Federated Learning (FL) is an emerging distributed machine learning (ML) technique that enables in-situ model training and inference on decentralized edge devices. We propose Totoro$^+$, a novel scalable FL system that enables massive FL applications to run simultaneously on edge networks. The key insight is to explore a distributed hash table (DHT)-based peer-to-peer (P2P) model to re-architect the centralized FL system design into a fully decentralized one. In contrast to previous studies where many FL applications shared one centralized parameter server, Totoro$^+$assigns a dedicated parameter server to each application. Any edge node can act as any application's coordinator, aggregator, client selector, worker (participant device), or any combination of the above, thereby radically improving scalability and adaptivity. Totoro$^+$introduces three innovations to realize its design: a locality-aware P2P multi-ring structure, a publish/subscribe-based forest abstraction, and a game-theoretic path planning model with a guarantee of an$\epsilon$-approximate Nash equilibrium. Real-world experiments on 500 Amazon EC2 servers show that Totoro$^+$scales gracefully with the number of FL applications and$N$edge nodes speeds up the total training time by$1.2\times -14.0\times$, achieves$\mathcal {O}(\log N)$hops for model dissemination and gradient aggregation with millions of nodes, and efficiently adapts to the practical edge networks and churns.

Satellite-assisted quantum network (SQN) is emerging as a promising solution to overcome the distance limitations of ground-based fiber quantum network (QN). However, each satellite and ground station has a limited number of transmitters and receivers, respectively, and the Entanglement Distribution Rate (EDR) decreases with the distance between satellites and ground stations, highlighting the need for effective resource allocation to serve requests in the SQN. In this paper, we present a novel optimization problem, termed ESOP, which aims to maximize the total EDR in the network while simultaneously considering the resource capacities of both satellites and ground stations, as well as the fidelity requirements of individual requests. To solve ESOP, we propose a (2 + ϵ)-approximation algorithm, AESOP, which combines a greedy approach with a tailored local search. Simulation results show that AESOP achieves up to 64% improvement in total EDR compared to the existing method.

Federated Learning (FL) enables multiple client devices to train a global model collaboratively by sharing only local gradient updates, thus preserving data privacy. Despite its advantages, FL remains vulnerable to Gradient Inversion Attack, where attackers exploit gradient updates to reconstruct private data. To counter these advanced threats, we propose Latent Interpolation Data Synthesis (LIDS), a client-side privacy-preserving training scheme that obfuscates per-sample gradient signals via structured batch construction and latent oversampling. LIDS groups semantically similar samples and generates synthesized variants, reducing the uniqueness of each gradient contribution while preserving model convergence. Experimental results show that LIDS reduces attacker reconstruction quality, with Peak Signal-to-Noise Ratio dropping from 27 to 17. On CIFAR-10, the attacker’s reconstruction label accuracy and fraction of successful reconstructions are reduced by up to 64% and 92%; on CIFAR-100, by 68% and 96%. These privacy gains come at a modest cost of only 11% test accuracy degradation in CIFAR-100.

Quantum computing is widely regarded as a promising technology due to its potential to offer exponential computational advantages over classical systems. However, practical deployment and realization of most quantum algorithms remain constrained by the limitations of current quantum architectures– most notably, the limited number of physical qubits available in quantum processing units (QPUs). Therefore, efficient utilization of quantum resources has become a crucial concern in the field. Thus, in this paper, we explore the nature of qubit reuse. Based on our findings, we formulate an optimization problem and propose a novel algorithm to minimize the number of qubits for scheduled quantum circuits. Finally, extensive simulation results on various practical benchmarks manifest that our algorithm can significantly outperform the existing approach in qubit reuse.

Quantum nodes exchange qubits to achieve largescale quantum computing applications in a quantum network (QN). However, due to environmental influences, entangled links used for data transmission would decohere over time. Besides, the scarcity of quantum memory in each quantum node necessitates frequent communication between nodes, leading to significant variations in the traffic matrix. Thus, providing real-time services in QNs becomes particularly important. To this end, this paper presents a new optimization problem and proposes a novel online algorithm, which combines the game-theoretic analysis results with the linear programming (LP) by the primal-dual technique. Finally, the simulation results demonstrate that the proposed algorithm outperforms the other baselines by up to 90 %.

With the enhanced security offered by quantum teleportation, quantum networks (QNs) are gradually gaining significant attention. However, previous research on QNs often exhausts all network resources to serve requests, hence overlooking the scarce resources in QN. This oversight easily leads to resource wastage and requests competing for limited resources. To address the above issues, this paper introduces a new optimization problem named EINS, which minimizes the maximum resource utilization and finds routing paths for each request through certain intermediate nodes. This design cleverly divides the routing path into segments to enhance routing path diversity. To solve the EINS, we design a novel algorithm named GOAL, which provides an$O(\log\vert V\vert)$bound for constraint deviation. It efficiently and equitably distributes requests, ensuring optimal network utilization. Finally, the simulation results manifest that GOAL can outperform the existing methods by up to 99%.

Quantum key distribution (QKD) is currently the only visible technology for secure symmetric key exchange between communicating parties. However, existing quantum photon sources (PSs) in QKD networks for generating keys between nodes are costly but offer limited achievable key rates. Moreover, achievable key rates across links diminish drastically with link distance, underscoring the importance of effectively identifying relay paths and strategically allocating PS resources. To optimize network costs, it is essential to jointly deploy PSs, allocate key rates across links, and determine relay paths based on anticipated traffic. To this end, we formulate a novel optimization problem, termed DAP, which simultaneously considers PS placement, key rate allocation, and relay path identification. Our proposed$O(\log\vert V\vert)$approximation algorithm, ADAP, leverages an advanced linear programming (LP) conversion with tailored rounding techniques. Simulation results manifest that ADAP achieves at least an 83 % reduction in the number of PSs.

Most quantum network schemes delay entanglement generation until a request arrives, causing slower processing. To this end, an approach of pre-establishing an initial shared state has emerged. However, the initial shared state must be versatile enough to accommodate all possible requests and may consume considerable qubits. It is crucial to minimize the number of qubits used while satisfying every possible request. We first introduce a 2 -approximation algorithm for the special case where each request consists of only one Bell or GHZ state requirement. Afterward, the 2 -approximation algorithm is extended to handle any possible request that may contain one or more requirements. Finally, via extensive simulation results, we show that our algorithm can outperform existing approaches by up to 29% in used qubits.

