GeoSwarm: A multi-source download scheme for peer-to-peer location-based services

Existing peer-to-peer (p2p) overlays for location-based services suffer from two major drawbacks: (i) they do not store data persistently under peer churn and (ii) they do not allow for the fast retrieval of large files, especially under asymmetric link conditions. This tremendously limits the use of current and future p2p location-based services as users are not able to share larger files such as high resolution pictures or video snippets. To overcome these two problems, we present GeoSwarm: a reliable multi-source download scheme for p2p location-based services. GeoSwarm combines the strengths of a BitTorrent-like download scheme with the locality awareness of an overlay for location-based search and its built-in replication mechanism. Thereby, a robust and fast download as well as a reliable storage of location-related multimedia content is achieved. Through extensive evaluation, we show that 95% of all downloads in GeoSwarm are carried out successfully even under churn, while downloads benefit from a 100% increased throughput in comparison to traditional single-source downloads.