Incentive Based Approach to Detect Selfish Nodes in Mobile P2P Network.

摘要

The growth of mobile devices led to the wide use of Mobile P2P networks. These networks are used in a wide variety of areas and hence there is lot of research in the field of mobile networks. Detecting selfish nodes is one of the research topics triggered due to the popularity of mobile P2P networks. It is necessary to detect selfish nodes in such networks to improve the efficiency of the network. In this thesis, an incentive based approach to detect selfish nodes is designed and evaluated. This approach differs from the existing work as it (i) can be used with any underlying routing protocol assuming there are no attacks due to routing protocol (ii) is able to detect selective behavior of nodes where nodes drop some packets and forward some (iii) prevents a wide variety of malicious activities or attacks by nodes in the network (iv) prevents false positives due to connectivity issues in the network. We assume the presence of some trusted nodes called Broker nodes and propose a way using which nodes in the network communicate. Each intermediate node sends a receipt to the Broker node which it uses to identify selfish nodes in the network. Each node has a currency assigned which it uses to pay others for the forwarding service. Currency of a node is changed based on the receipts sent by that node. When the currency level of a node below some threshold, it is designated as selfish node in the network. This approach is experimentally evaluated and is found to outperform some of the recent work in this area in terms of time to detect selfish nodes and overhead involved.