To address the issues of insufficient collaboration among computing resources and poor adaptability to task requirements in computing power networks, the computing power routing problem was modeled as a sequential decision problem. A deep reinforcement learning-based computing-aware routing algorithm was proposed for dynamic routing scheduling of computing network collaboration. The idea of hybrid expert models was drawn on and a differentiated expert network was designed based on an encoder-decoder structure for specialized optimization in three typical scenarios: delay-sensitive, ordinary, and computationally intensive. The routing selection space was constrained through an action masking mechanism to achieve efficient hop-by-hop decision-making and output a path containing the optimal computing node. The simulation experiment results show that compared with other routing scheduling algorithms, the proposed algorithm improves service success rate by about 17%, reduces end-to-end latency, optimizes load balancing between nodes, demonstrates good network topology adaptability, and can effectively meet the differentiated needs of diverse computing tasks.

This letter devises an AI-Inverter that pilots the use of a physics-informed neural network (PINN) to enable AI-based electromagnetic transient simulations (EMT) of grid-forming inverters. The contributions are threefold: (1) A PINN-enabled AI-Inverter is formulated; (2) An enhanced learning strategy, balanced-adaptive PINN, is devised; (3) extensive validations and comparative analysis of the accuracy and efficiency of AI-Inverter are made to show its superiority over the classical electromagnetic transient programs (EMTP).

In order to protect the data privacy of network users and solve the training difficulties caused by traffic distribution, this paper based on federal semi-supervised learning presents a traffic classification method to solve the problem of a small number of labeled traffic distributed in server, and a large number of non-labeled traffic distributed independently and identically in clients and not shared. On the one hand, this paper adopts the parameter decomposition strategy to avoid interference between different tasks. On the other hand, this paper uses consistency regularization between clients to maximize consensus between similar segment clients to solve the learning problem of variable small sample data. In addition, method in this paper only transfer parameter differences during the federated learning parameter transfer process. The experimental results show that the accuracy gap between our method and the supervised learning training method is minimal, which can effectively protect user privacy and does not require a large amount of labeled data and communication costs.

Traffic Engineering (TE) is applied to optimize network transmission efficiency by managing the routing of complicated traffic. Emerging Deep Reinforcement Learning (DRL) and Software-Defined Networking (SDN) provide flexible ability for traffic management and congestion control. However, existing methods either cannot reroute the network-wide traffic in an accurate way or encounter too much calculation latency due to the inference of mathematical optimization techniques. In this paper, we propose a flexible and link-reconfigurable TE solution called EfficientTE that effectively adjusts the traffic routing based on real-time traffic demands. By analyzing the characteristics of network topology and traffic, EfficientTE selects a few links that are critical for congestion in the network. Then, we propose the idea of virtual capacity that helps the DRL algorithm adjust to different link bandwidths. Based on the traffic demand and topology information collected by the SDN controller, the DRL algorithm is used to dynamically adjust the virtual capacity of the critical links to reshape the network. To ensure network performance with low disturbance, we selectively reroute the Top-K critical flows using the weighted K-shortest path algorithm, while forwarding the major flows with default rerouting. Experiments show that EfficientTE optimizes maximum link utilization and outperforms existing TE solutions by improving the load-balancing performance ratio by at least 6.13%, 18.78%, 16.20%, and 22.81% respectively in four network topologies.

Active defense is currently a key technology to reverse the asymmetry of offensive and defense in cyberspace. Honeypots, as one of the active defense technologies, are deployed in the internal network to attract attackers’ attacks, consume the attackers’ energy, and discover the attackers’ attack methods and attack intent. But at the same time, if there are security loopholes in the honeypot itself, the honeypot becomes the entrance for the attacker to attack the intranet, which will harm the security of the intranet. In this paper, aiming at the virtual machine escape scenario in the honeypot system, based on the mimicry defense idea, a dual mimicry mechanism and the honeypot architecture under this mechanism are proposed. This mechanism uses the heterogeneity of the underlying virtualization platform to resist the escaping vulnerabilities of the virtualization platform, and achieves level heterogeneity through honeypots to attract attackers, which is conducive to the complete collection of attackers' attack behaviors. Finally, the security test and performance test were carried out through the web implementation of the mimic honeypot.

Routing optimization has long been a problem in the networking field. With the rapid development of user applications, network traffic is continuously increasing in dynamicity, making optimization of the routing problem NP-hard. Traditional routing algorithms cannot ensure both accuracy and efficiency. Deep reinforcement learning (DRL) has recently shown great potential in solving networking problems. However, existing DRL-based routing solutions cannot process the graph-like information in the network topology and do not generalize well when the topology changes. In this paper, we propose AutoGNN, which combines a GNN and DRL for the automatic generation of routing policies. In AutoGNN, the traffic distribution in the network topology is processed by a GNN, while a DRL framework is used to train the parameters of neural networks without human expertise. Our experimental results show that AutoGNN can improve the average end-to-end delay of the network by up to 19.7% as well as present more robustness against topology changes.

