Self-Organizing Subsets: From Each According to His Abilities, to Each According to His Needs

Abstract

The key principles behind current peer-to-peer research include fully distributing service functionality among all nodes participating in the system and routing individual requests based on a small amount of locally maintained state. The goals extend much further than just improving raw system performance: such systems must survive massive concurrent failures, denial of service attacks, etc. These efforts are uncovering fundamental issues in the design and deployment of distributed services. However, the work ignores a number of practical issues with the deployment of general peer-to-peer systems, including i) the overhead of maintaining consistency among peers replicating mutable data and ii) the resource waste incurred by the replication necessary to counteract the loss in locality that results from random content distribution. This position paper argues that the key challenge in peer-to-peer research is not to distribute service functions among all participants, but rather to distribute functions to meet target levels of availability, survivability, and performance. In many cases, only a subset of participating hosts should take on server roles. The benefit of peerto-peer architectures then comes from massive diversity rather than massive decentralization: with high probability, there is always some node available to provide the required functionality should the need arise.

The nearest replica can be farther than you think

“…Replica Selection algorithms Geo-distributed storage systems tend to forward client's requests towards the "closest" replicas to minimize network delay and to provide the best performance. This task commonly occurs, e.g., in selforganizing overlays [39]. One of the primary tasks is to correctly compute or estimate the distance among the nodes; various systems have tackled this problem.…”

Efficient Recovery from Organizational Disconnects in SkipNet

“…As we and others [4,5,12] have noted, much of the data stored within peer-to-peer systems can still be expected to exhibit significant locality in clients' access patterns. Thus, for example, organizations may wish to make data globally available while still storing it locally in order to obtain better intra-organizational availability and performance.…”

Design of the P2P system based on heterogeneous characters

“…The question now is how to determine the appropriate subset of m to meet application performance and availability requirements. The law of conservation of network resources is given by [6] : nu=mde (1) Where e=nu/md; u is the average resource consumption of n nodes in P2P system, d is the ___________________________________ average resource provided by the subset of m nodes, and e is the efficiency of service delivery.…”