Data Sharing and Querying for Peer-to-Peer Data Management Systems

Abstract

In this thesis, we investigate mechanisms to support data sharing and querying in a peer-to-peer data management system, that is, a peer-to-peer system where each peer manages its own data. To support data sharing, we propose the use of mapping tables which list pairs of corresponding data values that reside in different peers. We argue why mapping tables are appropriate for data sharing in a peer-to-peer environment, and we discuss alternative semantics for these tables. We motivate why reasoning capabilities are needed to manage mapping tables, and we show the importance of inferring new tables from existing ones, or checking the consistency of existing tables. Finally, we present algorithms and experimental results which show that mapping tables may be managed efficiently in practice. In terms of querying, we propose a framework in which users pose select-project-join queries with respect to the schema of their local peer. Then, we provide a rewriting mechanism that uses mapping tables to translate a locally expressed query to a set of queries over the acquainted peers. We consider both sound translations (which only retrieve correct answers) and complete translations (which retrieve all correct answers, and no incorrect answers). We describe basic algorithms to compute such translations, and we offer optimized versions of these algorithms that offer savings both in terms of execution time and in terms of memory. Finally, we present experiments that show the performance and efficiency of our algorithms under various settings.