| Summary: | 碩士 === 國立成功大學 === 資訊工程學系碩博士班 === 98 === Packet classification is generally referred to the process of categorizing packets into flows in networking devices. All packets belonging to the same flow obey a pre-defined rule and are processed in a similar manner. Packet classification has wide applications such as unauthorized access prevention in firewalls and Quality of Service supported in Internet routers. The Classifier containing all the pre-defined rules are processed by the router for each incoming packet in order to find the best matching rule and take appropriate actions. However, the sizes of classifiers increase as quickly as the Internet traffic grows. Thus, how to reduce the memory usage for storing classifiers in routers is a very important problem. In this thesis, we focus on the range encoding problem in reducing the ternary content-addressable memory (TCAM) requirement. Existing range encoding schemes are usually 1-dimensional schemes that perform range encoding processes in the range fields independently. We propose an efficient 2-dimensional range encoding algorithm. The original 2-dimensional classifier is first divided into a set of elementary regions that will be mapped onto the vertices of a multi-dimensional hypercube. The elementary regions covered by each 2-D range can be also be mapped to a sub-cube in order to represent the 2-D range by only one ternary string. By using the proposed encoding algorithms, we can minimize the size of the hypercube. Our performance experiments on real-life classifiers show that the proposed 2-dimensional range encoding schemes uses less TCAM memory than the previously proposed 1-dimension-based schemes. Compared with the classifiers of 10k rules, our proposed schemes can reduce 18% ~ 25% TCAM memory usage than 1-Dimansional encoding scheme which has the best performance, and in classifiers of 5k rules, our proposed schemes can also reduce 20% ~ 25% TCAM memory usage.
|
|---|
