Current Issue

25 June 2021, Volume 42 Issue 6

Previous Issue    Next Issue

Papers

TDMA-based user scheduling policies for federated learning

Meixia TAO, Dong WANG, Rui SUN, Naifu ZHANG

2021, 42(6):  1-29.  doi:10.11959/j.issn.1000-436x.2021056

Asbtract ( 536 )   HTML ( 113)   PDF (2633KB) ( 632 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

To improve the communication efficiency in FL (federated learning), for the scenario with heterogeneous edge user's computing capacity and channel state, a class of time division multiple access (TDMA) based user scheduling policies were proposed for FL.The proposed policies aim to minimize the system delay in each round of model training subject to a given sample size constraint required for computing in each round.In addition, the convergence rate of the proposed scheduling algorithms was analyzed from a theoretical perspective to study the tradeoff between the convergence performance and the total system delay.The selection of the optimal batch size was further analyzed.Simulation results show that the convergence rate of the proposed algorithm is at least 30% higher than all the considered benchmarks.

Efficient routing strategy of blockchain-based payment channel network

Ru HUO, Dong NI, Hua LU, Yunfeng XIA, Shuo WANG, Tao HUANG, Yunjie LIU

2021, 42(6):  30-40.  doi:10.11959/j.issn.1000-436x.2021113

Asbtract ( 656 )   HTML ( 77)   PDF (1161KB) ( 588 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In order to solve the problems of the low transaction success rate and network imbalance of the payment channel network, an efficient routing strategy of blockchain-based payment channel network was proposed.This strategy established a dedicated payment channel for the high-priority services according to the service type and service priority, and divided the conventional business into multiple transaction unit.Furthermore, a channel balanced routing algorithm was designed to route each transaction unit, which could reduce the number of transactions on the blockchain and maintain long-term stable operation of the off-chain payment channel, as well as improve the transaction success rate.In addition, in order to avoid the temporary shortage of funds and unavailability of channels due to a certain link occupied by multiple transactions simultaneously, a transaction queuing mechanism in the payment channel network was designed.This mechanism established the forwarding rules for transactions by calculating the escrow amount between the node that transactions arrived and the next hop node, where the channel balanced routing algorithm was used to calculate the new forwarding path for the nodes that could not carry out capital injection within the queuing threshold.The simulation results show that the proposed strategy could improve the transaction success rate and realize the equilibrium of the payment channel network.

Random routing defense method based on deep deterministic policy gradient

Xiaoyu XU, Hao HU, Hongqi ZHANG, Yuling LIU

2021, 42(6):  41-51.  doi:10.11959/j.issn.1000-436x.2021093

Asbtract ( 354 )   HTML ( 56)   PDF (1017KB) ( 489 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

To solve the problem of the existing routing shuffling defenses, such as too coarse data flow splitting granularity, poor protection effect on legitimate QoS and the security against eavesdropping attacks needed to be improved, a random routing defense method based on DDPG was proposed.INT was used to monitor and obtain the network state in real time, DDPG algorithm was used to generate random routing scheme considering both security and QoS requirements, random routing scheme was implemented with programmable switch under P4 framework to realize real-time routing shuffling with packet level granularity.Experiment results show that compared with other typical routing shuffling defense methods, the security and QoS protection effect of the proposed method are improved.

Spatio-temporal data analysis and accessibility method for IoV in an urban scene

Jiujun CHENG, Guiyuan YUAN, Jie CUI, Aiguo ZHOU, Bo LYU, Guangyao LI

2021, 42(6):  52-61.  doi:10.11959/j.issn.1000-436x.2021110

Asbtract ( 306 )   HTML ( 40)   PDF (1431KB) ( 411 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In order to solve the problems of diversity spatio-temporal data and low connectivity efficiency in a single road side unit for Internet of vehicles (IoV) in an urban scene, a spatio-temporal data analysis and accessibility method was presented.First, a spatio-temporal data analysis method based on de-noising and data filling was introduced, and a tensor factor aggregation-based neural network was constructed to predict connectivity intensity among vehicles.Then, a connectivity intensity prediction-based accessibility method was proposed.The simulation results demonstrate that the proposed connectivity intensity prediction method can accurately predict connectivity intensity among vehicles, and the proposed accessibility method can effectively reduce connectivity redundancy and loads of road side units.

