Bit-shuffled trie: A new approach for IP address lookup

摘要

IP address lookup is a fundamental operation in packet forwarding. Using multi-level index tables to find out the next-hop value is an attractive approach due to its simplicity. However, memory efficiency is relatively low because prefixes are sparsely distributed in the address space. In this poster, we shall outline a new approach to construct memory efficient index tables based on a technique called bit-shuffling. The proposed method is evaluated using a real-life IPv4 routing table with 321K prefixes. The lookup tables occupy 0.8MB memory.