Peer-To-Peer Networking and Applications最新文献

In Underwater Acoustic Sensor Networks (UASNs), optimizing energy efficiency and minimizing void occurrences in routing is paramount. Due to the energy constraints of sensor nodes, low-power transmission is essential for conserving energy. Previous research highlighted the effectiveness of clustering and routing to enhance energy efficacy in UASNs. Therefore, the clustering and routing processes can be considered as optimization problems that are nondeterministic polynomial-time (NP) hard. These challenges can be tackled through the application of machine learning algorithms and meta-heuristics. In this context, K-means clustering is employed to partition the network into clusters, designating the centroid as an ideal Cluster Head (CH) location. This ensures a one-hop proximity between the CH and cluster members, reducing transmitting power and enhancing network energy efficiency. Subsequently, a potential CH is selected using a marine predator optimization (MPA) algorithm based on the derived multi-objective fitness function. The MPA algorithm not only determines the optimal CH but also moves the elected CH to the K-means centroid location. Consequently, Autonomous Underwater Vehicles (AUVs) are utilized to collect and route packets from the CH to the Base Station (BS), minimizing the occurrence of void nodes and avoiding obstacle collisions. An optimal routing path for AUV is established through a way-point-based navigation scheme to achieve high packet reliability. Additionally, the proposed method (DCARo) dynamically determines the optimal number of clusters using the elbow method, ensuring scalability according to network size. Extensive simulations affirm the superiority of the DCARo across various performance metrics.

Mobile crowdsensing (MCS) receives extensive interest due to enabling many novel applications at lower cost. However in pricing incentive scenes of MCS, utility from mobile users presents complex distributions due to mobility, changes of abilities and resource consumption such as device’s energy and memory, especially for submodular MCS with more general distributions. However, existing works only focus on homogeneous and heterogeneous MCS, whether multi-request pricing scene or single-request pricing scene. To the end, in this paper, we investigate online pricing issues for submodular MCS. Moreover, we apply a multiple-stage budget-limited process and robust mean estimators to design budget-limited pricing incentive for submodular MCS. Extensive simulations demonstrate that our mechanisms outweigh existing benchmarks.

Recently, a few Multi-Criteria Decision Making (MCDM)-based charging scheduling schemes have been proposed. However, these schemes have still connoted the problems from the viewpoint of assigning weights to multi-criteria and exploiting redundant capability of a Mobile Charger (MC). In this paper, we propose an efficient charging scheduling scheme using an integrated FCNP-TOPSIS to solve the above-mentioned problems. The proposed scheme firstly divides the whole network into sub-areas by using the Fuzzy C-Means (FCM) algorithm so as to evenly distribute charging request load into multiple MCs and assign a MC to each sub-area. Next, each MC draws up a charging schedule into on-demand or semi-on-demand charging scheduling scheme according to the MC’s charging capability and the number of charging Request Nodes (cRNs). In charging scheduling, first the Fuzzy Cognitive Network Process (FCNP) assigns the relative weights to multi-criteria to characterize the cRNs and predict the potential-to-be-Bottlenecked Nodes (pBNs). Then the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) selects the most suitable next charging location for on-demand charging scheduling and the proactive charging nodes among the predicted pBNs for semi-on-demand charging scheduling. While drawing up the on-demand charging schedule, the partial charging time at each charging location is calculated considering the weights of multi-criteria by FCNP. Extensive simulation experiments have been conducted to show that the proposed scheme greatly improves the charging and network performance at various performance metrics compared to existing schemes. In special, if the number of nodes is 650, the network lifetime of the proposed scheme is 129.4%, 239.8%, 282.5%, 283.2% and 293.6% longer compared to the FAHP-VWA-TOPSIS, FLCSD, AHP-TOPSIS, OPPC, and NJNP schemes, respectively.

