A Network Intrusion Detection System for Memory Architecture

Abstract

A network intrusion detection system is used to monitor network traffic for security threats by scanning packet payloads flowing through the network. The traditional software alone pattern matching approaches used for network security cannot meet the high throughput of today’s networking. So to manage the increasing number of attack patterns and to meet the throughput requirements a successful network intrusion detection system must have a memory-efficient pattern-matching algorithm and hardware design. A Multiple string matching technique is used here and it can compare hundreds of string patterns simultaneously. The aho-Corasick algorithms used in nowadays can process only one input character at a time. But the newly proposed method is achieving 21% of memory reduction compared with Aho-Corasick algorithm. In addition to that a 24% reduction in memory can be achieved by integrating this approach to the bit split algorithm. Filtering and exact matching technique is also adopted here The main advantages of this technique are easy reconfigurability and scalability.