Patent attributes
A method and apparatus for classification of packets is provided. The method includes classifying the packets based on the structured part of at least one packet classification rule, and classifying the packets based on the unstructured part of the at least one packet classification rule, the classification being done based on the structured classification results. The packet classification method provides a technique for splitting the n-dimensional space of the packet header fields into disjoint regions. The splitting is done such that all the packets falling into a region have the same packet classification result. The first packet falling into a region takes a longer time for classification, where the logical expression resulting from user-configured rule is solved. The classification of all subsequent packets falling into the same region gets accelerated and takes less time than the first packet to classify.