To provide real-time map information, roadside units (RSUs) near vehicles cache 3D map tiles to reduce transmission delays and alleviate backhaul congestion. Unfortunately, traditional cache problems primarily focus on maximizing the cache hit rate (i.e., popularity) while often neglecting other criteria, thus suppressing caching efficiency. For instance, the priority of each map tile may vary for different vehicles based on their route plans and current positions. Additionally, caching a greater number of small-scale map tiles with broader coverage helps serve more vehicles’ requests and reduce cache misses. However, these tiles may only be sufficient for vehicles that can tolerate lower detail levels, as they provide less information than large-scale tiles. Furthermore, ensuring the freshness of cached map tiles is essential for maintaining driving safety. To address the interplay among popularity, priority, map scale, and information freshness in online map caching, this paper formulates an online optimization problem and proposes a novel online algorithm called ADAM, which integrates mixed-integer linear programming (MILP) with deep reinforcement learning. The ADAM predicts future total costs to support caching decisions by leveraging an ingeniously designed MILP that emulates the behavior of a neural network. Finally, simulation results manifest that the proposed ADAM outperforms existing methods by an average of 50%.

Convolutional Neural Network (CNN)-based Intrusion Detection System (IDS) incorporated with Federated Learning (FL) facilitates collaborative model training to secure user privacy and enhance data diversity, emerging as a promising solution to defend malicious flows. Yet, FL is vulnerable to trigger backdoor attacks. Existing methods have limited ability on IDS with FL for detecting or repairing a converged attacked model as they either require access to the training process or are dedicated to specific applications. To tackle the limitations, we propose a novel Features Distraction Defense Framework (FDDF) that does not interfere with the model training and requires only a few additional model inferences to protect users from the aforesaid attack. The experiment results show that FDDF can recover the accuracy effectively and preserve the accuracy of normal models.

Quantum networks (QNs) gradually gain significant attention due to their higher security compared with classical networks. Conventional approaches in QN routing often aggregate multiple requests into a batch before determining their routing paths. However, this approach may overlook the limited lifetime of qubits, resulting in critical decoherence. In this paper, we present a new online entanglement routing architecture with an online optimization problem and propose a novel [1, O(log |V|)]-competitive algorithm supporting online requests with admission control, aiming to maximize the number of admitted requests. Finally, extensive simulation results show that our algorithm can outperform the existing approaches by up to 98%.

Quantum teleportation enables high-security communications via quantum entanglement. However, decoherence, signal decay, and environmental interference may cause imperfect entangled pairs with low fidelity. Such entangled pairs may easily generate errors when used to teleport data qubits and cause fatal computation errors on quantum computers. Fortunately, data qubits can be encoded via quantum error correction (QEC) code to recover the detected error to some extent. Nevertheless, letting any repeaters process data qubits is dangerous because malicious repeaters may peep at, destroy, or fake the data qubits. Thus, in this paper, we propose a novel QEC-enabled routing framework MOON facilitated with the concept of social networks (SNs) for quantum networks (QNs). MOON selects only trusted repeaters to process data qubits and exhibits elastic routing with QEC to maximize the throughput without errors. Last, simulation results show that our framework can outperform existing approaches by up to 34% on network throughput without errors.

Entangled pairs serve as the cornerstone for secure data transmission. All-optical-switching technology on nodes enables the entangling signals to bypass nodes and build ultra-long entangled pairs. Nevertheless, entangled pairs suffer from decoherence over distance, causing inadequate fidelity and potentially compromising transmission quality. To address the challenges, we employ entanglement purification to enhance fidelity to meet the threshold. However, the purification process consumes additional entangled pairs and may fail. Besides, the purification efficiency would be poor if the input pairs have low fidelity. Thus, it is unavoidable to divide the path into sub-paths with appropriate lengths for better purification efficiency and then merge them into a longer entangled pair by swapping. The novel optimization problem DOSP then emerges: maximizing the probability while adhering to fidelity constraints. To tackle the DOSP efficiently, we propose a (1–δ)-approximation algorithm NSPS to consider the probability and fidelity jointly, where is a positive user-defined constant. Finally, the simulation results manifest that the NSPS can outperform the existing methods by at least 70%.

Traditional end-to-end entanglement typically operates between two parties. However, such a setup may fall short when three parties are involved. GHZ states, a multi-qubit entangled state, provide an approach to such a problem. A fusion node is first chosen, and then the paths from all end nodes to the fusion node are fused to create an entangled state between all parties. The selection of the fusion node becomes critical since it highly affects the success probability of the whole process. Thus, in this paper, we explore the promising scenario for multiple requests of authorizable tripartite teleportation and introduce a novel optimization problem to maximize the (expected) total profit for 3-GHZ requests. Via extensive simulation results, we show that our algorithm can outperform existing approaches significantly.

Telecom carriers have announced new content services by making a partnership with content providers and offered popular video streaming to customers. Via the integration with edge networks, telecom carriers can procure various contents from content providers and place them on edge servers proximate to end users to serve real-time requests with high bandwidth and ultra-low latency. Nevertheless, it is challenging to consider the content procurement, placement, and services jointly due to the user preference, user distribution, storage capacity of edge server, economic costs, etc. Telecom carriers would like to balance the procuring, placing, and transfer costs. To address this problem, the paper formulates an optimization problem and then proposes an approximation algorithm. Finally, the simulation results manifest that our algorithm outperforms other baselines.

Establishing entangled links by entangling and swapping can enable end-to-end teleportation, avoiding eavesdropping and facilitating the critical applications, such as quantum key distribution (QKD) service. For these promising applications, a specific number of data qubits should be transmitted (periodically) to meet the quality of service (QoS) requirements for each request. Otherwise, the network provider cannot earn the profit from the request. In this paper, we present a new optimization problem and propose a novel approximation algorithm to maximize the (expected) total profit under the constraint of network resources. Finally, extensive simulation results manifest that our algorithm can outperform the existing approaches by up to 46%.

Dockless electric bike (E-bike) sharing has become a new urban modality of green transportation to offer convenient services. Typically, the service provider arranges a truck starting from the depot to visit multiple parking locations to replace low-energy batteries. However, visiting many parking locations may cause a considerable tour cost. One efficient way is to aggregate low-energy E-bikes together. Some incentive mechanisms are thus adopted to encourage E-bike users to move their bikes to suitable parking locations, but leading to an incentive cost. The service provider would like to balance the tour cost of the truck and the incentive cost of E-bike users. To address this problem, the paper formulates an optimization problem and then proposes an approximation algorithm. The simulation results with the real dataset show that our algorithm outperforms the other baselines.

