TheComputing Power Network(CPN) is a distributed system that integrates computing resources to optimize utilization, but ensuringQuality of Service(QoS) is challenging due to high demand and complex heterogeneous connections. Accurate workload prediction is essential for maintaining QoS, yet the diverse and complex user requirements in CPN make prediction difficult. To address this challenge, we propose an ensemble workload prediction model with fluctuation division control for workload prediction in CPN, comprising three key components. First, we use theThree-Way Decision(3WD) approach to partition workload fluctuations, controlling granularity thickness and applying clustering to capture dynamic workload characteristics. Second, we develop tailored prediction methods for each of the three partitioned regions and ensembles them to enhance overall prediction performance. Third, the ensemble prediction method is applied to each region to obtain the final predicted values. The proposed method introduces an innovative fluctuation division control strategy for characteristic mining to capture dynamic workload fluctuation patterns and designs the effective ensemble workload prediction model deal with the problem of non-stationary workload prediction in CPN. Experimental results on trace datasets from Alibaba and Dinda demonstrate that the proposed model improves the higher average prediction accuracy by up to 26.06%$\sim$66.4% than the comparison methods.

Gradient leakage attacks pose a significant privacy threat in distributed learning-as-a-service APIs. Existing literature on gradient leakage defense relies on gradient perturbation for preventing privacy leakage. However, determining where and how much to perturb the gradient offers different capabilities for preventing gradient leakage. This paper presents GradCloak, a principled approach to guiding gradient perturbation with theoretical robustness bounds in federated learning as a service, aiming to find the minimum required noise for simultaneously achieving privacy protection, competitive accuracy, and preventing gradient leakage attacks. The paper is organized into three major components.First, we formulate the gradient leakage threats and their adverse effect. We categorize the attack into two broad types: leakage during local training and leakage before global aggregation.Second, we investigate different gradient perturbation approaches. We analyze and compare these gradient perturbation methods, which are performed at the federated server, with those performed at the participating client(s).Third, we introduce three robustness properties of robust perturbation against gradient leakage threats, formulated bythe anonymization boundfor training data robustness,the perturbation boundfor gradient robustness, andthe distribution robustness boundfor perturbed gradients. We conduct extensive evaluations on eight benchmark datasets to demonstrate that specific settings of gradient perturbation exist that best balance privacy, accuracy, and leakage prevention. Code is available athttps://github.com/git-disl/GradCloak.

Identifying potential academic collaborators has become increasingly crucial to promote scientific development. Existing methods of recommending collaborators primarily focus on analyzing publication content or integrating both content and academic network information. However, these methods often suffer from the cold-start problem due to sparse interaction data among researchers. To this end, we propose a novel meta-learning enhanced model on heterogeneous academic networks for academic collaborator recommendation, named MACRec. This model considers the academic collaborator recommendation in both data and model levels to address the cold-start problem. We introduce a multi-view task constructor that uses academic data and meta-paths to capture semantic contexts within an academic network at the data level, and we also include an academic meta-learner that performs semantic-wise and task-wise adaptations, allowing for fast adaptation to new researcher tasks with few data at the model level. Extensive experiments on two real-world datasets show that MACRec performs significantly better than state-of-the-art baseline methods.

Cloud-edge collaborative computing (CECC) facilitates the sharing of computing resources by collaboratively scheduling tasks among servers, thereby maximizing task execution efficiency. Task scheduling and resource allocation (TS-RA) are two interrelated issues that significantly affect the efficient utilization of computing resources. In this paper, we decouple the joint optimization problem of TS-RA and propose a novel model based on multi-agent reinforcement learning (TRMARL), which is applicable to distributed task scheduling and resource allocation in a heterogeneous CECC system. TRMARL consists of two modules: 1) the task scheduling module, where we introduce a value factorization algorithm to maximize joint rewards of distributed scheduling actions; 2) the resource allocation module, where we present a proximal policy optimization (PPO) algorithm based mechanism to optimize resource allocation. TRMARL efficiently captures the state difference among heterogeneous servers through a graph attention network-based recurrent deep Q-network (GAT-based recurrent-DQN) architecture and learns different strategies for heterogeneous services through a multi-expert schema. The experimental results demonstrate that TRMARL effectively improves the task completion rate, reduces average system latency, and enhances convergence stability in a heterogeneous CECC system.

