An improved HSM packet classification algorithm using pruning trie and space compression

Hierarchical Space Mapping (HSM) [1] is excellent in the search time performance with the complexity of log(n), but it consumes great memory. So we propose a improved HSM(called C-HSM) using pruning trie to make the performance of search time independent of the filter number, and using heuristic to compress the space to save memory. The experiment results show that it can trade off time and space practically and enlarge the applicability.