International Journal of Computer Applications (0975 - 8887) Volume 128 - No.6, October 2015 A Reputation based System to Overcome Malicious Behavior in Peer-to-Peer Networks A.M. Anisul Huq Faculty Member, Department of CS, American International University - Bangladesh (AIUB), Dhaka - 1213, Bangladesh Mosharraf Hossain Khan Daffodil International University, Bangladesh ABSTRACT In recent times, the growth in the number of subscribers of peer to peer networks has been phenomenal. Anonymity being a character of such networks also gave rise to the number of free-riders and ma- licious behaviors. Though free riders consume network bandwidth and decrease the network performance by displaying selfish behav- ior, they are not a serious threat for the rest of the co-operative peers. Malicious peers on the other hand, spread viruses, worms, Trojans in the network, provide misleading feedback and try to dis- rupt the existing trust among the peers. Therefore, it is absolutely essential that a peer has reliable reputation information about other peers in order make informed decisions (e.g., who to download files from, who to serve content, which is a malicious node etc.). In this paper, we have proposed a reputation system that uses objective criteria to track each peer’s contribution in the system and allows peers to store their reputations locally. In our opinion, such mea- sures will eventually root out malicious peers from the system. 1 Keywords reputation, peer-to-peer network, trust, eclipse attack 1. INTRODUCTION The earliest application of peer-to-peer (P2P) was for newsgroups (USENET) and to exchange messages (FidoNet) [2]. Then Napster emerged. With its free music sharing platform and subsequent bat- tle with the big music corporations brought the whole concept of P2P networks into limelight. P2P networks are primarily used for sharing files and more recently for distributed computations. But studies have shown that the ma- jority of file sharing users do not offer any files for upload, but only download from others [4, 3, 18, 20, 21]. Those who do share are doing it mostly out of ignorance, for not even being aware of it. Or maybe they are indifferent about it, as their uplink bandwidth would simply go unused otherwise and their own download service 1 This paper is an extension of the work produced by the first author as a part of his Master’s course work done at Aalto University, Helsinki, Fin- land. It is available at: http://www.cse.hut.fi/en/publications/ B/5/papers/huq_final.pdf. Note that, this work is not a peer reviewed publication. quality does not suffer from uploads [3]. The presence of malicious peers is further complicating matters and this is the main concern of this paper. They pose a bigger threat because their main goal is to destroy data [25] or damage the infrastructure by propagating worms in the system [26]. As all the systems in a p2p network run the same software, it is very easy for an attacker to compromise the whole network by finding a single exploitable security hole in that software [17]. In order to address these two challenges, we have proposed a repu- tation mechanism. However, there is no universal agreement on the definition of reputation. In this paper, we have adopted the follow- ing working definition: Reputation: a peer’s belief in another peer’s capabilities, hon- esty and reliability based on recommendations received from other peers. This recommendation facility is also extended to individual files. Reputation can be centralized, computed by a trusted third party, like a Better Business Bureau; or in our case, it is decentral- ized, computed independently by each peer after asking other peers for recommendations. The structure of the paper is as follows: background knowledge is presented in Section 2. On the subsequent section, specific p2p at- tacks and respective defense mechanisms are described. In section 4, we discuss our proposed scheme of reputation. The fifth section analyzes the result and provides relevant comparison.Finally, sec- tion 6 concludes the paper. 2. BACKGROUND Peer-to-peer (P2P) networks have introduced a new paradigm in content distribution. Each peer is both a client and a server in these networks. Users are drawn to these networks due to the ability to locate a wide variety of multimedia content. Currently, there are several different architectures for P2P networks: (1) Centralized: There is a constantly-updated directory hosted at central locations. Nodes issue queries to this central directory server to locate which other nodes hold the desired files. Such centralized approaches do not scale well and have single points of failure. (2) Decentralized but Structured: Such systems do not have cen- tral directory server, but possess significant amount of structure by the way of P2P overlay topology. Such a topology is tightly controlled and files are placed at specific locations that enables 1