The rise of social bots poses a significant threat to online platforms, making their detection an urgent priority.Recent advancements in Graph Neural Networks (GNNs) have significantly improved bot detection by leveraging the rich relational data within social networks.However, existing approaches face two key challenges: imbalanced feature fusion across different modalities and edge heterophily, which limit their effectiveness.To address these issues, we propose BotBR, a novel bot detection framework.BotBR tackles feature imbalance by employing decision trees to extract behavioral patterns from user-related numerical features and utilizes an attention mechanism to seamlessly integrate multi-modal feature embeddings.To mitigate edge heterophily, BotBR incorporates an edge detector to differentiate between high-and low-reliability edges, optimizing the utilization of structural graph information.Furthermore, a homophily-based graph is introduced for consistency contrastive learning, enhancing the model's robustness.Experimental results on real-world bot detection benchmark datasets demonstrate that BotBR achieves state-of-the-art performance while maintaining efficiency comparable to classical methods.

Large Language models (LLMs) usually rely on extensive training datasets. In the financial domain, creating numerical reasoning datasets that include a mix of tables and long text often involves substantial manual annotation expenses. To address the limited data resources and reduce the annotation cost, we introduce FinLLMs, a method for generating financial question-answering (QA) data based on common financial formulas using LLMs. First, we compile a list of common financial formulas and construct a graph based on the variables these formulas employ. We then augment the formula set by combining those that share identical variables as new elements. Specifically, we explore formulas obtained by manual annotation and merge those formulas with shared variables by traversing the constructed graph. Finally, utilizing LLMs, we generate financial QA data that encompasses both tabular information and long textual content, building on the collected formula set. Our experiments demonstrate that the synthetic data generated by FinLLMs effectively enhances the performance of various numerical reasoning models in the financial domain, including both pre-trained language models (PLMs) and fine-tuned LLMs. This performance surpasses that of two established benchmark financial QA datasets.

Collaborating with researchers from different academic domains has become more accessible with the emergence of academic social networking platforms. However, finding the right collaborator remains a challenge. Current cross-domain collaborator recommendation methods fail to provide explanations for their suggestions, thus requiring additional effort from users. In this paper, we propose the Explainable Cross-Domain Collaborator Recommendation (ECDCR) problem, which aims to provide users with collaborator recommendations and corresponding explanations. Our approach, named ECRec, constructs an academic network by using external information and extracting users’ meta-path connections to provide comprehensive explanations for their collaborative behavior. To ensure consistency between recommendation and explanation, we employ an attention mechanism that integrates multiple meta-paths to predict collaboration and selects the most important meta-path connection as an explanation. We conduct extensive experiments on three real-world datasets and demonstrate that our method has excellent explainability and recommendation performance.

Graph Neural Networks (GNNs) have become an important graph feature learning paradigm that is extensively applied to graph inference tasks. However, GNNs still have limitations in some aspects such as representation capability, interpretability, etc. To this end, this paper introduces a novel hybrid Fuzzy-Logic Graph Neural Network (FL-GNN) that synergistically combines Fuzzy Neural Network (FNN) and GNN to effectively capture and aggregate local information flows within the graph. FL-GNN exhibits three distinctive features. First, we incorporate a specialized structure fuzzy rule to enhance FL-GNN’s graph inference capability, surpassing representative GNN models. Second, we augment the interpretability of FL-GNN by integrating analytical exploration methods from two perspectives: Fuzzy Inference System and Message Passing Algorithm (MPA). Lastly, we refine the structure of FL-GNN based on MPA to optimize its calculation complexity and consequently improve learning efficiency. Experimental results show that FL-GNN outperforms existing representative graph neural networks for graph inference tasks.

