A Study on the Bootstrapping Architectures for Scalable Private Reappearing Overlay Network

摘要

There has been a tremendous growth in the adoption of Peer-to-Peer (P2P) systems for media distribution and file sharing. However, P2P is now also being used as underlying technology for other purposes such as the distribution of software updates or as infrastructure for a distributed information repository. The research introduced in this paper is motivated by a use case in the latter area. This use case addresses the problem of document and information sharing between educational institutions in Malaysia. The additional challenge is that apart from not having any central instance peers have regular, daily down-times. During the night all peers disappear altogether. Hence the challenge is to find an effective way to set-up a P2P information and document sharing system taking into account the special characteristic of the dis- and re-appearing peers. In this paper, we investigate various bootstrapping architectures for the proposed Scalable Private Reappearing Overlay Network (SPRON). The bootstrapping needs to be self-organize, scalable and handle deterministic churn rate during its network formation period for peers (or computers) in various operational levels of the ministry of education. The deterministic churn rate is defined as the continuous process of node arrival and departure pattern of a P2P system. Within this paper we present the background information of the research theme, problem statements, and a proposed Scalable Private Reappearing Overlay Network (SPRON) design to illustrate a specialized P2P network and its unique set of requirements. The main focus of the paper is the presentation of the finding of our investigation regarding the bootstrapping performance of Chord, Chord with Aggressive Join, and Kademlia in order to establish the most appropriate P2P structure to base our system on. Through this we determine the suitability of these structured networks as the foundation in the proposed Scalable Private Reappearing Overlay Network design.