Economic Modelling and Incentive Mechanisms for Ecient Resource Provision in Peer-to-Peer Systems
Panayotis Antoniadis · 2006
Personal computers and mobile devices have today significant capabilities, they are connected through high-speed Internet connections or directly, forming ad-hoc networks, and thus provide the potential for the deployment of sophisticated distributed applications which seek to exploit the vast amount of aggregated underutilized resources available; the so-called peer-to-peer (p2p) applications. File sharing, which was the application responsible for the p2p hype started around the year 2000, is still the most popular and the only one widely deployed. But many more are being proposed and designed so as to exploit different types of resources such as computing power, storage, access bandwidth, and more. A fundamental difference of p2p applications compared to traditional distributed systems is the fact that decisions of the individual peers are based on their own self-interest and this in principle leads to inefficient system operation. In particular, a rational peer would participate in the system without contributing any resources following the so-called ‘free riding’ strategy, which could in many cases have a detrimental effect on system’s efficiency. So, suitable incentives should be given to peers in order to behave towards maximizing the overall efficiency. But there are very challenging research questions to be addressed in this context both in terms of economic modelling and implementation which have attracted lately many researchers from all the relevant research fields of computer science and economics. In this dissertation we identified and categorized the most important related concepts due to current practice and research activity, and we chose to focus on the issue of content provision in p2p file sharing systems as our main case study. Content availability is a non-rivalrous resource since files are not consumed by downloading them, and thus it has the main property of what is called in economics a ‘public good’. Hence, we model content provision in p2p systems as a problem of private provision of a public good. Three characteristics of p2p systems that play important role in our analysis are the following: the very large number of participants, the high degree of heterogeneity, and the very challenging implementation issues due to the fully distributed and untrusted environment. Notably, the latter is the one that gives new interest to this classic, and very challenging, problem in the economics literature. The large system size reduces, in general, the incentives to peers to contribute the efficient amount of resources for the provision of the public good because they feel less pivotal for its con-