Predicting information diffusion is a fundamental task in online social networks (OSNs).Recent studies mainly focus on the popularity prediction of specific content but ignore the correlation between multiple pieces of information.The topic is often used to correlate such information and can correspond to multi-source information.The popularity of a topic relies not only on information diffusion time but also on users' followership.Current solutions concentrate on hard time partition, lacking versatility.Meanwhile, the hop-based sampling adopted in state-of-the-art (SOTA) methods encounters redundant user followership.Moreover, many SOTA methods are not designed with good modularity and lack evaluation for each functional module and enlightening discussion.This paper presents a novel extensible framework, coined as HIF, for effective popularity prediction in OSNs with four original contributions.First, HIF adopts a soft partition of users and time intervals to better learn users' behavioral preferences over time.Second, HIF utilizes weighted sampling to optimize the construction of heterogeneous graphs and reduce redundancy.Furthermore, HIF supports multi-task collaborative optimization to improve its learning capability.Finally, as an extensible framework, HIF provides generic module slots to combine different submodules (e.g., RNNs,

In recent years,Coded Edge Computing(CEC) has been greatly studied as a promising technology to effectively mitigate the impact of stragglers and provide confidentiality in edge collaborative computing. It is crucial to verify the correctness of both intermediate results and the final result especially in untrustable and unreliable edge computing scenarios. However, the existing works on verification in CEC always verify and directly discard the whole incorrect intermediate results. In this paper, we propose thePartial Decode and Compare(PDC) verification scheme, which can fully utilize the correct part in the incorrect intermediate results to reduce the complexity and tolerate more abnormal edge devices. The PDC verification scheme consists of two parts:Final Result Verification(FRV) andAbnormal Edge Device Identification(AEDI). By deeply analyzing the decoding impact of the intermediate results on the final result, the PDC verification scheme divides the intermediate results and final results intosubresult vectors. It decodes, compares, and verifies the final result in units of subresult vectors. In this way, the obtained parts which verified to be correct do not need to participate in the following verification. Therefore, it can significantly reduce the verification overhead including both the number of required decoding rounds and the complexity of each decoding round. Based on the correct final result verified by the PDC verification scheme, we also propose anAbnormal Edge Devices Identificationscheme to identify all abnormal edge devices that return incorrect intermediate results. We then present extensive theoretical analyses and simulation experiments of the PDC verification scheme, which demonstrates that the PDC verification scheme can tolerate a higher ratio of incorrect intermediate results and achieve lower verification overhead than the state-of-the-art verification works. Therefore, the proposed PDC verification scheme enables CEC to provide reliable services in unstable and unreliable edge computing scenarios.

Competitive influence maximization (CIM) has been studied for years due to its wide application in many domains. Most current studies primarily focus on the microlevel optimization by designing policies for one competitor to defeat its opponents. Furthermore, current studies ignore the fact that many influential nodes have their own starting prices, which may lead to inefficient budget allocation. In this article, we propose a novel competitive bidding influence maximization (CBIM) problem, where the competitors allocate budgets to bid for the seeds attributed to the platform during multiple bidding rounds. To solve the CBIM problem, we propose a fairness-aware multiagent CBIM (FMCBIM) framework. In this framework, we present a multiagent bidding particle environment (MBE) to model the competitors’ interactions and design a starting price adjustment mechanism to model the dynamic bidding environment. Moreover, we put forward a novel multiagent CBIM (MCBIM) algorithm to optimize competitors’ bidding policies. Extensive experiments on five datasets show that our work has good efficiency and effectiveness.

The efficiency of cloud computing is significantly impacted by the capabilities of data center networks (DCNs). However, with the increasing demands of applications, the number of servers in DCNs is growing exponentially. In general, networks become more vulnerable as they expand in size. Consequently, it is essential to assess the network's reliability. Subsystem reliability, a crucial parameter of network reliability, is the likelihood that a faultless subsystem with a specific size would function normally. Even though several networks have the same order, their subsystem reliability may differ. In this article, we compare two distinct subsystems of$m$-port$k$-dimensional DCNs BCube that have the same number of vertices and ascertain the critical time so that the reliability bound is robust. In addition, we discover that, when comparing two subsystems of BCube that have the same number of vertices, the smaller the dimension$k$, the greater the subsystem reliability of BCube will be for$m\geq 2$and$k\geq 1$. This offers a theoretical foundation that allows us to choose a more reliable system in BCube that have the same number of vertices. Finally, we use numerical experiments to verify our findings.

