Algorithms and Data Structures for Ip Lookups

Ioannidis, Ioannis. Ph.D., Purdue University, May, 2005. Algorithms and DataStructures for IP Lookups. Major Professor: Ananth Grama. The problem of optimizing access mechanisms for IP routing tables is an impor-tant and well studied one. Several techniques have been proposed for structuringand managing routing tables, with special emphasis on backbone, high-throughputrouting. Trie-based data structures have emerged as the most promising alternative,due to their efficiency, their conceptual simplicity and their adaptability to changesin the Internet Protocol and implementation details of the routing procedure. In thisthesis, we present three optimization methods for trie-based routing. The first gen-eralizes the well known level compression (LC) method so that all available memoryof a router can be utilized to lower the average lookup time in an LC trie. This isthe first method for generalized LC tries for which there exists an established per-formance bound. The second method reorganizes an LC trie into a directed acyclicgraph (DAG) with a view to packing trie parts together and further reducing depth.Our experiments with actual Internet backbone routing tables and a combination ofthe above two techniques show that the performance of LC tries can be improved byup to 25% using only 1% memory space in excess of that used by the original routingtable. Finally, we study the effect of caching on the access time for tries. We show that Least Recently Used (LRU) caching can be used to accelerate trie lookups, withthe use of intelligent main memory placement techniques. We prove that partition-ing a trie to minimize the expected number of distinct cache lines accessed duringa lookup optimizes the hit ratio of an LRU cache for an important class of memoryplacement techniques. We describe the optimal cache replacement policy / memoryplacement technique for trie lookups and show experimentally that LRU caching can