The paradigm of sensor networks involves connecting wireless electronic devices through small sensor nodes to gather and sense surrounding information. As these networks have limited resources, it is crucial to optimize their usage to enhance network performance. To achieve this, software-defined network technology is integrated into wireless sensor networks to efficiently utilize network resources. Furthermore, optimized clustering and energy-aware routing techniques are employed to evenly distribute network traffic and enable energy-efficient data transmission in SDN-enabled WSNs. However, the issue of hotspots or energy holes consistently persists in cluster-based routing protocols. This research aims to develop an energy-aware routing protocol incorporating a mobile sink, aiming to achieve energy consumption equilibrium and extend the network's lifespan. To address these concerns and ensure the longer sustainability of SDN-enabled WSNs, a novel approach called mobile sink-based energy-aware routing is proposed for energy-efficient data delivery. It utilizes optimized sink mobility to resolve the hotspot issue based on a proposed fitness function and centroid point approach. The fitness function considers essential parameters such as energy, control node density, and distances from control nodes. The flow rules are also generated based on the rank-based tree topology for multi-hop data transmission. The proposed approach is executed with an ONOS controller to implement SDN policies, and the performance of the heterogeneous network is evaluated through simulation using the ns-3 simulator. Furthermore, the proposed MS-EAR demonstrates significant improvements in the network lifespan compared to existing techniques such as GM-WOA, GMPSO, and FJAPSO, with increases of (18.0mathbf{%},47.5mathbf{%}) and (94.0mathbf{%}), respectively. It also outperforms the current state-of-the-art by considering various performance metrics, including stability period, number of alive nodes, network residual energy, packets transmitted to the control server, and average delay.

Unmanned Aerial vehicles (UAV) are high-speed moving machines that attained rapid growth in various activities and are considered an integral component in the Satellite-Air -Ground-Sea (SAGS) incorporated network. However, UAVs are affected by communication delays and malicious attacks. Therefore, an adequate and secure communication routing and attack detection model is necessary for UAV communication networks. This research described a novel approach for initiating secure communication in UAV networks namely Gannet Walruses Optimization Algorithm + Sheppard Convolutional Spinal Network (GWOA + ShCSpinalNet). Initially, the UAV network is simulated, and the data packets are transmitted among the nodes using optimal routing paths. An optimal routing path is computed using the Gannet Walruses Optimization Algorithm (GWOA) by considering some multi-objective functions through the Deep Recurrent Neural Network (DRNN). The developed GWAO integrates Gannet Optimization (GOA) and Walruses Optimization (WaOA). The data communication is done through monitoring agents. The newly devised Sheppard Convolutional Spinal Network (ShCSpinalNet) is utilized as a decision-making agent for malicious attack detection. The attributes considered for decision-making are round trip time, packet delivery ratio, the strength of the signal, the size of the packet, and the number of incoming packets. Once the SpinalNet categorizes the normal and attacked nodes the defense agent is implemented for attack migration. The ShCSpinalNet is devised by the combination of the Sheppard Convolutional Neural Network and Spinal Network. The GWOA + ShCSpinalNet accomplishes a diminished delay of 0.614 s, an increased detection rate of 0.930%, an energy of 0.439 J, and a Packet Delivery Ratio (PDR) of 0.749.

Double auction provides a cost-effective manner for sellers/buyers in smart grid. Due to concerns about information leakage, the asks/bids from sellers/buyers are sealed, making it challenging to select potential winners. To address this problem, the concept of public key encryption with equality test is deployed in double auction, since it is able to perform information retrieval over secure asks/bids. However, previous solutions suffer from the following two limitations: (i) unable to check inconsistent secure asks/bids due to the lack of tester-verifiable mechanism; (ii) incurring high matching time costs caused by one-to-one secure asks/bids. Therefore, we propose the VerDA, a secure double auction retrieval system with verifiable equality retrieval towards multiple secure asks/bids. Technically, to achieve the property of consistency over secure asks/bids, we develop the tester-verifiable technology by combining the decryption module and test module. To improve the efficiency of retrieval, we introduce secure multi-asks/bids testing function by augmenting the number of inputs in a same retrieval process. Moreover, we implement VerDA based on the PJM dataset in real cloud environment, where the experimental results show practical performance with encryption and test costs amounting to only 57.4% and 18.7% compared to state-of-the-art solution.