Information diffusion can be regarded as the process of multi-user collaboration to deliver information. How to predict the cascade size is a fundamental task and has many applications such as rumor detection, product marketing, etc. Recent works attempt to mine temporal and structural characteristics hidden in the information cascade based on deep learning models. As we know, the complicated interactions between nodes are critical for cascade size prediction, and these interactions are usually hidden in the multiple types of cascade relationships. However, the cascade relationship diversity has not been comprehensively exploited by current studies. In this paper, we propose a novel model named CTformer, which leverages the global receptive field of Transformer to make accurate prediction. Specifically, CTformer takes advantage of both global position encoding and bias matrices to explore cascade relationship diversity. Extensive evaluation results on multiple real-world datasets show that CTformer achieves significant performance gains over the state-of-the-art methods.

Abstract As a kind of conditional connectivity, component connectivity is an improvement of traditional connectivity, which is conducive to enhance the reliability of the network. To be specific, the $r$-component connectivity of a network $G$, written as $c\kappa _{r}(G)$, is defined as the minimum number of all node cuts whose removal causes the remaining network to have at least $r$ components. Component diagnosability, as another measure of network reliability, is usually related to the number of components in the remaining network. The $r$-component diagnosability, written as $ct_{r}(G)$, is defined as the maximum number of faulty sets such that at least $r$ components in the surviving network and all faulty nodes can be diagnosed. This paper mainly explores the relationship between component connectivity and component diagnosability of some regular networks. Once knowing the component connectivity of such a network, with the help of this relationship, we can easily obtain the component diagnosability of the network. Furthermore, we apply this relationship to some famous regular networks to obtain their component diagnosabilities under the PMC model.

This paper presents a holistic approach to gradient leakage resilient distributed Stochastic Gradient Descent (SGD).First, we analyze two types of strategies for privacy-enhanced federated learning: (i) gradient pruning with random selection or low-rank filtering and (ii) gradient perturbation with additive random noise or differential privacy noise. We analyze the inherent limitations of these approaches and their underlying impact on privacy guarantee, model accuracy, and attack resilience.Next, we present a gradient leakage resilient approach to securing distributed SGD in federated learning, with differential privacy controlled noise as the tool. Unlike conventional methods with the per-client federated noise injection and fixed noise parameter strategy, our approach keeps track of the trend of per-example gradient updates. It makes adaptive noise injection closely aligned throughout the federated model training.Finally, we provide an empirical privacy analysis on the privacy guarantee, model utility, and attack resilience of the proposed approach. Extensive evaluation using five benchmark datasets demonstrates that our gradient leakage resilient approach can outperform the state-of-the-art methods with competitive accuracy performance, strong differential privacy guarantee, and high resilience against gradient leakage attacks.

Popularity prediction is to predict the number of social network users involved in information diffusion. Recently, deep learning methods for popularity prediction advance traditional approaches that rely on hand-crafted features. However, existing approaches ignore the multi-source cascade that consists of multiple sub-cascades with different content but under the same topic. Different from single-source cascade, more cascading information can be observed from multi-source cascade and they are potentially correlated. How to correlate the diverse information and take advantage of them from both temporal and spatial aspects is critical for prediction. To this end, we propose a novel framework, called HEterogeneous Recurrent Integrated Graph Convolutional Neural Network (HERI-GCN). Specifically, we construct a heterogeneous cascade graph to model the multi-source cascade where time intervals are treated as heterogeneous time nodes. Besides, we propose a heterogeneous GCN to learn rich features from the multi-source cascade. RNN is organically integrated into the heterogeneous GCN to overcome the limited learning ability toward temporal and spatial data. We evaluate HERI-GCN through comparative experiments on three datasets. The experimental evaluation shows that HERI-GCN outperforms the state-of-the-art baseline methods.

