Chinese Journal of Computers   Full Text
Title	Redundancy Schemes for High Availability in DHTs
Authors	CHEN Gui-Hai1) WU Fan2) LI Hong-Xing1) QIU Tong-Qing3)
Address	1)(State Key Laboratory of Novel Software Technology, Nanjing University, Nanjing 210093) 2)(Department of Computer Science and Engineering, SUNY at Buffalo, Buffalo, NY 14260) 3)(College of Computing, Georgia Institute of Technology, Atlanta, GA 30332)
Year	2008
Issue	No.10(1695—1704)
Abstract &   Background	Abstract High availability in peer-to-peer DHTs requires data redundancy. This paper takes user download behavior into account to evaluate redundancy schemes in data storage and share systems. Furthermore, it proposes a hybrid redundancy scheme of replication and erasure coding. Experimental results show that replication scheme saves more bandwidth than erasure coding scheme, although it requires more storage space, when average node availability is higher than 47%. The hybrid scheme saves more maintenance bandwidth with acceptable redundancy factor. Keywords DHT； peer-to-peer； availability； redundancy； erasure-coding； replication Background This work is supported by many research funding agencies both from China and from the US, including China NSF grants (No.60573131，No.60673154), China 973 projects (No.2002CB312002, No.2006CB303000), U.S. NSF grant ACI-0203592 and U.S. NSF grant CCF-0524030. The purpose of the work aims to establish scalable techniques for large-scale structured P2P systems. The Sino-US joint research team have kept friendly collaboration for more than 4 years and produced fruitful results in the related fields. For example, the recently-proposed constant-degree P2P overlay network (Cycloid) was published in Chinese Journal of Computer in July 2005. This paper presents the authors’ recent achievement in the study of DHT overlay networks for highly available file systems. A preliminary version of this paper was once presented on 3rd International Symposium on Parallel and Distributed Processing and Applications in 2005.