Searching resources in peer-to-peer network using friend and path result - Sharing searching concepts.

A new searching technique is proposed in this paper for unstructured Peer-to- Peer file sharing system which uses the search path result sharing concept and also uses the friend concept. As per this new concepts, a friend set or groups are formed by joining the closely neighbor nodes in the unstructured peer to peer network system. Many friend sets can be formed and each friend set has one Head node, heading all other nodes. This head node is not a centralized node and just maintains a list of all other nodes in that friend set and their meta data in its own table. The friend set can be formed within certain radius and the head node contains the IP addresses of all nodes in own friend group or friend set. The search radius is seven hop count distance. Since the head node contains the meta data of all nodes in its friend set, it knows which resource or file is available in which node in its own group. Whenever a source node finds a resource by sending queries, it stores that path information of that resource in its own path table and also shares that path information with all other nodes in its friend set and head node. When searching, if the source node itself has the searching resource, it will directly display that resource or file to user so that he can use it or if at least it has that path information of that resource in its own path table, it will take that path information from the table and display that link on the screen so that user download it. If not, it will route the query messages only to all head nodes in the network simultaneously. The head nodes, send query to all nodes in its own set and will response to the source node on behalf of all other nodes in their own set. Inside the network, the generation of query messages is greatly reduced. Similarly, as far as reply is concerned, the head node bundled all the reply messages belong to its set and send to source as a single response message. Thus the response message going back to source is also reduced. But if a head node fails due to some reasons, the entire set of nodes in that friend set cannot be searched. So this algorithm uses dynamic search algorithm to search the uncovered nodes. The main advantage of this algorithm is all nodes in friend set gets information about resources from their friend nodes without searching. This algorithm also reduces the search time, search cost and bandwidth consumption. This algorithm achieved 75 % reduction of query messages, 50 % reduction of response messages, 50 % reduction in bandwidth and 50 % reduction in search cost. [ABSTRACT FROM AUTHOR]