Abstract Connectivity and diagnosability are two crucial subjects for a network’s ability to tolerate and diagnose faulty processors. The $r$-component connectivity $c\kappa _{r}(G)$ of a network $G$ is the minimum number of vertices whose deletion results in a graph with at least $r$ components. The $r$-component diagnosability $ct_{r}(G)$ of a network $G$ is the maximum number of faulty vertices that the system can guarantee to identify under the condition that there exist at least $r$ fault-free components. This paper first establishes that the $(r+1)$-component connectivity of $k$-ary $n$-cube $Q^{k}_{n}$ is $c\kappa _{r+1}(Q^{k}_{n})=-\frac{1}{2}r^{2}+\Big(2n-\frac{1}{2}\Big)r+1$ for $n\geq 2$, $k\geq 4$ and $1\leq r\leq n$. In view of $c\kappa _{r+1}(Q^{k}_{n})$, we prove that the $(r+1)$-component diagnosabilities of $k$-ary $n$-cube $Q^{k}_{n}$ under the PMC model and MM* model are $ct_{r+1}(Q^{k}_{n})=-\frac{1}{2}r^{2}+\Big(2n-\frac{3}{2}\Big)r+2n$ for $n\geq 4$, $k\geq 4$ and $1\leq r\leq n-1$.

In recent years, edge computing (EC), as an emerging technology, has been widely used in various industries. It can meet the needs of industries in real-time business, application intelligence, security and privacy protection. However, edge devices may not always be trustworthy in the edge computing environment. Moreover, traditional edge computing systems have ignored the fact that the computation capability of user device can also be used. In this paper, we propose the Minimum Computation Latency Secure Edge Computing (MCLSEC) scheme to minimize computation latency and provide the security of computing data by utilizing linear coding and the resources of both edge devices and user device. Specifically, we consider the matrix multiplication as a computation task, which is an important module in many application operations, such as machine learning, big data analysis, etc. We firstly theoretically analyze the total computation latency of edge devices and user device in the coded edge computing. We then give the design of the MCLSEC scheme, which includes of the coding scheme and the task allocation scheme. Moreover, we also give theoretical analysis to show the proposed MCLSEC scheme is secure and optimal. Finally, we conduct extensive simulation experiments to show the effectiveness of the proposed scheme. Compared with the existing schemes, MCLSEC scheme significantly reduces the computation latency of edge computing while ensuring data confidentiality.

Abstract Connectivity is a classic measure for reliability of a multiprocessor system in the case of processor failures. Extra connectivity and component connectivity are two important indicators of the reliability of a multiprocessor system in presence of failing processors. The $h$-extra connectivity $\kappa _{h}(G)$ of a graph $G$ is the minimum number of nodes whose removal will disconnect $G$, and every remaining component has at least $h+1$ nodes. Moreover, the $h$-component connectivity $c\kappa _{h}(G)$ of $G$ is the minimum number of nodes whose deletion results in a graph with at least $h$ components. However, the extra connectivity and component connectivity of many well-known networks have been independently investigated. In this paper, we determine the relationship between extra connectivity and component connectivity of general networks. As applications, the extra connectivity and component connectivity are explored for some well-known networks, including complete cubic networks, hierarchical cubic networks, generalized exchanged hypercubes, dual-cube-like networks, Cayley graphs generated by transposition trees and hierarchical hypercubes as well.

Abstract As the number of servers in a data center network (DCN) increases, the probability of server failures is significantly increased. Traditional connectivity is an important metric to measure the reliability of DCN. However, the traditional connectivity of a DCN based on the condition of arbitrary faulty servers is generally lower. Therefore, it is important to increase the connectivity of a DCN by adding some limited conditions for the faulty server set. As a result, $g$-restricted connectivity and $h$-extra connectivity, which are two crucial subjects for a DCN’s ability to tolerate faulty servers, were proposed in the literature. In this paper, we study the $g$-restricted connectivity and $h$-extra connectivity of a new server-centric DCN, called BCDC, based on crossed cube with excellent performance. We prove that the $g$-restricted connectivity of BCDC is 4 for $n=3$ and $2n+g(n-2)-2$ for $n\geq 4$, where $0\leq g\leq n-3$, and the $h$-extra connectivity of BCDC is 4 for $n=3$ and $2n+h(n-2)-2$ for $n\geq 4$, where $0\leq h\leq n-3$.

