P2P bot detection system based on map reduce

Among the various sorts of malware, botnets are rising briskly as one of the most serious threat against cyber-security. Botnets supply a distributed platform for many prohibited activities like launching distributed denial of service attacks against crucial targets, phishing, malware dissemination, click fraud etc. The characteristic of botnets is that they use command and control channels to keep themselves updated and directed. P2P Botnets are one of the most malevolent threats to the Internet users due to their resiliency against takedown. This paper proposes a scalable model which can be used for botnet detection in P2P network. This proposed system treats network traffic as a data stream, segregating the traffic into two parallel streams. The detection is based on failure traffic, communication traffic and frequency of traffic. The traffic is analyzed during small time window, and the infected hosts are reported immediately.