Node selection method in federated learning based on deep reinforcement learning

Wenchen HE, Shaoyong GUO, Xuesong QIU, Liandong CHEN, Suxiang ZHANG

2021, 42(6):  62-71.  doi:10.11959/j.issn.1000-436x.2021111

Asbtract ( 1168 )   HTML ( 142)   PDF (1208KB) ( 875 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

To cope with the impact of different device computing capabilities and non-independent uniformly distributed data on federated learning performance, and to efficiently schedule terminal devices to complete model aggregation, a method of node selection based on deep reinforcement learning was proposed.It considered training quality and efficiency of heterogeneous terminal devices, and filtrate malicious nodes to guarantee higher model accuracy and shorter training delay of federated learning.Firstly, according to characteristics of model distributed training in federated learning, a node selection system model based on deep reinforcement learning was constructed.Secondly, considering such factors as device training delay, model transmission delay and accuracy, an optimization model of accuracy for node selection was proposed.Finally, the problem model was constructed as a Markov decision process and a node selection algorithm based on distributed proximal strategy optimization was designed to obtain a reasonable set of devices before each training iteration to complete model aggregation.Simulation results demonstrate that the proposed method significantly improves the accuracy and training speed of federated learning, and its convergence and robustness are also well.

Key node identification algorithm for complex network based on improved grey wolf optimization

Qiuyang GU, Bao WU, Zhaoyang SUN, Renyong CHI

2021, 42(6):  72-83.  doi:10.11959/j.issn.1000-436x.2021088

Asbtract ( 318 )   HTML ( 56)   PDF (984KB) ( 393 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In recent years, how to select the most influential key node for identification has become the most cutting-edge hot direction in network science.Formulating the problem of maximizing the influence of complex network nodes as an optimization problem whose cost function was expressed as the influence of nodes and the distance between them, measures user influence using Shannon entropy, and solved this problem using an improved gray wolf optimization algorithm.Finally, numerical examples were performed with real complex network datasets.The experimental results show that the proposed algorithm is more accurate and computationally efficient than the existing method.

Reduction algorithm based on supervised discriminant projection for network security data

Fangfang GUO, Hongwu LYU, Weilin REN, Ruini WANG

2021, 42(6):  84-93.  doi:10.11959/j.issn.1000-436x.2021117

Asbtract ( 211 )   HTML ( 25)   PDF (1424KB) ( 377 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In response to the problem that for dimensionality reduction, traditional manifold learning algorithm did not consider the raw data category information, and the degree of clustering was generally at a low level, a manifold learning dimensionality reduction algorithm with supervised discriminant projection (SDP) was proposed to improve the dimensionality reduction effects of network security data.On the basis of the nearest neighbor matrix, the label information of the raw data category was exploited to construct a supervised discriminant matrix in order to translate unsupervised popular learning into supervised learning.The target was to find a low dimensional projective space with both maximum global divergence matrix and minimum local divergence matrix, ensuring that the same kind of data was concentrated and heterogeneous data was scattered after dimensionality reduction projection.The experimental results show that the SDP algorithm, compared with the traditional dimensionality reduction algorithms, can effectively remove redundant data with low time complexity.Meanwhile the data after dimensionality reduction is more concentrated, and the heterogeneous samples are more dispersed, suitable for the actual network security data analysis model.

