Efficient and robust hash table for connection management in high-speed networks

The continual growth of network traffic rates has driven an increasing demand that hash table for network connection management must perform at fast speeds for real-time packet processing. This paper proposes an efficient and robust hash table, PRH-MTF (Pseudo Random Hashing-Move to Front), to maintain large numbers of simultaneous connections in high-speed networks. A hash function PRH is designed by computing connection identifier with selected operators in consideration of algorithmic complexity attacks. To resolve hash collision efficiently, we apply the MTF heuristic to improve the classical chaining resolution in virtue of network traffic locality. The proposed PRH-MTF hash table is employed to implement connection management, and its average search length is analyzed with packet train model as packet arrival pattern. Finally, we evaluate the performance of the PRH-MTF hash table under diverse conditions with real high-speed network traffic. Experimental results indicate that the PRH-MTF hash table performs better than the traditional IPX-SC one in terms of efficiency and robustness. © 2012 Praise Worthy Prize S.r.l. - All rights reserved.