NBDT: an efficient P2P indexing scheme for web service discovery

Abstract

In this paper, we propose a new infrastructure for web services discovery in P2P networks, the Nested Balanced Distributed Tree (NBDT). Peers that store web services information, such as data item descriptions, are efficiently located using a scalable and robust data indexing structure based on a NBDT. The key innovation is that the solution is based on a totally new infrastructure, which is not vulnerable to classic disadvantages. We present a theoretical analysis backed up by experimental results, which shows that the communication cost of query and update operations scale sublogarithmically in the worst case with the number of NBDT nodes, outperforming three of the most popular decentralised infrastructures: Chord (and some of its successors), BATON (and its successor) and Skip Graphs. Furthermore, we show that the network is robust to failures providing quality of web services requirements.