Region based multi attribute traffic impact based efficient low rate DDoS attack detection

Abstract

Modern users access internet technology for many day to day activities. Information technology has seen a big growth in this decade providing various location independent services. Such services have no newlineconstraint in accessing through the internet. The user can access those newlineservices from his location and perform his activity. However, the service newlineprovided and the providers have certain limitations. However, there is a limit newlineon the strength and the volume of data that can be handled by the server. The newlineusers are restricted and informed to access the services in a proper manner newlinewith limited data in a protocol. But humans are not trustworthy and even a newlinegenuine registered user can perform malicious activity. When the number of newlineservice request or the volume of data grows to more than a limit allowed, it newlinebecomes a denial of service attack. The generic denial of service attack is newlineperformed by a single node or group of nodes which are distributed newlinethroughout the network. It is not all the time that the malicious nodes are involved in denial of service attack in a distributed manner. Even a single genuine user would newlineperform denial of service attack in limited manner to be named as low rate newlineDistributed Denial of Service (DDoS) attack. Unlike a generic DDOS attack, newlinelow rate attacks are appear in a very limited manner but produce a huge newlineimpact in the network performance. In the case of connection based flooding, newlinethe denial of service detection is performed on the basis of on a connection newlinethreshold. It is denied for the service access only when the malicious user newlinecrosses specified limits but if he knows the limit, he can produce such newlineflooding up to the limit of threshold. The same can be performed in the case newlineof data flooding. Identifying such low rate attacks is a challenging issue in newlinemodern days. newline newline