Highly Adaptive Lookup Systems for Peer-to-Peer Computing

Abstract

The performance of file sharing peer-to-peer systems depends to a large degree on the speed of lookup operation. A number of proposed solutions rely on distributed hashing techniques. Traditionally nodes are assigned fixed length identifiers which does not allow the table to expand or shrink with the increase or decrease in the node count. With the fixed length identifiers, the performance deteriorates when the number of nodes reaches a high value. On the other hand, the overhead of maintaining per-node state (i.e., information of all adjacent nodes) can be unnecessarily large if the number of nodes is small. The focus of our study is to propose a way to combine the distributed and dynamic nature of the system in a way that allows large and unpredictable changes in both the number and the distribution of nodes while providing scalability and good performance. Our approach is based on dynamic distributed hashing techniques; node identifier length varies with the number of nodes in the system. Hence, the system can adapt to the changing conditions and maintain good performance. In this paper, we describe the operations of the proposed adaptive system and verify its performance through simulations.