| Summary: | 碩士 === 中興大學 === 資訊科學系所 === 95 === Due to the fast development of Internet, the routing table in backbone router of core network increases exponentially and reaches 200,000 entries formally. This means that the simple algorithm supporting fast incremental update with high pipelined hardware architecture is suitable to process huge IP routing lookups.
In the thesis, we present a data structure based on Level Smart-Compression Trie[2] scheme for IP lookup, which is called Prefix Dominant String (PDS), to decrease a large number of memory space requirement for big routing table. Therefore, it can raise the performance of original scheme and be expected to achieve fast IP lookup, low memory space requirement, supporting incremental update, and scalability. To utilize our scheme with 281,032 routing table entries (N=4, M=8) in popular platform and compare with LS-Compression Trie[2] scheme, the usage of memory space decreases from 6,220 KB to 601 KB, which saves 90.33% memory space. The performance for IP lookup increases from 45 MPPS to 100 MPPS.
|
|---|