Partial interference elimination based retrospective interference alignment scheme in the downlink MIMO broadcast channel

Jingfu LI, Wenjiang FENG, Wenshou WANG, Weiheng JIANG, Chonghai YANG

2021, 42(6):  94-106.  doi:10.11959/j.issn.1000-436x.2021092

Asbtract ( 265 )   HTML ( 24)   PDF (939KB) ( 486 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

For interference networks, the degree of delay (DoD) was proposed to quantify the information delay among traditional interference alignment (IA) schemes.It indicated that networks suffer from the issue of latency caused by IA schemes.Herein, for the MIMO downlink broadcast channel (BC), the partial interference elimination based retrospective interference alignment scheme (PIE-RIA) was put forward.In the scheme, relay technique was adopted to eliminate partial interference signals so that part of desired symbols could be decoded promptly in user sides.Meanwhile, in the last slot, the eliminated signals could be regenerated in base station side with the retrospective interference alignment (RIA) scheme .The simulation results show that PIE-RIA scheme maintains degree of freedom (DoF) gain of RIA scheme and in the meantime, achieves lower DoD than the other schemes.

Approximation method of multiple consistency constraint under differential privacy

Jianping CAI, Ximeng LIU, Jinbo XIONG, Zuobin YING, Yingjie WU

2021, 42(6):  107-117.  doi:10.11959/j.issn.1000-436x.2021122

Asbtract ( 614 )   HTML ( 34)   PDF (937KB) ( 499 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Under differential privacy, to solve the optimal publishing problem with multiple consistency constraints, an approximation method of multiple consistency constraints was proposed by the theoretical analysis of the principle of optimal consistency release.The main idea was to divide the consistency constraint problem into several consistency constraint sub-problems and then achieve the original problem's optimal consistency release by solving each consistency constraint sub-problem repeatedly and independently.The advantage was that after the consistency constraint problem divided, the sub-problems were often easier to solve, or the technology to achieve optimal and consistent release of sub-problems is quite mature.Therefore more complex differential privacy optimal release problem could be solved.After analysis, the approximation method's convergence was fully demonstrated, ensuring that any partition of consistency constrained sub-problems can always achieve the optimal consistency release of the original problem.Furthermore, taking the sales histogram publishing as an example, based on the approximation method of multiple consistency constraints, a parallel algorithm was designed with optimal consistency release under differential privacy.The experimental results show that the algorithm's efficiency is 400 times higher than that of the general solution, and the algorithm can process millions of large-scale data.

Cooperative service caching and peer offloading in Internet of vehicles based on multi-agent meta-reinforcement learning

Zhaolong NING, Kaiyuan ZHANG, Xiaojie WANG, Lei GUO

2021, 42(6):  118-130.  doi:10.11959/j.issn.1000-436x.2021104

Asbtract ( 690 )   HTML ( 90)   PDF (1302KB) ( 579 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In order to reduce computation complexity, a two-layer mutli-RSU (road side unit) service caching and peer offloading algorithm (MPO) was proposed to decouple the optimization problem.In the designed MPO, the outer layer utilized multi-agent meta-reinforcement learning, which established long short-term memory (LSTM) network as the meta-agent to balance decisions of local agents and accelerate learning progress.The inner layer utilized lagrange multiplier method to achieve optimal decision for RSU peer offloading.Experimental results based on real traffic data in Hangzhou demonstrate that the proposed method outperforms other methods and remains robust under large-scale workloads.

Software-defined network packet forwarding verification scheme based on attribute-based signatures identification

Chaowen CHANG, Jianshu JIN, Peisheng HAN, Xianwei ZHU

2021, 42(6):  131-144.  doi:10.11959/j.issn.1000-436x.2021079

Asbtract ( 259 )   HTML ( 34)   PDF (1232KB) ( 712 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Aiming at the lack of effective forwarding verification mechanism for packet in software defined network (SDN), a data packet forwarding verification scheme based on attributed-based signatures identification was proposed.First, the attribute signature identification was generated according to the user's identity attribute, and the data packet was marked by the attribute signature identification.Then, the P4 forwarding device was used to control accurately and sample the data packet.The controller verified the attribute signature of the sampled data packet.The OpenFlow forwarding device processes the abnormal data packets according to the flow table issued by the controller.Finally, a multi-controllers architecture was constructed to avoid the single point failure of the controller.The results of the experiment indicate that the scheme can achieve accurate control and sampling of data packet, effectively detect the forwarding abnormal behaviors such as packet tampering and forgery, and the network delay is within the range of feasible communication delay.

