Web Service's Discovery Based On Peer-to-Peer

Peer-to-Peer network breaks the traditional model of Client/Server.P2P networkemphasizes each node's role is equal. Each node not only plays a role of server thatprovides resources, but also plays a role of client that uses resources. P2P network P2Pnetwork has characters such as good extension, robustness, use distributed resourcessufficiently. It can be applied in file sharing, resource locating and distributedcomputing.Web Service as a module technology Internet-based promotes distributedapplications. The discovery is the basic of Web Service. UDDI is a popular solutionsupports the discovery of Web Service which is model of Client/Server yet.After analyzing the possibility of Web Service associates with P2P network, thispaper propose a Peer-to-Peer approach of Web Service discovery. Paper first use{keyword, WSDL's URL} two elements array to describe a specified Web Service.Constructing structured P2P network as the infrastructure. Use consistent hashing tospecify identifiers of nodes and keywords and map keywords to correspond node. Chordis the basic location protocol of paper which is used by query keyword in order to getthe WSDL's URL.Original Chord reduces path length by half in each query step to find keywords.After analyzing the original Chord's performance, the paper give an algorithm toimprove on the efficiency of query which increase the size of finger table and decreasethe hops of query to reduce the path length of query. The simulation result shows thatthe algorithm is effective.The different frequency of word using will lead to load's imbalance. Afteranalyzing the origin of load imbalance, the paper give two algorithms which amongnodes and inner node. The simulation result shows that these two algorithms canimprove load balance effectively.MIT's P2PSim is a basic tool of paper, and it adds some functions to suffice thedemand of research.