In recent years, the computation security of edge computing has been raised as a major concern since the edge devices are often distributed on the edge of the network, less trustworthy than cloud servers and have limited storage/ computation/ communication resources. Recently, coded computing has been proposed to protect the confidentiality of computing data under edge device's independent attack and minimize the total cost (resource consumption) of edge system. In this paper, for the cooperative attack, we design an efficient scheme to ensure the information-theory security (ITS) of user's data and further reduce the total cost of edge system. Specifically, we take matrix multiplication as an example, which is an important module appeared in many application operations. Moreover, we theoretically analyze the necessary and sufficient conditions for the existence of feasible scheme, prove the security and decodeability of the proposed scheme. We also prove the effectiveness of the proposed scheme through considerable simulation experiments. Compared with the existing schemes, the proposed scheme further reduces the total cost of edge system. The experiments also show a trade-off between storage and communication.

In recent years, edge computing has attracted increasing attention for its capability of facilitating delay-sensitive applications. In the implementation of edge computing, however, data confidentiality has been raised as a major concern because edge devices may be untrustable. In this paper, we propose a design of secure and efficient edge computing by linear coding. In general, linear coding can achieve data confidentiality by adding random information to the original data before they are distributed to edge devices. To this end, it is important to carefully design code such that the user can successfully decode the final result while achieving security requirements. Meanwhile, task allocation, which selects a set of edge devices to participate in a computation task, affects not only the total resource consumption, including computation, storage, and communication, but also coding design. In this paper, we study task allocation and coding design, two highly-coupled problems in secure coded edge computing, in a unified framework. In particular, we take matrix multiplication, a fundamental building block of many distributed machine learning algorithms, as the representative computation task, and study optimal task allocation and coding design to minimize resource consumption while achieving information-theoretic security.

Edge computing is attracting more and more attention in recent years to fulfill the requirements of latency-critical and computation-intensive applications. By using the coding redundancy, coded edge computing has emerged to optimize the total computation latency. Compared with the servers in cloud computing, edge devices located at the edge of network may not be reliable and trustworthy. In coded edge computing, even one incorrect intermediate result will lead to the incorrect final result. Therefore, considering the low computation capabilities of edge devices and low latency requirements of user, we study the result verification problem for coded edge computing. Specifically, we propose an efficient Orthogonal Mark (OM) verification scheme by the properties of linear space. We also conduct solid theoretical analysis to show the successful verification probabilities under two kinds of attack models, respectively. Finally, we conduct extensive simulations to show the effectiveness of the proposed OM verification scheme when comparing with basic coded edge computing scheme and Decoding Comparison (DC) scheme.

This paper designs an Intelligent Ashcan Management System (IAMS) which is one of the most important part in a smart city. Traditional ashcan management is inefficient and has many disadvantages, because managers cannot obtain the realtime state of ashcans efficiently. As a result, it often happens that ashcans are full but not collected in time. Moreover, in special cases that ashcans fall, catch fire, etc., managers should find these ashcans and handle these emergencies as soon as possible. To manage ashcans efficiently, economically and intelligently, in this paper, we propose an edge computing based IAMS. Specifically, in IAMS, each ashcan has an intelligent user equipment (UE) equipped with sensors to measure distance, temperature, smog and tilt. For data transmission, IAMS uses Narrow Band Internet of Things (NB-IoT), which has advantages of large transmission range and low cost. Combined with edge computing, the collected data can be processed rapidly and managers can obtain the real-time state of each ashcan. Moreover, managers can view global information through web browser and mobile devices. According to the real-time state of ashcans, we also design an IAMS algorithm to get an efficient garbage collection path based on genetic algorithm. Finally, we deployed the proposed IAMS in Soochow University and experimental results show its efficiency and stability.