Distributed storage causal consistency model with trusted constraint

Junfeng TIAN, Juntao ZHANG, Yanbiao WANG

2021, 42(6):  145-157.  doi:10.11959/j.issn.1000-436x.2021091

Asbtract ( 237 )   HTML ( 22)   PDF (899KB) ( 376 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

At present, there are few mature solutions to consider security risks in the research field of distributed storage causal consistency.On the basis of hybrid logic clock and HashGraph, combined with trusted cloud alliance technology in trusted cloud platform, a distributed storage causal consistency model (CCT model) with trust constraints was proposed.The CCT model designed identity authentication and consistent data trust verification mechanism on the client side and the server side respectively, and imposed security constraints on the process of data synchronization between data replicas in the cloud storage cluster.Through the simulation experiment, CCT model can identify and verify the identity signature forgery, illegal third party and other security risks in the client and server, and provide the trusted constraint for the system on the premise of causing small performance cost.

Optimal transmission strategy of diamond channel with limited battery capacity of source node

Taoshen LI, Li SUN, Zhe WANG

2021, 42(6):  158-170.  doi:10.11959/j.issn.1000-436x.2021069

Asbtract ( 147 )   HTML ( 14)   PDF (968KB) ( 364 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In order to solve the problem that the energy harvesting uncertainty of node in wireless energy harvesting network makes it difficult for the system to design a reasonable transmission strategy, a diamond channel with limited battery capacity of the source node were constructed.The goal was to maximize the end-to-end throughput of the system.Under the conditions of meeting the constraints of energy and data causality between transmission nodes and no battery energy overflow, a model to maximize the end-to-end throughput of the system was built, and an optimal offline transmission power and rate allocation strategy was proposed.Firstly, based on theoretical analysis and formula derivation, the strategy transformed the optimal transmission problem into the optimal transmission problem on the side of the broadcast channel, and used the water-filling algorithm to solve the optimal total power from the source node to the relay node.Then, the throughput of the source node to each relay node was analyzed based on the cutoff power level in the broadcast channel throughput maximization problem.Finally, according to the data arrival of the relay node, the iterative water-filling method was used to solve the optimal transmission power from the relay node to the destination node, and the optimal rate was solved by extending the optimal rate of the multiple access channel.Simulation results demonstrate the feasibility and correctness of the proposed optimization policies.

Android application privacy protection mechanism based on virtual machine bytecode injection

Yubo SONG, Qi CHEN, Rui SONG, Aiqun HU

2021, 42(6):  171-181.  doi:10.11959/j.issn.1000-436x.2021115

Asbtract ( 275 )   HTML ( 38)   PDF (864KB) ( 387 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

To solve the abuse of the Android application permission mechanism, a method of Android application access control based on virtual machine bytecode injection technology was proposed.The security policy in the form of virtual machine bytecode was generated according to the user’s security requirement and usage scenario, and injected into the coding unit of Android application that involves dangerous permission request and sensitive data access, to realize dynamic application behavior control.Tests on applications crawled from four mainstream domestic App stores show that the method can effectively intercept sensitive API calls and dangerous permission requests of legitimate App programs and implement control according to pre-specified security policies.Also, after injecting virtual machine bytecode, most of the App program operation is not affected by the injected code, and the robustness is guaranteed.The proposed method has a good universality.

