Görevdeş ağlarda sık bulunan öğe setinin bulunması

Several P2P applications require a global view of system information such as data access frequencies, item frequencies, query and event counts, that are available locally and partially at peers. Frequent item set discovery (FID) in a distributed environment is a common problem requiring global information computation. Items that globally occur more than a threshold value are referred as frequent or popular and the number of diverse applications that need globally frequent items is increasing expeditiously in today's P2P networks. Therefore, efficiently discovering frequent items would be a valuable service for peers. Being significant for P2P systems, FID problem is also applicable to distributed database applications, cache management, data replication, sensor networks, and security mechanisms in which identifying frequently occurring items in the entire system is useful.In this thesis, we propose and develop a gossip-based distributed approach, namely ProFID, for discovering frequent items in unstructured P2P networks. In contrast to the prior studies, our solution progresses in a fully distributed manner using an atomic averaging function to discover frequent items. Utilizing averaging function with gossip-based aggregation in frequent item set discovery problem and a practical convergence rule are novel and beneficial features of our approach. We make the following contributions to the current state of the art. First, we propose a fully distributed Protocol for Frequent Item Discovery (ProFID) where the result is produced at every peer. ProFID uses a novel pairwise averaging function and network size estimation together to discover frequent items in an unstructured P2P network. We also propose a practical rule for convergence of the algorithm. In contrast to previous works, each peer gives local decision for convergence based on the change of updated local state. Moreover, we developed a model of ProFID in PeerSim and performed various experiments to compare and evaluate its efficiency, scalability, applicability. Finally, we compared the accuracy and scalability of ProFID with adaptive Push-Sum algorithm. The comparison results show that ProFID outperforms adaptive Push-sum in terms of accuracy, convergence speed and message overhead.