The rapid growth of Internet of Things (IoT) applications has spurred the need for efficient data collection mechanisms. Traditional approaches relying on fixed infrastructure have limitations in coverage, scalability, and deployment costs. Unmanned Aerial Vehicles (UAVs) have emerged as a promising alternative due to their mobility and flexibility. In this paper, we aim to minimize the number of UAVs deployed to collect data in IoT networks while considering a delay budget for energy limitation and data freshness. To this end, we propose a novel 3-approximation dynamic-programming-based algorithm called GPUDA to address the challenges of efficient data collection from IoT devices via UAVs for real-world scenarios where the number of UAVs owned by an individual or organization is unlikely to be excessive, improving the best-known approximation ratio of 4. GPUDA is a geometric partition-based method that incorporates data rounding techniques. The experimental results demonstrate that the proposed algorithm requires 35.01% to 58.55% fewer deployed UAVs than the existing algorithms on average.

Vehicle-to-everything (V2X) communication is one of the key technologies of 5G New Radio to support emerging applications such as autonomous driving. Due to the high density of vehicles, Remote Radio Heads (RRHs) will be deployed as Road Side Units to support V2X. Nevertheless, activation of all RRHs during low-traffic off-peak hours may cause energy wasting. The proper activation of RRH and association between vehicles and RRHs while maintaining the required service quality are the keys to reducing energy consumption. In this work, we first formulate the problem as an Integer Linear Programming optimization problem and prove that the problem is NP-hard. Then, we propose two novel algorithms, referred to as “Least Delete (LD)” and “Largest-First Rounding with Capacity Constraints (LFRCC).” The simulation results show that the proposed algorithms can achieve significantly better performance compared with existing solutions and are competitive with the optimal solution. Specifically, the LD and LFRCC algorithms can reduce the number of activated RRHs by 86$\%$and 89$\%$in low-density scenarios. In high-density scenarios, the LD algorithm can reduce the number of activated RRHs by 90$\%$. In addition, the solution of LFRCC is larger than that of the optimal solution within 7$\%$on average.

For critical applications such as emergency medical rescue missions or telehealth in remote areas, stable network connectivity is vital for patient state monitoring and proper temporary care. Unexpected connection interruption or network lag can cause trouble for skilled doctors in remote care centers to predict the progress of a patient’s condition. Network quality is variable in many areas because of signal power degradation (zones without purple coverage) in rural areas with many building obstacles. As a result, many current emergency services still rely on on-site first aid efforts. The idea of unmanned aerial vehicles (UAVs) serving as aerial relaying stations to provide connectivity for ground users has received much attention over the years. However, controlling UAVs via cellular networks is still a challenging issue. In this work, we consider the mission of dispatching UAVbased relaying stations as a path-planning scheme, where the UAVs go to planned locations and serve the EVs with a certain connectivity requirement. The core novelty of this work is a novel searching scheme that can suggest a deployment plan for the swarm of UAVs at the time of the EVs’ departure. The search is also robust for path planning with real-time applications or dynamic environments.

The applications of Automatic Speech Recognition (ASR) on Internet-of-Things (IoT) devices have increased significantly in recent years, and Federated Learning (FL) is often used to improve ASR performance since its decentralized training mechanism ensures users' data privacy. However, FL is vulnerable to various attacks. The most challenging one to detect and defend against is trigger backdoor attack. Adversaries inject the trigger into the training audio data and participate in the FL training, causing the converged global model to mispredict the poisoned data. Unlike previous defense methods filtering suspicious models during model aggregation, we propose the Knowledge Distillation Defense Framework (KDDF) to detect and remove features of the potential triggers during the inference. KDDF utilizes Knowledge Distillation (KD) to train a validation model on each IoT device, which is used to identify suspicious data. Then, KDDF would try to eliminate the injected trigger during the model inference if the data is suspicious. Experimental results show that KDDF can effectively distinguish between benign and suspicious data and recover the classification results of suspicious data.

The conventional quantum teleportation schemes enable high-security network communications by establishing end-to-end entangled paths. However, those schemes focus on time synchronization and thus cause lots of idle time. Recent research suggests adopting an opportunistic scheme to forward data qubits as far as it can. However, this scheme lacks security since data qubits may be stored at malicious repeaters, which may peek at, destroy, or fake the data qubits. To this end, we design a new scheme called SOAR that considers trusted repeaters via social networks. Moreover, SOAR promotes the parallelism of swapping processes and thus leads to a less idle time of network resources than the other existing schemes. Furthermore, we design an algorithm called SAGE that can best fit SOAR by linking multiple subpaths via appropriate trusted repeaters to get an ideal path and augmenting least-hop paths to utilize the resources in quantum networks better. Simulation results manifest that SOAR outperforms the other schemes by 54%-89%; SAGE outperforms the other routing algorithms by 50% on average on SOAR.

The increasing diversification of network attacks has posed many security threats. Even within a local area network, different hosts may encounter distinct attacks. Leveraging the intrusion data dispersed across various hosts is crucial to achieving more comprehensive intrusion detection. Decentralized learning has emerged as a promising solution by enabling hosts to share information in a peer-to-peer manner. However, the imbalanced nature of intrusion data and varying data distributions between hosts can significantly impact model performance. To address the challenges of imbalanced and non-IID data, we propose a Decentralized Learning-based Intrusion Detection System (DLIDS). It rebalances training data to mitigate the model's bias towards the majority class and periodically substitutes the training model to facilitate knowledge acquisition. Moreover, the ensemble method is incorporated to integrate diverse perspectives and generate unbiased predictions. Finally, the experiment results on CSE-CIC-IDS2018 dataset show that the proposed method performs well even under imbalanced and non-IID data conditions.

Federated Learning (FL) allows Internet-of-Things (IoT) devices to train a global model collaboratively and keep their data locally to address privacy concerns. However, the current FL framework has three main drawbacks, high communication cost, single point of failure, and low accuracy on non-independent and identically distributed (non-IID) data. To this end, we propose a novel FL framework, IHC-FL, to 1) group devices into clusters based on communication cost and model distance, 2) distribute model aggregation over cluster heads, and 3) construct a topology to guide cluster heads to exchange model updates. To the best of our knowledge, this paper makes the first attempt to jointly optimize grouping user devices into clusters and exchanging model updates among cluster heads to enhance model performance. The numeric results show that IHC-FL can reduce 38%~89% of total communication cost over time than other heuristics with non-IID data on FMNIST and CIFAR-10 to achieve the target accuracy.

Federated Learning (FL) provides a decentralized training mechanism that ensures users' data privacy. However, FL is vulnerable to backdoor attacks, a type of data poisoning attack. The adversaries tampered with the local models by injecting a trigger into a subset of training data. After the aggregation process, the global model would be poisoned and mispredict the input images that injected a trigger designed by an adversary. Unlike the existing defense methods attempting to identify and remove the abnormal model updates on the aggregation step, this paper proposes a Successive Interference Cancellation-based Defense Framework (SICDF) to detect and eliminate the trigger during model inference. SICDF first employs Explainable AI to infer where the trigger is and then uses image processing skills to eliminate potential trigger effects. Experiment results show that SICDF can effectively recover the poisoned data while only slightly reducing the accuracy of the clean model and benign data.

