|Autor:||K. Jünemann, P. Andelfinger, H. Hartenstein||Links:||Volltext als PDF|
|Quelle:||Workshop on Grid and P2P Systems and Applications (GridPeer 2011), Proceedings of the 20th International Conference on Computer Communications and Networks (ICCCN), Maui, Hawaii, USA, August 2011|
Distributed Hash Tables (DHTs) prove valuable for distributed architectures by providing distributed information lookup, routing, and data storage while promising good scalability and robustness at the same time. From this point of view, a DHT could be seen as a basic service that can be used to build distributed applications. Whereas today no widely deployed and publicly accessible basic DHT service exists and thus DHT-based applications have to deploy their very own DHT networks, DHTs consisting of millions of peers are formed by file sharing clients. Although the interfaces of typical DHTs used for file sharing are too narrow, a basic DHT service could probably be created by bundling a suitable client implementation with file sharing software. In this paper, we evaluate whether a basic DHT service could suit the needs of DHT-based applications in terms of stability, number of participating peers, the peers' session lengths, geographical distribution and peer connectivity when deployed similar to DHTs driven by file sharing. As these metrics mostly depend on end user behavior rather than on the DHT protocol, we report on measurement results gathered from monitoring of the BitTorrent Mainline client's DHT over six months. We analyze which metrics would fit a basic DHT service and which could prove problematic. Furthermore, we discuss resulting technical requirements for an appropriate DHT protocol.