) issue is critical for e‐commerce applications. However, many existing studies have little work in situations that have insufficient or no historical information regarding QoWS data. In this study, we propose a trusted QoWS metric approach, that is, Bayesian Approach with Maximum Entropy Principle. The key of our proposed approach is to extract QoWS prior distribution of Web service by using Maximum Entropy Principle and then to infer QoWS posterior distribution of Web service by using Bayesian Approach. On the basis of the obtained QoWS posterior distribution, trusted QoWS can be measured. We conduct extensive simulations to evaluate our proposed approach. The simulation results demonstrate that our proposed approach can obtain trusted QoWS effectively. Copyright © 2012 John Wiley & Sons, Ltd.

Most of trustworthy web service selection simply focus on individual reputation and ignore the collaboration reputation between services. To enhance the collaboration trust during web service selection, a reputation model called collaboration reputation is proposed. The reputation model is built on web service collaboration network(WSCN), which is constructed in terms of the composite service execution log. Thus, the WSCN aims to maintain the trustworthy collaboration alliance among web services, In WSCN, the collaboration reputation can be assessed by two metrics, one called invoking reputation is computed by recommendation, which is selected from the community structure hiding in WSCN, the other is assessed by the invoked web service. In addition, the web service selection based on WSCN is designed.

In order to obtain trustworthy QoS (Quality of Service) in situations which have little or no information regarding a service's QoS, we propose a QoS measure approach, namely Bayesian Approach with Maximum-Entropy Principle (BA-MEP). BA-MEP firstly extracts the QoS prior distribution from the objective data (such as historical statistics data) and subjective data (such as the service providers and QoS experts) by Maximum Entropy Principle, and then Bayesian Approach is used to infer the QoS posterior distribution, finally, trustworthy QoS can be obtained from the QoS model. In addition, we also propose a trustworthy expert algorithm (TEA) and analysis three reasons that the QoS data of Web services are not always true. Some experiments are illustrated to show the effectiveness of BA-MEP.