Honeypot is a proactive defence technology introduced by the defence to change the asymmetry of the network offensive and defence game. Through deploying security resources without any production purpose, the defenders can deceive attackers to illegally use them to capture and analyse the attack behaviour to understand the attack tools, methods, speculate the attack intentions and motivations. This paper introduces the development process and the latest development direction of honeypot technology, comprehensively analyses the application situation of artificial intelligence technology in the key mechanisms of honeypot technology, and summarises the latest application of artificial intelligence technology in the deception environment construction, threat data capture and threat data analysis. Finally, the application of artificial intelligence technology in honeypot technology is discussed.

With the development of modern browsing, the convenience brought by rich browser features has also produced a large number of features, which are called browser fingerprints. This article surveys the latest research results on browser fingerprinting, hoping to provide a convenient navigation for newcomers to research or apply this technology in the future. This paper first briefly introduces the browser fingerprinting technology itself, then classifies the related research on browsers, and analyzes the development of different research directions of browser fingerprinting in detail. And through the analysis of the existing results, the problems faced by different research directions are pointed out. After that, this paper introduces the application of browser fingerprint technology in detail and discusses the application achievements and technical challenges of this technology. Next, this paper introduces the theoretical tools related to the research of browser fingerprinting technology and introduces the application of different theoretical tools and practical significance. Finally, the research achievements of browser fingerprint recognition are summarized, and the future development trend is pointed out.

Log anomaly detection based on deep learning is one of the research hotspots in the field of computer security. It is foreseeable that the mimicry theory proposed by Academician Wu Jiangxing will further improve the detection capabilities of deep learning models, but will also bring high resource consumption and difficulty in application. Therefore, this paper proposes a mimic model construction method that uses the output of complex models as prior knowledge to train lightweight heterogeneous execution bodies and then integrates them. Finally, it is based on DPCNN and TextCNN as complex models and lightweight executions respectively. The experiment of the body structure mimic model proves that while reducing the number of parameters from millions to thousands, its detection accuracy and F1 value are only about 2% and 4% lower than the original model, which greatly retains the original model. The detection capability.

蜜网作为一种重要的主动防御机制,对增强网络的防护能力有重要意义.但传统蜜网存在配置单一、拓扑固定、伪装能力差等自我隐藏机制不完备、不智能的缺陷.提出一种网络欺骗机制,通过IP地址跳变、针对性配置蜜罐等手段,显著提高了攻击者攻击的难度并增强了蜜网获取攻击信息的能力.同时,利用SDN可编程的特点与灵活的控制机制,使用mininet仿真实验平台搭建了基于网络欺骗机制的SDN蜜网,并通过博弈论证明所提蜜网的有效性.实验结果证明,基于网络欺骗机制的SDN蜜网能够在不降低服务质量的同时有效欺骗攻击者.

Existing artificial intelligence-based routing schemes are limited in generalization performance, and fail to adapt to the topological changes of networks.This paper proposes an intelligent routing strategy named SmartRoute based on deep reinforcement learning.SmartRoute can dynamically adjust the routing strategy by sensing the network traffic distribution in real time.Additionally, it combines the topology information perception ability of graph neural network and the self-training ability of deep reinforcement learning to improve the intelligence of network routing strategy.Experimental results show that SmartRoute saves up to 9.6% of end-to-end delay, and exhibits higher robustness than DRL-TE, TIDE and other schemes.

The rapid development of the Internet has also brought opportunities for some illegal elements. Network attackers steal sensitive information from victims through phishing webpages to obtain economic benefits. Currently, the commonly used detection methods for phishing webpages, based on blacklist detection and webpage content feature detection, have the problems of being unable to detect newly emerging phishing webpages or requiring manual extraction of webpage features. Therefore, researchers have used Convolution Neural Network (CNN) to detect phishing webpages by automatically extracting URL features. However, its method has some limitations: (1) The memory is limited when the URL is transformed into the feature matrix, and the embedding vector of new words cannot be obtained or the effective information of sensitive words is lost; (2) the long-distance dependent feature of the URL cannot be obtained. In response to the above challenges, we proposes a phishing detection method based on CNN and Bi-directional Long Short-Term Memory (Bi-LSTM) based on existing work: based on sensitive word segmentation-- comprehensively using two existing URL segmentation methods before converting URL into eigenvector matrix; adding Bi-LSTM on the basis of convolutional neural network to obtain URL long-distance dependent features. Experimental results show that this method can achieve high accuracy, recall rate and F1 value.