A Detailed Evaluation of Threshold Algorithms for Answering Top-k queries in Peer-to-Peer Networks
Ioannis Chrysakis, Constantinos Chalkidis, Dimitris Plexousakis, Vassilika Vouton · 2010
Ranking queries, also known as top-k queries, have drawn considerable attention due to their usability in various applications. Several algorithms have been proposed for the evaluation of top-k queries. A large percentage of them follow the Threshold Approach. In p2p networks, top-k query processing can provide a lot of advantages both in time and bandwidth consumption. We focus on the main adaptations of threshold algorithms fulfilling the requirements of modern p2p applications. We introduce two algorithms optimized for ranking queries in p2p networks and present their characteristics. In the setting of a simulation of large p2p networks, we evaluate the performance of Threshold Algorithms. Our experiments demonstrate that in some cases a threshold algorithm can improve top-k query processing, while in others it is far more costly. The results show that distributed query processing can be more effective than a simple threshold algorithm in a p2p network.