Network security situational awareness model based on threat intelligence

Hongbin ZHANG, Yan YIN, Dongmei ZHAO, Bin LIU

2021, 42(6):  182-194.  doi:10.11959/j.issn.1000-436x.2021106

Asbtract ( 630 )   HTML ( 91)   PDF (1209KB) ( 606 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In order to deal with the problems that the increasing scale of the network in the real environment leads to the continuous high incidence of network attacks, the threat intelligence was applied to situational awareness, and the situational awareness model based on random game was constructed.Threat perception of the target system was performed by comparing the similarity between the exogenous threat intelligence and the internal security events of the system.At the same time, internal threat intelligence was generated based on the threat information inside the system.In this process, game theory was used to quantify the current network security situation of the system, evaluate the security status of the network.Finally, the prediction of the network security situation was realized.The experimental results show that the network security situation awareness method based on threat intelligence can reflect the changes in the network security situation and predict attack behaviors accurately.

Comprehensive Review

Survey on privacy protection in non-aggregated data sharing

Youhuizi LI, Yuyu YIN, Honghao GAO, Yi JIN, Xinheng WANG

2021, 42(6):  195-212.  doi:10.11959/j.issn.1000-436x.2021120

Asbtract ( 662 )   HTML ( 115)   PDF (991KB) ( 921 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Although there is a great value hidden in the massive data, it can also easily expose user privacy.Aiming at efficiently and securely sharing data from multiple parties and avoiding leakage of user private information, the development of related research and technologies on the non-aggregated data sharing field was introduced.Firstly, secure multi-party computing and its technologies were briefly described, including homomorphic encryption, oblivious transfer, secret sharing, etc.Secondly, the federated learning architecture was analyzed from the aspects of source data nodes and transmission optimization.Finally, the existing non-aggregated data sharing frameworks were listed and compared.In addition, the challenges and future potential research directions were summarized, such as complex multi-party scenarios, the balance between optimization and cost, as well as related security risks.

Correspondences

Verifiably secure fast group authentication protocol with anonymous traceability for Internet of vehicles

Haibo ZHANG, Hongwu HUANG, Kaijian LIU, Xiaofan HE

2021, 42(6):  213-225.  doi:10.11959/j.issn.1000-436x.2021073

Asbtract ( 448 )   HTML ( 68)   PDF (1112KB) ( 666 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Identity authentication is the first line of defense for vehicles to access IoV (Internet of vehicles).However, the existing schemes cannot meet the requirements of efficient authentication of IoV well, nor can they realize fast anonymous traceability.In view of this, a bidirectional anonymous traceability group authentication protocol was proposed in IoV.In this protocol, a number of RSU (road side unit) were grouped quickly and dynamically.And the vehicles entering the RSU group were authenticated with the one-way trap gating property and semi-group property of Chebyshev chaotic map.When the vehicle switched between the RSU within a group, the reverse hash chain was used for fast handover authentication.In addition, the RSU within a group can trace anonymously and revoke the identity of malicious vehicles quickly by using blockchain, and can also freely change the ID of users who reveal their real identity.At the same time, the semantic security of the proposed protocol is proved by using the random predictor model.Finally, simulation results show the proposed scheme has good security and effectiveness.

Latency model of neighbor discovery based on Bluetooth low energy 5.0

Bingqing LUO, Peipei WANG, Zhengkang WANG, Zhixin SUN

2021, 42(6):  226-237.  doi:10.11959/j.issn.1000-436x.2021105

Asbtract ( 300 )   HTML ( 47)   PDF (28298KB) ( 151 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

For Bluetooth low energy neighbor discovery protocol, it was usually difficult to assess the discovery latency performance.A latency performance model of neighbor discovery based on Chinese remainder theorem for Bluetooth 5.0 standard was proposed.A mathematical relationship between the discovery latency and key parameters was put forward, including advertising interval, scan interval and scan window.The effect of different parameter configurations on the neighbor discovery latency was verified.Experimental results show that the model can effectively predict the delay peak under the influence of parameter configuration, and provide parameter configuration verification and guidance for different application scenarios.