Systematic analysis on peer-to-peer botnet attack detection

Abstract

"Botnet” refers to a network of compromised machines that the bot master remotely controls to prosecute innumerable malicious activities through a CC server and mis cellaneous slave machines. It is possible to categorize botnets as centralized (CC) or decentralized (P2P). According to their distributed functionality,recently P2P botnets is the most significant risks to network security . In this paper, we sys tematically analyze and compare some very recent peer-to-peer botnet algorithms and methods such as Honeypots, AutoBotCatcher, SDN, and PeerGrep to ascertain the most appropriate one for real-world applications. To perform this comparison, we examine AutuBotCatcher, an algorithm that utilizes the community detection method, Honeypot system, where we focus on the Nepethesis honeypot method. Additionally, the PeerGrep system integrates the PeerGrep algorithm, CART algo rithm, and P2P traffic in SDN to automate and flexibly manage flow entries through machine learning.