A detection mechanism of DoS attack using adaptive NSA algorithm in cloud environment

Security of any distributed system is not only complex in nature, it also needs much more attention as most of the applications being used and developed in recent past are on distributed platform. Denial of Service (DoS) attack causes drop in quality of service and may also reach to entire absence of service for some `real´ users. Identifying some users as attackers also need appropriate algorithm. Negative selection algorithm (NSA) is a very effective approach in identifying some user as attacker. However declaring some `real´ user as an attacker is a very common limitation of these types of algorithms unless and until the mechanism of detection is updated at regular intervals. In this research work we have modified NSA algorithm to take into account the necessity of updating the detector set from time to time. We have introduced a second detection module to accommodate the updation. Both the algorithms are implemented on common data set and comparative study is presented. Our proposed algorithm comes out with much improved results and significantly reduces false positive (false alarm) cases.