ransub-usits03 - Using Random Subsets to Build Scalable...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Using Random Subsets to Build Scalable Network Services Dejan Kosti´c, Adolfo Rodriguez, Jeannie Albrecht, Abhijeet Bhirud, and Amin Vahdat Department of Computer Science Duke University dkostic,razor,albrecht,abhi,vahdat @cs.duke.edu Abstract In this paper, we argue that a broad range of large-scale network services would benefit from a scalable mecha- nism for delivering state about a random subset of global participants. Key to this approach is ensuring that mem- bership in the subset changes periodically and with uni- form representation over all participants. Random sub- sets could help overcome inherent scaling limitations to services that maintain global state and perform global network probing. It could further improve the routing performance of peer-to-peer distributed hash tables by locating topologically-close nodes. This paper presents the design, implementation, and evaluation of RanSub , a scalable protocol for delivering such state. As a first demonstration of the RanSub utility, we con- struct SARO, a scalable and adaptive application-layer overlay tree. SARO uses RanSub state information to locate appropriate peers for meeting application-specific delay and bandwidth targets and to dynamically adapt to changing network conditions. A large-scale evaluation of 1000 overlay nodes participating in an emulated 20,000- node wide-area network topology demonstrate both the adaptivity and scalability (in terms of per-node state and network overhead) of both RanSub and SARO. Finally, we use an existing streaming media server to distribute content through SARO running on top of the PlanetLab Internet testbed. 1 Introduction Many distributed services must track the characteristics of a subset of their peers. This information is used for failure detection, routing, application-layer multicast, re- source discovery, or update propagation. Ideally, the size This research is supported in part by the National Science Founda- tion (EIA-9972879, ITR-0082912), Hewlett Packard, IBM, Intel, and Microsoft. Albrecht is also supported by an NSF graduate fellow- ship and Vahdat is also supported by an NSF CAREER award (CCR- 9984328). of this subset would equal the number of all global par- ticipants to provide each node with the highest quality in- formation. Unfortunately, this approach breaks down be- yond a few tens of nodes across the wide-area, encounter- ing scalability limitations both in terms of per-node state and network overhead. Recent work suggests building scalable distributed systems on top of a location infras- tructure where each node can quickly (in steps) locate any remote node while maintaining only local state [22, 24, 26, 30]. This approach holds promise for scaling to distributed systems consisting of millions of participating nodes. While existing techniques track the characteristics of a
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 14

ransub-usits03 - Using Random Subsets to Build Scalable...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online