Influence Maximization (IM) has been extensively applied to many fields, and the viral marketing in today's online social networks (OSNs) is one of the most famous applications, where a group of seed users are selected to activate more users in a distributed cascading fashion. Many prior work explore the IM problem based on the assumption of given budget. However, the budget assumption does not hold in many practical scenarios, since companies might have no sufficient prior knowledge about the market. Moreover, companies prefer a moderately controllable viral marketing that allows them to adjust marketing decision according to the market reaction. In this paper, we propose a new problem, called Controllable social influence maximization (Cosin), to find a set of seed users inside a controllable scope to maximize the benefit given an expected return on investment (ROI). Like the IM problem, the Cosin problem is also NP-hard. We present a distributed multi-hop based framework for the influence estimation, and design a (1/2 + ϵ)-approximate algorithm based on the proposed framework. Moreover, we further present a distributed implementation to accelerate the execution of algorithm for large-scale social networks. Extensive experiments with a billion-scale social network indicate that the proposed algorithms outperform state-of-the-art algorithms in both benefit and running time.

Nowadays people tend to create accounts with multiple social networks (SNs) to enjoy a variety of social network services. User identity linkage (UIL) aims to identify those multiple accounts belonging to a same person. UIL is of great importance to user behavior understanding and prediction, information dissemination, viral marketing, wellness diagnosis, etc. Most of existing solutions typically rely on the embedding of either user's attributes or behaviors into a latent vector space, and establish anchor link based on vector distance. However, these efforts still face challenges originated from the heterogeneity of SNs, incompleteness of user information and lack of enough known anchor links. In this paper, we investigate the UIL problem by presenting a translation-modeling approach-TransLink. It jointly embeds both users and interactive behaviors of various SNs into a unified low-dimensional representation space according to a set of known anchor links. More specifically, we primarily study three typical SNs, e.g, Twitter, Foursquare and Instagram. Before embedding, we abstract schemas of three SNs and extract interaction metapaths for each SN. By doing this we can efficiently address the first two challenges. Furthermore, iterative linkage can ensure linkage performance by using a very small set of known anchor links. Experiment results on two real-world datasets demonstrate the superiority of TransLink over the state-of-the-art approaches.

Summary Currently, Mobile edge computing (MEC) is facing a great challenge that is how to make full use of edge resources to provide a seamless support for compute‐intensive latency‐sensitive applications. Prior studies often make a simple assumption that tasks can be executed upon every edge server, but the assumption does not hold in practical scenarios. Because a specific application task often corresponds to a certain service that provides the corresponding running environment, whereas an edge server only has limited resources and cannot offer too many services. How to decide service deployment of so many types of services among multiple edge servers is also a big challenge. To address the challenge, we study dynamic service deployment for latency‐sensitive applications. We first model the long‐term budget‐constrained latency minimization problem as a multi‐slot latency minimization problem based on the Lyapunov framework. By doing this, the hardness of a problem is significantly reduced, since we never require future information to solve the long‐term optimization. Furthermore, we extend our study by joining the task scheduling optimization, where every edge server is fully utilized in an even more efficient collaborative manner. Our extensive experiments show that the proposed algorithms can bring short latency with low cost.

In cognitive radio networks (CRNs), secondary users need to first discover neighbours and form communication links, referred to as the rendezvous process. Rendezvous between any two secondary users can only be achieved on the same channel. However, the nature of the CRN makes this a challenging problem. Specifically in CRN, not only the network is multi- channel, but the channels available at different nodes may be different. While most of the existing works study pair-wise rendezvous and design channel hopping sequence, in this paper, we focus on the performance improvement of the rendezvous process based on the existing channel hopping sequences with multiple users in CRN. We propose a new strategy, called Group Based Strategy (GBS) to achieve the acceleration, which is flexible to incorporate the existing sequence generation algorithms. Our basic idea is to group the encountered users and schedule rendezvous for them. With the purpose to increase rendezvous diversity, other users or groups can join the group if they get the group rendezvous information. Experiments are conducted to evaluate the proposed scheme. Overall, the performance can be improved by more than 50% under symmetric model or asymmetric model using our accelerating strategy.

