In recent years, the peer-to-peer (p2p) computing paradigms have attracted tremendous attention from both industry and research communities, since they are appropriate for building large-scale distributed indexing structures encompassing huge amount of users and massive digital contents across continents. However, most existing peer-to-peer systems cannot directly or efficiently support complex queries.
Structured p2p systems, also known as Distributed Hash Tables based p2p systems, are designed for exact-key searching, and the support of complex queries in such systems is still a challenging problem. In this dissertation, we extend the distributed hash tables to efficiently support complex queries. Based on efficient space mapping and query routing mechanisms, our architecture can provide a general platform to simultaneously support any number of indexes with different data types. Significantly, this architecture does not need to generate or maintain search trees for each index scheme. Instead, it exploits the embedded trees in the underlying distributed hash tables to refine and deliver queries or disseminate data objects. To deal with skewed data distribution, we provide load-balancing mechanisms to statically or dynamically adjust loads among nodes and ensure that no node in the system is unduly loaded.
Unstructured p2p systems can support any complex queries. However, such systems are low efficient, either a large amount of nodes have to be probed to get a high recall rate or some relevant data objects would be missed, due to the loose control of the data location and the network topology. In this dissertation, we propose a distributed, content-based and heuristic feedback mechanism, which allows peers to keep track of recent queries and learn from the assessment of answers to previous queries, so as to self-adaptively route the subsequent query to nodes which have the most relevant data objects to the query. Therefore, a high recall rate can be achieved by probing only a small amount of peers.
The performance of our designs has been evaluated through extensive simulations with a variety of metrics. The experimental results show that our proposed architectures are scalable, efficient in routing and processing complex queries and can improve the quality of query results.