Quantum teleportation through quantum entanglement over optical fiber channels enables secure communications. However, physical obstacles such as oceans and mountains may block optical fiber channels, and thus quantum networks (QNs) may have multiple disjoint domains. To overcome the issue, in this paper, we first propose a promising framework termed SSR to leverage satellite-based Free-Space Optical (FSO) channels to create inter-domain entangled paths for long-distance requests. Still, FSO channels may be intermittent due to satellite sparsity and atmospheric turbulence. Then, we introduce two algorithms, named RAIN and IDOL, to efficiently utilize FSO channels in SSR. RAIN estimates the existence probability of FSO channels to plan an appropriate inter-domain routing for each request while balancing domain loads. IDOL selects desired intra-domain paths and trusted repeaters for each request to teleport data qubits. Finally, simulation results manifest that SSR runs efficiently and outperforms existing approaches by 31%–57% in throughput.

Promising vehicle-to-everything (V2X) communication technologies can increase road safety by periodically broad-casting Cooperative Awareness Messages (sCAMs) that contain vehicles' status and attribute information, such as time, location, velocity, motion state, and vehicle type, to all nearby vehicles. However, out-of-date information and prediction deviations may cause potential risks and severe vehicle safety problems. In this paper, we propose an efficient safety-aware CAM broadcast rate control algorithm termed DESBRAC for vehicles to consider more safety metrics and determine the CAM broadcast rates cooperatively. Furthermore, we introduce Over-the-Air Computation (AirComp) to help vehicles aggregate information from their nearby vehicles instantly for metric estimation and cooperative CAM broadcast rate determination. Finally, the simulation results based on a simple and a realistic scenarios of vehicular networks show that our algorithm can achieve an improvement of about 31% in driving safety compared to the state-of-the-art algorithms.

Federated learning (FL) emerges to mitigate the privacy concerns in machine learning-based services and applications, and personalized federated learning (PFL) evolves to alleviate the issue of data heterogeneity. However, FL and PFL usually rest on two assumptions: the users' data is well-labeled, or the personalized goals align with sufficient local data. Unfortunately, the two assumptions may not hold in most cases, where data labeling is costly, or most users have no sufficient local data to satisfy their personalized needs. To this end, we first formulate the problem, DoLP, that studies the issue of insufficient and partially-labeled data on FL-based services. DoLP aims to maximize two service objectives: 1) personalized classification objective and 2) the personalized labeling objective for each user within the constraint of training time over wireless networks. Then, we propose a PFL-based service system DoFed-SPP to solve DoLP. The DoFed-SPP's novelty is two-fold. First, we devise an inference-based first-order approximation metric, similarity ratio, to identify the similarity between users' local data. Second, we design an approximation algorithm to determine the appropriate size and set of users for uploading in each round. Extensive experiments show DoFed-SPP outperforms the state-of-the-art in final accuracy and time-to-accuracy performance on CIFAR10/100 and DBPedia.

Quantum teleportation via quantum entanglement enables high-security communications in networks. However, if two quantum nodes are far away, it may be difficult to create an entangled path due to the low success probability. Besides, existing approaches neglect social relations among nodes' owners. In this paper, we propose a new framework SEER to minimize the waiting time of all source-destination (SD) pairs' requests. SEER has two promising features: 1) Social-relation Consideration. SEER makes the first attempt to select trusted owners' nodes via social networks as intermediate nodes for requests to temporarily store data qubits to increase the success probability. 2) Starvation Mitigation. SEER divides long SD pairs and slices resources reasonably to remedy starvation due to their low success probability. To this end, we design RATE and PLAN to find the proper intermediate nodes and reduce the average waiting time. Simulation results manifest SEER outperforms others by 37%.

Cooperative intelligent transport systems (C-ITS) enable fast communication among vehicles, infrastructure, and other road users. Via C-ITS, vehicles can acquire the real-time traffic information from various C-ITS stations, especially the information of the highly dynamic data in the local dynamic map (LDM). In the paper, we consider the message broadcast from the trustworthy roadside unit (RSU) to vehicles, where the messages have different popularity levels (i.e., how many vehicles request for this message), generation time of last-received corresponding messages, and priority towards each vehicle. Besides, the freshness of information of messages needs to be as fresh as possible. To explore the non-trivial order of transmission in time under the limited bandwidth, we formulate an optimization problem named ROAD to find an efficient transmission schedule and propose an approximation algorithm termed ABS. Experiment results manifest that ABS outperforms traditional approaches by 40%.

A powerful Bit-Flip Attack (BFA), based on Row Hammer Attack (RHA), can precisely flip the most vulnerable bits in the memory system (i.e., DRAM) to crash Convolutional Neural Networks (CNNs) run on Internet-of-Things (IoT) devices. However, it is very difficult to detect BFA since most devices are usually with limited computation capability and unaware of the security issue. Therefore, BFA becomes one of the most crucial threats to IoT networks. To this end, we design a novel defense system termed Resilient Dual-mode Defense System (RIDES) to encourage IoT devices with social relationships (i.e., Social IoT (SIoT)) to collaborate on BFA detection in an online manner. Subsequently, a new online problem, Online Computing Unit Assignment Problem (OMAR), is formulated to optimize the total inference rate for detecting BFA. To address OMAR's challenges, we present an online algorithm, Socially-aware Checker Assign-ment Algorithm (SCAN), to achieve the optimal competitive ratio. Extensive experiment and simulation results manifest that RIDES effectively detects BFA and in average, SCAN increases the total inference rate by 8%-515% and reduces the average overhead per image by 31%-55% compared with other solutions.

Flow consolidation has been proposed for merging multiple flows from different services into an aggregate flow to remedy the state explosion problem in software-defined networks (SDN). However, we observe that the Quality of Service (QoS) requirements are no longer sustained in aggregate flows since the bandwidth decided by TCP is usually different from the desired rate of each service. Therefore, this article explores an idea to control the rates of only a few service flows so that the rates of all uncontrolled flows allocated by TCP will meet their QoS requirements. We design a new architecture, called Scalable Per-Flow Rate Allocation (SPFRA), and formulate a new optimization problem, termed Scalable Rate Allocation for Aggregate Flows (SRAF), to find a minimum number of controlled flows to increase the scalability of SDN with diverse service requirements. We prove the NP-hardness and inapproximability of SRAF. To solve the problem, we design an algorithm, named Aggregate Flow Selection and Flow Release (AFSFR), to achieve the tightest bound and extend it to support distributed computation and dynamic traffic for instant services. Simulations and implementation on an SDN testbed manifest that AFSFR performs nearly optimally in real networks, and the number of controlled flows can be effectively reduced by 50 percent.

