Efficient Dynamic IPv4/IPv6 Lookup Scheme Based on the Most Specific Prefixes

IP lookup schemes in Internet routers can be broadly classified into two categories: static and dynamic. While static schemes improve lookup speed and reduce memory requirement by using pre-computations, their drawback is that when prefixes are added or deleted, the entire data structure may need to be rebuilt and thus very slow. Dynamic schemes can perform fast prefix insertions and deletions in real time. In this paper, we develop a dynamic scheme called Most Specific Prefix Tree (MSPT). MSPT is a balanced binary search tree based on the most specific prefixes in the routing table. The prefixes other than the most specific ones are allocated to the enclosure sets of some nodes in MSPT. The worst-case time complexities of MSPT search, insertion, and deletion algorithms are O(logM), O(logM + W), and O(logM×logW), respectively, where M is the number of nodes in MSPT. However, in practice, the time complexities of the MSPT search, insertion, and deletion are all O(logM) because M is very close to the routing table size. Our IPv4/IPv6 experiments show that MSPT not only performs better than the existing dynamic and static schemes in update speed, but also better than the existing dynamic schemes in lookup speed.