Deeply proficient honeypotdevelopment strategies for detecting denial of service attacks

Abstract

Honeypots are the computer network environments that help to newlineprotect the network resources and services from unauthorized network newlineaccesses from the intruders. Honeypot establishes a system environment newlinewhich looks like legitimate environment to attract the external attackers newlineand attack the network and their services. By this way, honeypot systems newlinedetect various attacker activities and their motives. This practice is newlinemainly employs to detect the vulnerable security points of entire network. newlineConsequently, the security principles are implemented to resolve the newlinenetwork security failures in future. Particularly, this research work creates newlinevarious honeypot development strategies using Machine Learning (ML) newlineand Deep Learning (DL) techniques. The intelligent honeypot systems newlineare configured to analyze effective real time attacks injected in to the newlinenetworks. Notably, the proposed research work deals with the variants of newlinecrucial Denial of Service (Dos) attacks. For developing intelligent newlinehoneypot systems, the proposed research work proposes three different newlinetypes of honeypot development strategies. Each strategy is implemented newlineusing highly adaptive honeypot services to detect Dos attacks. In newlineaddition, this research work compares the performance of proposed newlinehoneypot strategies with various existing techniques. In first strategy, newlineReinforcement Learning (RL) techniques are used with DL procedures. newlineThe Deep RL (DRL) procedures are used for implementing Network newlineIntrusion Detection (NIDS) by using IDS agents. The configured IDS newlineagents are implemented to build enterprise honeypot systems for newlinedetecting Dos attacks at run time. The second approach is implemented newlinefor enabling Generative Adversarial Networks (GAN) based honeypot newlineprinciples. newline