Copy the page URI to the clipboard
Rahman, R.; Hales, D.; Vinkó, T.; Pouwelse, J. and Sips, H.
(2010).
DOI: https://doi.org/10.1109/HPCS.2010.5547112
Abstract
Many peer-to-peer file sharing communities implement credit policies to incentivise users to contribute upload resources. Such policies implicitly assume a user model - how the user controlling each peer behaves. We show using an agent-based model that credit policies, based on bandwidth contribution, and a selfish user model, can lead to both “crunches” and “crashes” where the system seizes completely due to too little credit or too much credit. We explore the conditions that lead to these system pathologies and present a theoretical analysis that allows us to determine if a community is sustainable or will eventually crunch or crash. Finally we apply the analysis to produce a novel adaptive credit system that automatically adjusts credit policies to maintain sustainability.