RQrRi--A Rsrouce Quality Rating And Reputation Iterating Strategy For P2p Systems

Peer-to-peer network, which knowing as one of the Real-world Applications, has been playing a more and more important role in people's everyday Internet life. Although the p2p technology has done a lot in solving some inherent problems of traditional Client/Server based systems, but there still exists many problems in the p2p technology to be researched since it is still in its infancy. For now, most p2p research works are concentrating on the structures or the querying strategies of the p2p system rather than dealing with peers'interactivities and resources'qualities. That causes many different p2p implementations never showing their convenience and effectiveness as people's desire. Traditional p2p search engines simply allow peers requesting resources by sending their queries and receiving reponses that match the queries from other peers. But the efficiency of such kind of search is very poor. This can be addressed in two aspects. First, resources provided by peers usually do not have enough metadata information about their qualities, and most p2p systems only use keywords extracted from the search query to match resources'filenames and file extentions b. But as we all know, information provided by the filename is very limited as for the resource's content, so, in fact we can not precisely locate the desired resources. Second, being lacking of information about resources'qualities, it's quite hard for the end user to choose the appropriate one of the items from the resource list to download, because the resource list may contain handreds of resource items received from deferent peers. Actually, some p2p research shows that the peer-to-peer system is the one which structure and behaviour resemble the real world most in all kinds of network systems, while peers of a p2p system are also the best representatives of the end user——human being. Consequently, by borrowing the model of the real world society and merits of some useful technologies, like personalized web search, we propose one strategy for rating resources'qualities and managing peers'reputation——rQrRi(Resource Quality Rating and Reputation Iterating Strategy). In this paper, by combining users'interests and resources'metadata information, we address the strategies and algorithms of rQrRi. We also analyse the rQrRi's feasibility basing on several proposed theories and rQrRi's robustness under the attacks of several security threat models.