Fang Qi, Haiying Shen, Harrison Chandler, Guoxin Liu, and Ze Li
In recent years, peer-to-peer (P2P) networks have gained popularity in many large-scale, distributed Internet applications. P2P networks enable the sharing of globally scattered computer resources, allowing them to be collectively used in a cooperative manner for different applications such as file sharing [ 1‒4], instant messaging , audio conferencing , and distributed computing . Node cooperation is critical to achieving reliable P2P performance but proves challenging since networks feature autonomous nodes without preexisting trust relationships. Additionally, some internal nodes may be compromised, misbehaving, selfish, or even malicious. Then, a critical problem arises: How can a resource requester choose a resource provider that is trustworthy and provides high quality of service (QoS) from among many resource providers?
The main way to address this problem is with reputation management systems. Like the reputation systems in the eBay , Amazon  and Overstock  e-commerce platforms, reputation systems employed in P2P networks compute and publish global reputation values for each node based on a collection of local ratings from others in order to provide guidance in selecting trustworthy nodes. They thwart the intentions of uncooperative and dishonest nodes and provide incentives for high ...