Centralized data silos are not only becoming prohibitively expensive but also raise issues of data ownership and data availability. These developments are affecting the industry, researchers, and ultimately society in general. Decentralized storage solutions present a promising alternative. Furthermore, such systems can become a crucial layer for new paradigms of edge-centric computing and web3 applications. Decentralized storage solutions based on p2p networks can enable scalable and self-sustaining open-source infrastructures. However, like other p2p systems, they require well-designed incentive mechanisms for participating peers. These mechanisms should be not only effective but also fair in regard to individual participants. Even though several such systems have been studied in deployment, there is still a lack of systematic understanding regarding these issues. We investigate the interplay between incentive mechanisms, network characteristics, and fairness of peer rewards. In particular, we identify and evaluate three core and up-to-date reward mechanisms for moving data in p2p networks: distance-based payments, reciprocity, and time-limited free service. Distance-based payments are relevant since libp2p Kademlia, which enables distance-based algorithms for content lookup and retrieval, is part of various modern p2p systems. We base our model on the Swarm network that uses a combination of the three mechanisms and serves as inspiration for our Tit-for-Token model. We present our Tit-for-Token model and develop a tool to explore the behaviors of these payment mechanisms. Our evaluation provides novel insights into the functioning and interplay of these mechanisms and helps. Based on these insights, we propose modifications to these mechanisms that better address fairness concerns and outline improvement proposals for the Swarm network.
翻译:暂无翻译