Query Processing in Structured Peer-to-Peer Networks

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

Recently a new generation of P2P systems, offering distributed hash table (DHT) functionality, have been proposed. These systems greatly improve the scalability and exact-match accuracy of P2P systems, but offer only the exact-match query facility. This proposal outlines a research agenda for building complex query facilities on top of these DHT-based P2P systems. There are three defining directions of the proposed research agenda: Core Algebra: Explore the implementation of a core set of algebraic query operators that run over DHTs in a P2P network. DHT APIs: Identify the minimum set of primitives and functions that need to be implemented by a P2P network in order to efficiently support the query operators. Query Optimization: P2P networks are often distributed across slow network links, and may benefit from distributed database optimizations like the use of semi-joins. However, it is not clear how the various ideas in the query processing literature map into the P2P context, particularly when using DHTs for routing during query processing. This proposal intends to study the costs and tradeoffs of various alternatives for the query plan space in this context. Simulation of large networks will be used to study the proposed techniques at Internet scale.