A scalable storage system becomes more important today for online social networks (OSNs) as the volume of user data increases rapidly. Key-value store uses consistent hashing to save data in a distributed manner. As a defacto standard, it has been widely used in production environments of many OSNs. However, the random nature of hashing always leads to high inter-server traffic. Recently, partitioning and replication are respectively proposed in many existing works where the former aims to minimize the inter-server read traffic and the latter aims to optimize the inter-server write traffic. Nevertheless, the separated manners of optimization cannot efficiently reduce the traffic. Because the inter-server read traffic is changed during replication. In this paper, we suggest that performing partitioning and replication simultaneously could provide probability to further optimize traffic. Then we formulate the problem as a revised graph partitioning with overlaps, since overlaps partitioning naturally corresponds to replication. To solve the problem, we propose a Joint Partitioning and Replication (JPR) scheme. Through extensive experiments with a real world Facebook trace, we evaluate that JPR significantly reduces inter-server traffic with slightly sacrificing storage cost compared to hashing, and preserves a good load balancing across servers as well.

Summary With the increasing number of users and a huge scale of data, the service providers of Online Social Networks (OSNs) are facing the problem of how to place users' data to multiple servers. Key‐value stores solve the problem based on consistent hashing, and have become a defacto standard. However, random placement manner of hashing cannot preserve social locality, which leads to high intra‐data center traffic and unpredictable response time. Many existing works solve the problem by using graph partitioning algorithms. These works have two drawbacks: First, the social graph is constructed with ordinary pairwise graph that cannot fully reflect multi‐participant interactions often occurring in OSNs. Second, the underlying network topologies of data center have never been considered. This paper investigates the problem of traffic minimization for OSNs data storage. Motivated by maximally preserving both social locality and distance locality, we formulate the problem as two sub‐problems — hypergraph partitioning and partition‐to‐server mapping, and propose a two‐phase data placement (TDP) scheme. Specifically we present two algorithms to solve partition‐to‐server mapping over two widely used network topologies (i.e.,tree and BCube). Evaluations with a large scale Facebook trace show that TDP significantly reduces intra‐data center traffic as well as load balancing across servers. Copyright © 2016 John Wiley & Sons, Ltd.

Summary Personal cloud storage provides users with convenient data access services. Service providers build distributed storage systems by utilizing cloud resources with distributed hash table (DHT), so as to enhance system scalability. Efficient resource provisioning could not only guarantee service performance, but help providers to save cost. However, the interactions among servers in a DHT‐based cloud storage system depend on the routing process, which makes its execution logic more complicated than traditional multi‐tier applications. In addition, production data centers often comprise heterogeneous machines with different capacities. Few studies have fully considered the heterogeneity of cloud resources, which brings new challenges to resource provisioning. To address these challenges, this paper presents a novel resource provisioning model for service providers. The model utilizes queuing network for analysis of both service performance and cost estimation. Then, the problem is defined as a cost optimization with performance constraints. We propose a cost‐efficient algorithm to decompose the original problem into a sub‐optimization one. Furthermore, we implement a prototype system on top of an infrastructure platform built with OpenStack. It has been deployed in our campus network. Based on real‐world traces collected from our system and Dropbox, we validate the efficiency of our proposed algorithms by extensive experiments. Copyright © 2016 John Wiley & Sons, Ltd.

The complete binary tree is an important network structure for parallel and distributed computing, which has many nice properties and is often used to be embedded into other interconnection architectures. The locally twisted cube LTQn is an important variant of the hypercube Qn. It has many better properties than Qn with the same number of edges and vertices. In this paper, we prove that the complete binary tree CBTn can be embedded with dilation 2 and congestion 1 into LTQn. Furthermore, it is proven that if there exists an arbitrary faulty node in LTQn, then both the dilation and congestion values will become 2 after reconfiguring CBTn, while if there are two arbitrary faulty nodes in LTQn, then both the dilation and congestion values will become 3 after reconfiguration.

Grid service composition has been recognized as a flexible way for resource sharing and application integration since appearance of service-oriented architecture. Approaches are needed to select service candidates with various Quality of Service (QoS) levels according to use’s performance requirements. In this paper We model this problem as the Multi-Constrained Optimal Path Selection Problem (MCOP), due to the dynamic property of grid service, adaptive mechanism is introduced to ensure the whole QoS when some service candidates fail. An algorithm QAGSS is proposed. Simulation results show that QAGSS has greater success rate and lower cost than previous algorithms.