Delay tolerant networks (DTNs) are a kind of sporadically connected mobile networks in which the network is intermittent, and end-to-end path is hard to establish. However, as devices in DTNs may often have limited energy and buffer, the network performance will be inevitably affected, especially in our application domain, i.e. the post-disaster scenario. Thus, to start with, we present an appropriate energy and buffer efficient routing protocol (EBRout) for efficient message transmission over a smartphone based DTN. Due to limited battery and storage capacity in mobile devices, a major problem in DTNs is to convince forwarder nodes to participate in forwarding messages. Thus, for improving cooperation among the nodes, an incentivizing scheme is proposed which works in two steps. As the first step, we propose an optimization model to find the minimum incentive. Next, we propose a blockchain-based incentive allocation model that uses Ethereum platform built on top of a DTN-Blockchain integrated environment. The use of blockchain helps to create an immutable and globally accessible record for incentive allocation. The performance of the entire scheme is estimated through extensive simulation in ONE simulator, Python PuLP and Ethereum platform. Performance analyses indicate that the average incentive paid using our proposed optimization model is much lower than the average incentive paid without using the optimization model. Also, the results substantiate the efficiency of the proposed scheme over the competing schemes, in terms of delivery ratio, energy and message overhead without negotiating the blockchain performance in terms of processing time and gas consumption.

Ethereum is a blockchain network that allows developers to create smart contracts and programs that run on the blockchain. Smart contracts contain logic to transfer assets based on pre-defined conditions. With over 100,000 new smart contracts being deployed every day, the potential for coding errors is high, making the contracts vulnerable to exploits. A key limitation is that once deployed, smart contracts are immutable and cannot be updated, even if flaws are found. This inflexibility puts funds at risk of theft and loss. The rapid pace of deployment outpaces security audits, increasing vulnerabilities that put users’ cryptocurrency at risk. To reduce the risk caused by smart contract vulnerabilities, we applied deep learning techniques. To develop a deep learning model capable of detecting vulnerabilities, we first created a dataset by replaying real transactions on the Ethereum Mainnet, collecting opcode sequences from real Ethereum contracts, and labeling them using the SODA plugin. We pre-processed this opcode data by removing duplicates, normalizing sequence lengths, simplifying opcodes into representative groups, and converting sequences into numerical vectors to ultimately obtain an optimal representation of the data. We then trained and evaluated three different neural network architectures on this dataset. Our best-performing model achieved an average accuracy of 88% in detecting seven types of vulnerabilities. Further analysis showed that the model was effective at identifying potential problems in smart contracts, which was an important capability for securing funds and executing logic in live contracts.

While it is well known that privacy-preserving cox regression generally consists of a semi-honest cloud service provider (CSP) who performs curious-but-honest computations on ciphertexts to train the cox model. No one can verify the behaviors of CSP when he performs computations dishonestly in reality. Focusing on this problem, we propose a verifiable privacy-preserving cox regression algorithm tailored with the semi-malicious CSP, where all his behaviors are recorded on a witness tape fulfilling the requirement of transparency. To be specific, a multi-key fully homomorphic encryption (FHE) is used to protect the information of different data owners. The verifiability of our proposed multi-key homomorphic message authenticator (HMAC) ensures CSP sends correct results back to data owners. Furthermore, the compactness of FHE and succinctness of HMAC both under multi keys make the cox regression scheme more feasible. The efficiency of our proposed cox regression scheme is also proved by both theoretical analyses and experimental evaluations. After 21 iterations, it costs no more than 10 min to evaluate our cox regression scheme.

Resilient of routing processes is one of the biggest challenges for data transmission in mobile networks without infrastructure. Communication under current routing protocols is through a communication path that, although the shortest, may not perform satisfactorily in terms of resilient. Routing and communication within such a path may take place using nodes that are malicious or inappropriate in the communication process due to malicious or poor technical state. This paper presents a new algorithm for various uses of mobile ad hoc networks not only in edge networks with infrastructure but also with the possibility of being used in the cloud solutions. We have modified decentralized blockchain technology and artificial intelligence using deep learning methods that have been implemented in routing processes. The objective of this algorithm was to select the most resilient communication path from the source to the destination node. Such a communication path selection consisted of selecting the nodes that were most suitable in terms of resilience, where the selection nodes was provided through a network and technical parameters. The key quality of service metrics, throughput, total delay, number of delivered signaling and data packets and the ratio between them were used to evaluate the proposed resilient routing algorithm. Modified resilient routing protocols achieved improvement in all the analyzed parameters compared to the original routing protocols. The improvement in these parameters led to an increase in the resilience of the routing process based on the actual data obtained from each node in the network and previous communications.