| As more and more people using P2P systems, more and more resources are stored in them, and how to provide services to quickly search and locate on the Internet has increasingly become an important issue. As the key technology of P2P systems, searching resource and location algorithm have always been the researching focus. There are some classic P2P algorithms such as Chord, CAN, Pastry, Tapestry and Kademlia, these algorithms generally use DHT (Distributed Hash Table) which organize the nodes in the network, its essence is that the resources and the nodes which resources are stored are evenly distributed to some value space through hash function, each node is responsible for the value of the small part of the resources, thereby to realize the whole DHT overlay network addressing and storage. In the whole DHT algorithms, Kademlia algorithm is the most common application due to its simplicity and efficiency. But as the representative P2P algorithm, Kademlia is general lack of quality of service (QoS) support, and it can not meet the needs of practical application. Therefore, it has important practical significance to increase the QoS support based on the study of the correlation algorithm.This paper focused on P2P routing algorithm, and the most representative Kademlia algorithm of P2P algorithms to study deeply, proposed storage algorithm and resource location algorithm based on QoS on the basis of original algorithm, that is the resource is stored redundantly when released, the node identifier and resource identifier must have common prefix, the resource is positioned to common prefix length as the location of the target. In order to check the effect of the algorithm, this paper proposes a QoS model. Through the simulation in the general overlay network, the new algorithm reduced 10% to 20% in the total number of messages than the original algorithm, the average delay was reduced by approximately 10%, the searching success rate can even reach 100%. |
PDF Full Text Request