Virtual network embedding arranges virtual services onto substrate components. The performance of algorithms determines the effectiveness and efficiency a virtualized network, making it critical part virtualization technology. To achieve better performance, algorithm needs to automatically detect status which is complicated changes in time-varying manner, dynamically provide solutions that can best fit current status. However, most existing fail automatic an acceptable running time. In this...

With the rapidly emerging encryption techniques for network traffic, classification of encrypted traffic has increasingly become significantly important in management and security. In this paper, we propose a novel deep neural that combines both convolutional recurrent to improve accuracy results. The is used extract packet features single packet. trained pick out flow based on inputs any three consecutive packets flow. proposed model surpasses existing studies which ask first flow, it...

Penetration testing is a vital practice for identifying and mitigating vulnerabilities in cybersecurity systems, but its manual execution labor-intensive time-consuming. Existing large language model (LLM)-assisted or automated penetration approaches often suffer from inefficiencies, such as lack of contextual understanding excessive, unstructured data generation. This paper presents VulnBot, an framework that leverages LLMs to simulate the collaborative workflow human teams through...

The widespread use of encryption protocols and increasing privacy demands have significantly increased encrypted traffic, creating new challenges for network monitoring threat detection. Current methods struggle with diverse scenarios distinguish between subtle traffic patterns within webpages the same application. To address these challenges, we introduce ANT-ET, an end-to-end multimodal framework designed fine-grained webpage fingerprinting. ANT-ET leverages a transformer to model payload...

Communication networks are becoming increasingly complex toward 6G. Manual management is no longer an option for network operators. Network automation has been widely discussed in the networking community, and it a sensible means to manage communication network. Deep learning models developed enable given operation practices have limitations of lack explainability inapplicability across different and/or settings. To tackle above issues, this article we propose new knowledge-powered framework...

With the rapid development of Internet, more and devices are being connected to making up Internet-of-Things (IoT). The accountability privacy two important but contradictory factors ensure security IoT networks. How provide an accountable anonymous access networks is a challenging task. Since network largely driven by services, in this paper we propose new efficient architecture achieve based on services. In architecture, self-certifying identifier proposed efficiently identify service....

Summary OpenFlow has become the key standard and technology for software defined networking, which been widely adopted in various environments. However, global deployment of encountered several issues, such as increasing number fields complex structure flow entries, making size table switches explosively grows, results hardware implementation difficulty. To this end, paper presents modelling on minimisation storage space proposes a Heuristic Storage Optimisation algorithm Flow Table (H‐SOFT)...

Virtual network embedding (VNE) that instantiates virtualized networks on a substrate infrastructure, is one of the key research problems for virtualization. Most existing VNE approaches, however, focus current virtual request (VNR) and treat all VNRs equally, which disregard long-term impact waste many resources process infeasible (i.e., cannot be embedded completely). To address these problems, proactive algorithm based hierarchical reinforcement learning, VNE-HRL, proposed in this paper....

Logs are used to record runtime states and significant events for a software system. They widely anomaly detection. produced by most of the real-world systems show clear characteristics imbalanced data because number samples in different classes varies sharply. The distribution makes classifier bias toward majority class, so it is difficult learn detect anomalies correctly. Most existing methods log-based detection ignore this important problem, they perform poorly on systems. In paper, we...

In recent years, attacks on the DNS continue to proliferate due lack of security mechanisms. over HTTPS (DoH) is a standard developed for encrypting plaintext protect user privacy, but attackers may use this protocol bypass enterprise firewalls and exfiltrate private data through establishment tunnels. Traditional tunnel detection methods based packet inspection are no longer applicable because encryption. Although popular machine learning widely used, it require expert knowledge extract...

In this paper, a PbD (Programming by Demonstration) experiment platform is setup based on dual arm robot and an motion tracking system. A method optical system implemented tested platform. This uses sensor to capture operator arms then let imitate motion. At the same time, logged, optimized program generated automatically computing device or controller itself. It can speed up programming reduce learning difficulty of greatly just demonstrating action sequences robot, furthermore, it save...

Multicast transmission can effectively reduce the cost of network resources in delivering packets to a group receivers. However, it is difficult adjust path generated multicast spanning tree (MST) if source and members are unchanged. OpenFlow has been key technology realize software defined networking makes flow control feasible. This paper presents an OpenFlow-based solution dynamically MST by configuring paths according requirements loads specific conditions. In addition, we propose...