Publications

Abstract : Network Intrusion Detection Systems (NIDS) have become integral to today’s computer networks as the information transferred through the network is highly vulnerable to cyber-attacks. The implementation of the same involves an efficient string matching algorithm which will compare certain parameters and features of the received packets depending on the protocols used for communication. As the system incorporates a huge number of pattern comparisons, the string matching algorithm stands as the heart of entire NIDS system. Software implementations such as SNORT are inadequate for Ethernet backbones with bit rates of hundreds of Gigabits where high throughputs are required. Small improvements at the algorithmic level can boost the performance of the system drastically. In order to handle the large payloads of current network packets, simple and commonly used algorithms such as Aho-Corasick lack performance. In this work, the Z algorithm which is generally used for finding large subsequences in DNA strands is adapted for string matching for intrusion detection. This new algorithm shows noticeable improvements in performance and scalability. © Research India Publications.