GPGPU based Hybrid Multi-Pattern Algorithm design for high-speed Intrusion Detection System

Networks expand every day at a large scale and operate at high speeds. People have become increasingly dependent on the Internet to perform various tasks. Networks require a system that protects against intrusion activities. Therefore, Intrusion Detection Systems (IDSs) have been created. Subsequently, IDSs have become an important part of modern networks to detect suspicious attacks. Modern networks expand and operate at high speeds, which is why handling all network traffic without overlooking security issues is difficult for IDS that run on a single processor. In this paper, we proposed a shared memory design called Hybrid Multi-Pattern Algorithm (HMA) as extension of our previous work, multi-core based HMA. The experiments show that GPGPU based HMA performs better than multi-core based HMA. This finding proves that GPGPU-based HMA is robust and can function properly in a high-speed network.