Online Decentralized Learning (ODL) is suitable for Internet-of-Things (IoT) devices since only parameter updates are exchanged with neighbors to avoid uploading private data to a central server and the training data is allowed to arrive at the devices sequentially. However, the current ODL frameworks cannot support the emerging Social IoT (SIoT) paradigm favorably since the SIoT devices exchange parameter updates with only trust-worthy neighbors based on specific social relations (e.g., parental object relation and ownership object relation). Conversely, sharing parameter updates with untrustworthy neighbors could speed up the training process but may violate social relations. Differential privacy (DP) is thus used to ensure data security while excessive devices engaging DP may downgrade the training performance. However, most research neglects the effect of neighbor selection for each device based on social networks, physical networks, and DP. Thus, in this paper, we innovate an ODL framework ODLF-PDP to allow only a part of devices to engage DP (i.e., partially DP) to improve training performance. Then, an algorithm BeTTa is proposed to build an adequate communication topology based on the interplay among the social networks, physical networks, and DP. Last, the experiment results manifest that ODLF-PDP saves more than 20% physical training time compared to the current frameworks via the benchmark of MNIST.

Federated learning (FL) enables multiple user de-vices to collaboratively train a global machine learning (ML) model by uploading their local models to the central server for aggregation. However, attackers may upload tampered local models (e.g., label-flipping attack) to corrupt the global model. Existing defense methods focus on outlier detection, but they are computationally intensive and can be circumvented by advanced model tampering. We employ a shuffling-based defense model to isolate the attackers from ordinary users. To explore the intrinsic properties, we simplify the defense model problem and formulate it as a Markov Decision Problem (MDP) to find the optimal policy. Then, we introduce a novel notion, (re)grouping, into the defense model to propose a new cost-efficient defense framework termed SAGE. Experiment results manifest that SAGE can effectively mitigate the impact of attacks in FL by efficiently decreasing the ratio of attacker devices to ordinary user devices. SAGE increases the testing accuracy of the targeted class by at most 40%.

Federated Learning (FL) allows Internet-of-Things (IoT) devices to train a global model collaboratively and circumvent the security issue. However, the current FL framework has three main drawbacks, the huge network overhead, single point of failure, and accuracy degradation in non-independent-and-identically-distributed (non-IID) data distribution. We propose a novel Deep Reinforcement Learning (DRL) based Decentralized Learning (DL) framework, DeepSelect, to 1) reduce the network overhead of conventional FL, 2) construct a good communication topology adaptively to mitigate the effect of non-IID data, and 3) accelerate the DL training by balancing the effects of hitting time (HT) and data bias. Moreover, DeepSelect with a subtly-designed DRL agent is reusable with different levels of non-IID data distributions. To the best of our knowledge, this paper is the first one to indicate that proper neighbor selection for exchanging parameters (not raw data) can counterbalance the data bias's effect and improve the DL convergence with non-IID data. The experiment results show that DeepSelect can reduce 18%-51% training rounds than the other heuristics on FashionMNIST and CIFAR-10 with non-IID data distributions.

Deep Neural Networks (DNNs) are widely used to analyze the abundance of data collected by massive Internet-of-Thing (IoT) devices. The traditional approaches usually send the data to the cloud and process the DNN inference on the powerful cloud servers but suffer from long network latency. Therefore, edge computing has emerged to reduce network latency by offloading the computation from the cloud to the edge. However, a single resource-constrained edge device is unable to process real-time DNN inference. Thus, we devise a collaborative edge computing system CoopAI to distribute DNN inference over several edge devices with a novel model partition technique to allow the edge devices to prefetch the required data in advance to compute the inference cooperatively in parallel without exchanging data. Subsequently, we present a new optimization problem to minimize the completion time of distributed DNN inference. An innovative algorithm is then proposed to intelligently partition the model into the proper number and sizes of blocks, deploy them on a suitable number of edge devices, and run them in different rounds. The numerical results manifest that our algorithm outperforms the traditional approach by 20%−30% on the completion time.

Decentralized learning (DL) allows IoT devices to exchange local model updates with only their neighboring devices instead of sending their model updates to a central server for aggregation. However, current DL frameworks cannot support the emerging Social IoT(SIoT) paradigm since SIoT devices exchange model updates with only social neighbors based on specific social relations (e.g., ownership and parental relationships). Conversely, sharing model updates with non-social neighbors can improve training performance but may violate social relations. Differential privacy (DP) is thus engaged with DL to ensure data security, while excessive devices engaging DP may downgrade the training performance. However, most research neglects the effect of neighbor selection for each device based on social networks, physical networks, and DP. Therefore, in this paper, we explore the non-trivial relation among the above factors to present a DL framework, DeepPrivacy, and prove its convergence rate and DP. Then, we formulate a novel optimization problem, CoTOPO, to find an efficient communication topology1for model updates exchange among devices in DL, and propose an algorithm, AutoTag, for CoTOPO. Last, experiment results manifest that DeepPrivacy and AutoTag combined outperform the state of the art in terms of convergence rate and physical training time significantly on CIFAR10 and FMNIST.

Vehicle platooning is a promising technology to bring up significant benefits of improved fuel economy and fewer traffic collisions. However, many security attacks such as beacon message falsification have been exposed, creating grave concerns about maintaining a vehicle platoon stably. This work introduces a robust positioning-based verification scheme, namely PVS, to enhance reliability of vehicle platoon control in vehicular networks. By exploiting geographic and maneuver information from 5G radio-based positioning, PVS can detect whether a vehicle is honest in reporting its location for platoon joining preparation or collision avoidance, with up to 96% accuracy.

Recently, autopilot-like vehicles have become more pervasive. To maintain inter-vehicle distance stably, Cooperative Adaptive Cruise Control (CACC) is then proposed to make each autonomous vehicle exchange its dynamic state with neighboring vehicles via vehicle-to-vehicle (V2V) communication. However, longitudinal platooning via CACC systems alone is not enough to improve traffic throughput since each vehicle has its own desired speed and only considers itself to optimize its traveling. Therefore, we develop a novel framework MANA to determine the suitable platoon-merge maneuver, lane-change maneuver, and space-reserve maneuver. Extensive simulation results manifest that MANA can avoid collisions effectively, converge fast, and save fuel consumption and CO2emissions by 24% and 20%.

