P2P-DIET: One-Time and Continuous Queries in Super-Peer Networks

摘要

In peer-to-peer (P2P) systems a very large number of autonomous computing nodes (the peers) pool together their resources and rely on each other for data and services. P2P systems are application level virtual or overlay networks that have emerged as a natural way to share data and resources. The main application scenario considered in recent P2P data sharing systems is that of one-time querying: a user poses a query (e.g., "I want music by Moby") and the system returns a list of pointers to matching files owned by various peers in the network. Then, the user can go ahead and download files of interest. The complementary scenario of selective dissemination of information (SDI) or selective information push is also very interesting. In an SDI scenario, a user posts a continuous query to the system to receive notifications whenever certain resources of interest appear in the system (e.g., when a song of Moby becomes available). SDI can be as useful as one-time querying in many target applications of P2P networks ranging from file sharing, to more advanced applications such as alert systems for digital libraries, e-commerce networks etc. At the Intelligent Systems Laboratory of the Technical University of Crete, we have recently concentrated on the problem of SDI in P2P networks in the context of project DIET (http://www.dfki.de/diet). Our work, summarized in, has culminated in the implementation of P2P-DIET, a service that unifies one-time and continuous query processing in P2P networks with super-peers. P2P-DIET is a direct descendant of DIAS, a Distributed Information Alert System for digital libraries, that was presented in. P2P-DIET combines one-time querying as found in other super-peer networks and SDI as proposed in DIAS.