an infinitely expanding number of network appliances are utilising packet classifiers to fulfil Quality of Service, security, and traffic engineering tasks. Packet classification is an important role of firewalls and routers. Internet ?rewalls, routers and service providers perform different operations at different ?ows. All the packets have to be classified methodically for emerging broadband internet services, and applications such as Internet TV/Radio, gaming, Video on Demand (VoD) and e-businesses, which are in perpetual demand for a higher degree of transmission bandwidth, complex security, and specific Quality of Service (QoS).
This paper presents a comparative evaluation packet classification of algorithms, HiCut, based on a decision tree structure, and the Recursive Dimensional Cutting (DimCut). The comparison has been conducted on operations based on similar principles and design choices. Performances measurements have been obtained by placing the implemented classifiers in the same test conditions scenario. In particular, the comparison aims at achieving a good arrangement between performance, memory usage and flexibility.
The Recursive Dimensional Cutting (DimCut) is the extension of HiCut algorithm with new heuristics ideas, implementing techniques and parallel programming features which classify packets while retaining HiCut’s basic framework. After testing the DimCut algorithm, for classifying packets based on five header ?elds, it is observed that the algorithm can classify packets rapidly. The DimCut algorithm has two separated levels, pre-processing level (tree construction and making index table) and search level.