In a shared software-defined network (SDN), the controller should provide isolation guarantees across flows (users) to predict network performance and minimize disruption from some malicious flows. In an SDN, packets are forwarded by flow rules installed in flow tables, and the capacity of flow tables is usually limited by power and cost constraints so that a limited number of flows can be accommodated. To date, OpenFlow 1.4.0 introduces the flow rule replacement, which allows replacing existing flow rules with new ones once the flow table is full. This is called flow table overflow. Although flow table overflow may lead to an increase in packet delay, our experiments on an SDN testbed show that the network performance could benefit by admitting more flows through slightly overbooking the flow table resource. In this paper, we address the Flow table Overbooking isoLation guArantees problem (FOLA), which aims to maximize minimum progress of flows and minimize maximum flow table overflow. To that end, an algorithm with guaranteed minimum progress and bounded maximum flow table overflow is proposed. Trace-driven experiments on an SDN testbed show that our solution outperforms state-of-the-art methods for maximizing minimum progress in terms of the minimum progress and network throughput.

With growing concerns about privacy issues of machine learning, collaborative learning (CL) is developed to offer on-device training. However, adversarial behaviors of model inversion (MI) are undermining privacy of training data. Specifically, adversaries act as ordinary participants in CL and reproduce private data of a class in training data by training generative adversarial networks (GAN) on the fly, unknowingly. To this end, we design a novel model partition defense, PAMPAS, over user devices and trustworthy edge server to resist GAN attack, and formulate a new optimization problem, TENSOR, to optimize training time. To address the challenges that come with PAMPAS, we propose an algorithm TESLA that yields the optimal solution. Experiment and simulation results manifest that PAMPAS effectively defend GAN attack and TESLA reduces training time by 50% compared with other solutions.

Data privacy preservation has drawn much attention in emerging machine learning applications. Decentralized learning is thus developed to guarantee data security and get rid of the involvement of parameter server to avoid transmission bottleneck. However, the previous research focuses on data compression and exchange rules of model parameters among smart devices but neglects the interplay between link cardinality and transmission power consumption. To jointly optimize these issues, in this paper, we first formulate a new optimization problem, named GreenDL, prove its hardness, and then propose an approximation algorithm termed CoTRAIN. Experiment and simulation results manifest that CoTRAIN reduces more than 20% power compared with traditional methods without sacrificing the convergence rate.

Inference acceleration has drawn much attention to cope with the real-time requirement of artificial intelligence (AI) applications. To this end, model partition for Deep Neural Networks (DNN) has been proposed to utilize the parallel and distributed computing units. However, the previous works focus on the load balancing among servers but may overlook the interplay between the computing and communication. This issue makes the existing approaches less efficient especially in mobile edge networks at which smart devices usually with limited computing capacity have to offload the tasks via limited bandwidth capacity to nearby servers. In this paper, therefore, we innovate a new system and formulate a new optimization problem, CONVENE, to minimize the completion time of inference for the smart devices with one or more antennas. To explore the intrinsic properties, we first study CONVENE with Single Antenna and derive an algorithm termed THREAD-SA to foster the optimum solution. Then, an extension, THREAD, is proposed to subtly utilize multiple antennas to further reduce completion time. Simulation results manifest that our algorithm outperforms others by 100%.

Data privacy preservation has drawn much attention with emerging machine learning applications. Federated Learning is thus developed to offer decentralized learning on user devices. However, it is difficult to jointly address multiple issues such as device selection, upload scheduling, and payment minimization. To jointly optimize the issues above, we first formulate a new optimization problem, named TRAIN, to minimize the training cost (including incentive payment and upload time) while ensuring the data requirement. We then prove the NP-hardness and propose a 3-approximation algorithm, named DETECT to obtain a near-optimal solution. Simulation results manifest that DETECT reduces the training cost by 50% compared with other traditional methods and achieves high accuracy and short convergence time.

Artificial intelligence (AI) on chips has recently driven the expansion of the Social Internet of Things (SIoT), where a group of SIoT devices with social relations can collaboratively identify and handle local events without the help of remote servers. On the other hand, mobile-edge computing (MEC) is a favorable way to locally process SIoT data for reducing data transmission and computation among SIoT devices and backhaul networks. Nevertheless, the load sharing among SIoT devices, MEC, and remote servers brings about new challenges for the communication and computation tradeoff, cross-layer design in SIoT, and forwarding and aggregation tradeoff. To tackle these issues, we formulate a new optimization problem, namely, SIoT collaborative group and device selection problem (SCGDSP), and prove the NP-hardness. We first explore the intrinsic properties of a fundamental SCGDSP case by finding the optimal collaborative group for each user request. Then, we design an approximation algorithm for the general SCGDSP that first evaluates candidate collaborative groups under different social relations, and then selects the collaborative groups and SIoT devices properly. For scalability, the proposed algorithm also supports dynamic user requests and can be distributionally deployed in massive networks enabling collaborative MEC. Moreover, it also sustains local SIoT services, where the computation only involves SIoT devices and MEC servers. Simulation results demonstrate that effective SIoT and collaborative group selection (ESCGS) can reduce by more than 50% of the total communication and computation costs compared with baseline schemes in the real networks from topology zoo. Moreover, the distributed ESCGS reduces by 87% of the running time with total 16.5-MB message overhead, requiring no more than 0.05-ms transmission delay in a 100-Gb/s backbone network with eight MEC servers, 1000 SIoTs, and 800 monitored locations.

Recently, the notions of partitioning the Deep Neural Network (DNN) model over the multi-level computing units and making a fast inference with the early- inference technique have been proposed to shorten the inference time. Such computing units form a hierarchical mobile network to provide locality-aware computation, and the early-inference technique allows the prediction results to early exit the model with a probability. However, an inadequate model partition and misapply early inference may prolong response time. Previous studies focus on the classifier design for early inference, and thus, the optimal model partition with classifier deployment has not been explored. In this paper, we study DEMAND-OPE to consider response time and throughput. We first design the COLT for the simplified DEMAND-OPE without Optional Exit Points (DEMAND) to carefully balance the computing time and data transfer time. Then, an extension termed COLT- OPE is developed to achieve the lower response time. Simulation results show that our algorithms (COLT- OPE) outperform previous methods by 200%.

Traffic engineering (TE) and efficient network updating have been considered as separate problems in previous SDN research. Traffic engineering mostly focuses on static traffic and does not consider the rerouting overheads to support dynamic traffic. Efficient network updating assumes the new routing is provided by TE and focuses on minimizing only the rerouting overheads, and therefore, the improved new routing with bandwidth consumption similar to the new routing from TE but much lower rerouting overheads has not been explored. In this paper, we explore Multi-tree Low-overhead Multicast Rerouting (MLMR) to jointly solve both problems for SDN multicast. We prove that MLMR is NP-hard and design a new approximation algorithm, named Multicast Rerouting and Update Scheduling Algorithm (MRUSA). Equipped with the notions of deterioration indicator, motivator, and inhibitor, MRUSA provides incremental tree updating and multi-tree update scheduling to address the trade-off between the bandwidth consumption and rerouting overheads. Frequent rerouting due to tiny changes of multicast users can be effectively avoided, because rerouting time for each group can be correctly identified. Simulations and implementation on real SDNs with YouTube traffic manifest that the total cost can be reduced by at least 35% compared with SPT and ST, and the computation time is small for massive SDN.

Live video streaming applications are expected to proliferate very rapidly in the next generation network, i.e., 5G networks. To efficiently serve all users watching the same live video in 5G networks, the multicast technique plays an important role in providing scalable and high-performance services. Software-Defined Networking (SDN), contained in 5G networks, further facilitates the development of multicast techniques due to its flexibility of updating routing rules. However, few of conventional multicast mechanisms in SDN take into account the rule update overhead resulting from handovers of mobile users, which leads to tremendous network overhead. In this paper, we adopt Segment Routing (SR) as the first step to alleviate the rule update overhead, and then consider the rule update cost while maintaining the multicast tree to deal with user handovers. Thus, we first formulate Handover-aware Multicast Tree (HMT) problem and then show that HMT is NP-hard and does not admit any approximation algorithm. We then propose a heuristic algorithm called Mobility Aware Multicast Tree Algorithm (MAMTA). MAMTA takes advantage of user movement prediction to assign a base station that could provide longer service to a user, which leads to infrequent rule updates. Simulation results show that MAMTA significantly outperforms the shortest path tree and Steiner tree algorithms.

With the emergence of multi-view 3D videos, network operators now face a new challenge to resolve the dramatic increase of the network bandwidth required to support all subscribed views (typically 16 or 32) of a video. Recently, Depth-Image-Based Rendering (DIBR) in Computer Vision has been demonstrated as a promising way for efficient multi-view 3D video multicast, because many views can be synthesized in mobile devices and no longer need to be transmitted. Nevertheless, DIBR is computationally intensive and incurs additional power consumption in mobile devices, and unsubscribed views need to be transmitted to mobile users for DIBR. In this paper, therefore, we aim to leverage Mobile Edge Computing (MEC) for DIBR to foster effective and efficient multi-view video multicast. We formulate a new problem, named Multi-view Multicast Synthesis and Delivery (MMSD) and prove the NP-Hardness. We design an approximation algorithm, named Merge Search Forest Algorithm (MSFA), to choose the view to be synthesized and build a multicast topology including a low-cost forest for supporting each subscribed view. Simulation results manifest that MSFA outperforms the existing approaches by 30% to 50% of the total communication and computation cost.

Mobile Edge Computing (MEC) has recently emerged as a primary candidate for big data processing to reduce the latency and jitter. On the other hand, Software-Defined Internet of Things (SD-IoT) has been proposed to effectively and flexibly collect and process big IoT data. Nevertheless, minimizing the energy consumption in SD-IoT with big data processing (e.g., data aggregation and reconstruction) has not been explored before. In this paper, therefore, we explore the sensor data selection and routing problem in SD-IoT with big data processing. Specifically, given 1) a set of sensors, 2) a set of observed locations, 3) the network topology, and 4) the energy consumption model of big data processing and forwarding, we formulate a new optimization problem, named Sensor Data Selection, Processing, and Routing Problem (SDSPRP), to minimize the total energy consumption in SD-IoT. We prove that the emphasized problem is NP-hard and inapproximable within O(log|K|). To solve the problem, we propose an αlog|K|- approximation algorithm, called Energy Efficient Sensor Selection and Routing (ESR), to minimize the energy consumption by jointly considering the sensor selection and the energy consumption in traffic engineering and data processing. The proposed algorithm is evaluated on two real networks, and the results manifest that the energy consumption in SD- IoT can be reduced by more than 56%.

Previous research on SDN traffic engineering mostly focuses on static traffic, whereas dynamic traffic, though more practical, has drawn much less attention. Especially, online SDN multicast that supports IETF dynamic group membership (i.e., any user can join or leave at any time) has not been explored. Different from traditional shortest-path trees (SPT) and graph theoretical Steiner trees (ST), which concentrate on routing one tree at any instant, online SDN multicast traffic engineering is more challenging because it needs to support dynamic group membership and optimize a sequence of correlated trees without the knowledge of future join and leave, whereas the scalability of SDN due to limited TCAM is also crucial. In this paper, therefore, we formulate a new optimization problem, named Online Branch-aware Steiner Tree (OBST), to jointly consider the bandwidth consumption, SDN multicast scalability, and rerouting overhead. We prove that OBST is NP-hard and does not have a |Dmax|1-ε-competitive algorithm for any , where |Dmax| is the largest group size at any time. We design a |Dmax|-competitive algorithm equipped with the notion of the budget, the deposit, and Reference Tree to achieve the tightest bound. The simulations and implementation on real SDNs with YouTube traffic manifest that the total cost can be reduced by at least 25% compared with SPT and ST, and the computation time is small for massive SDN.

Interference mitigation among body area networks (BANs) has been a critical challenge due to their mobility and fully distributed nature. When two or more nearby BANs transmit simultaneously, collisions may occur and body sensors therefore consume more energy for re-transmission. Each BAN thus spends more channel time, which lowers the total network capacity. To reduce energy consumption and better the network capacity, in this paper, we propose a CSMA/CA-based Load Adaptive MAC Protocol for inter-BAN, called LAMP. Based on Markov chain analysis, LAMP can dynamically choose a proper contention window size according to the load in the operating channel and automatically switch to another channel when the current channel is near congested. Simulation results show that LAMP outperforms traditional CSMA/CA protocols in terms of throughput and energy consumption. Meanwhile, it keeps high fairness.

Most scheduling algorithms in the literature for cellular networks are concerned with throughput, fairness, or cost optimization. Recently, however, wireless surveillance in cellular networks has become increasingly important, and more and more institutions and companies have adopted commercial cellular surveillance cameras due to their low installation cost and the wide network coverage. In this paper, therefore, we first explore the resource allocation problem for a multi-camera surveillance system in cellular networks. We minimize the number of allocated resource blocks (RBs) while simultaneously ensuring the coverage requirement for the surveillance system in cellular networks. Specifically, we first describe our system model and then formulate the Camera Set Resource Allocation Problem (CSRAP). Next, we prove that the problem is NP-hard and inapproximable within ln n, where n is the number of surveillance targets. To solve the problem, we propose an approximation algorithm for the general case of CSRAP and then we find the optimal solutions of three deployments of cameras in the Manhattan Street Network to find the intrinsic characteristics of camera selections. The simulation results, based on two real surveillance maps and synthetic datasets, show that the number of allocated RBs can be effectively reduced compared to the existing approach for cellular networks.

Marketing through mobile social networks is convenient, low-cost, and beneficial for small companies seeking to expand their customer numbers. In the literature, many studies address the influence maximization problem, which selects initial consumers (seeds) to spread the product information such that the number of consumers receiving the product information (the influenced consumers) is maximized. However, to date, none of these schemes take the beliefs of other persons that could significantly change the consumer's behavioral intention into account. In this paper, we fill this gap by proposing a new variant of the influence maximization problem, the Budgeted Seed Selection (BSS) problem, which asks for a set of seeds with the total cost not greater than a given budget in a mobile social network such that the total expected behavioral intentions of the consumers influenced by the selected seeds are maximized. In addition, we propose an approximation algorithm for the BSS problem. We also conduct simulations to evaluate the performance of our algorithm using real traces and synthesis data. Experimental results show that our algorithm evaluates an approximately optimal seed set for the BSS problem and outperforms several greedy algorithms.

Software-Defined Networking and Network Function Virtualisation (SDN/NFV) can provide flexible resource allocation to support innovative security solutions in a central manner. To mitigate HTTP DDoS attacks, shuffling-based moving target defense has been regarded as one of the most effective ways by redirecting user traffic among a group of virtualized service functions. However, previous work did not notice that frequent changes of user traffic will significantly intensify the control overhead of SDN. In this paper, therefore, we first model the effectiveness and cost for shuffling in SDN/NFV networking with Multi-Objective Markov Decision Processes to find the optimal tradeoff between the effectiveness and cost. We then propose a cost-effective approximation algorithm with a guarantee performance bound to solve the problem. Simulation and implementation on an experimental SDN/NFV network manifest that, given 100 attackers among 1000 users and 50 virtualized functions of a web service, our algorithm achieves the approximation ratio of 0.68 and imposes only 2.4s rule modification latency for each shuffle.

Network Function Virtualization (NFV) on Software-Defined Networks (SDN) can effectively optimize the allocation of Virtual Network Functions (VNFs) and the routing of network flows simultaneously. Nevertheless, most previous studies on NFV focus on unicast service chains and thereby are not scalable to support a large number of destinations in multicast. On the other hand, the allocation of VNFs has not been supported in the current SDN multicast routing algorithms. In this paper, therefore, we make the first attempt to tackle a new challenging problem for finding a service forest with multiple service trees, where each tree contains multiple VNFs required by each destination. Specifically, we formulate a new optimization, named Service Overlay Forest (SOF), to minimize the total cost of all allocated VNFs and all multicast trees in the forest. We design a new 3ρST-approximation algorithm to solve the problem, where ρSTdenotes the best approximation ratio of the Steiner Tree problem, and the distributed implementation of the algorithm is also presented. Simulation results on real networks for data centers manifest that the proposed algorithm outperforms the existing ones by over 25%. Moreover, the implementation of an experimental SDN with HP OpenFlow switches indicates that SOF can significantly improve the QoE of the Youtube service.

With software-defined network (SDN) and network function virtualization (NFV) techniques, we can embed the service chain consisting of a sequence of virtualized network functions (VNFs), i.e., we can determine the flow path and deploy the VNFs contained in the service chain at any place on the path. In the literature, the methods of service chain embedding bound the number of VNFs at a node, whereas the link capacities are disregarded and the amount of flows is not considered, which could cause serious congestion. In addition, according to our experiment, the process overhead on a computation node is linear to the total amount of flows processed. In this paper, we propose a method of service chain embedding to maximize the total amount of flows while bounding the process overhead of the flows on a node by its computation capability and the total amount of flows on an link by its bandwidth capacity. To our knowledge, our method is the first approximation algorithm of service chain embedding with considering flow in the literature. Simulations show our algorithm has good performance in terms of the total amount of flows.

Using the double-ruling technique, many data replication and retrieval schemes achieve low data retrieval latency. However, none of these schemes are location-free schemes in mobile ad hoc networks (MANETs). In this paper, we propose a zero-knowledge double-ruling-based location-free data replication and retrieval scheme (MobiMark) in MANETs. Our primary idea is to label the grid-like structured landmarks in the network using the landmark-labeling, dynamically designate the node that is nearest to a landmark as the landmark broker, and transmit the consumers' interests (or producers' data) to all horizontal (or vertical) landmark brokers using the double-ruling technique. Simulations show that MobiMark achieves good performance in terms of data retrieval rate and data retrieval latency.

The MapReduce/Hadoop architecture has become very important and effective in cloud systems because many data-intensive applications are usually required to process big data. In such environments, big data is partitioned and stored over several data nodes; thus, the total completion time of a task would be delayed if the maximum access latency among all pairs of a data node and its assigned computation node is not bounded. Moreover, the computation nodes usually need to communicate with each other for aggregating the computation results; therefore, the maximum access latency among all pairs of assigned computation nodes also needs to be bounded. In the literature, it has been proved that the placement problem of computation nodes (virtual machines) to minimize the maximum access latency among all pairs of a data node and its assigned computation node and among all pairs of assigned computation nodes does not admit any approximation algorithm with a factor smaller than two, whereas no approximation algorithms have been proposed so far. In this paper, we first propose a 3-approximation algorithm for resolving the problem. Subsequently, we close the gap by proposing a 2-approximation algorithm, that is, an optimal approximation algorithm, for resolving the problem in the price of higher time complexity. Finally, we conduct simulations for evaluating the performance of our algorithms.

We study the information brokerage schemes in wireless sensor networks, which allow consumers to obtain data from producers by replicating and retrieving data in a certain set of sensors, and propose a novel information brokerage scheme, termed RDRIB. Unlike existing information brokerage schemes, RDRIB guarantees successful data retrieval without using any boundary detection algorithm and the geographic location information acquired by the global positioning system (GPS). In RDRIB, the double-ruling technique is used to replicate and retrieve the data within a constructed virtual boundary, and simulations show that RDRIB has good performance in terms of the replication memory overhead, the replication message overhead, the retrieval message overhead, the retrieval latency, and the construction message overhead.

Because the global positioning system (GPS) consumes a large amount of power and does not work indoors, many GPS-free information brokerage schemes are proposed for wireless sensor networks. Each of them, however, either cannot guarantee successful data retrieval or demands a great deal of message overhead to replicate the data. In this paper, we propose a GPS-free information brokerage scheme, RDRIB, in which the double-ruling technique is used to replicate and retrieve the data. RDRIB guarantees successful data retrieval, and, in addition, simulations show that RDRIB has good performance in terms of the replication message overhead and the construction message overhead.