Towards Mobile Computational Application Steering: Visualizing The Spatial Characteristics of Metropolitan Area Wireless Networks

摘要

Computational application steering from mobile devices is an attractive proposition and one that is being actively explored. In wireless networks, particularly metropolitan area ifrastructural networks, a range of unique temporal and spatial performance characteristics may affect the ability to perform computational steering. The network coverage area contains zones, termed trouble spots, that are distinct pockets where garbled, poor, or no network coverage exist. They have been experimentally attributed to geographic features (e.g. tunnels and tall buildings). Point-to-point manifestations of trouble spots include stalled web pages, slow ftp connections and data corruption which, although detrimental, do not compromise system usability. However, computational applications such as steering are highly susceptible, and could suffer from serious operational and semantic problems in the presence of trouble spots. Previous experimental work has verified not only the existence of trouble spots in a range of wireless networks, but has also identified some of the issues surrounding their detectability. One of the difficulties encountered during the initial study was the collection of reliable data, primarily due to a lack of tool support. To alleviate this, a visualization package, termed RadioTool has been developed; the underlying goal being to investigate the nature of trouble spots more reliably. It is envisaged that the tool will eventually serve as a detection and display mechanism for visualizing network quality as a user moves about a wireless network, and thereby complement and support computational steering applications. This paper describes the features incorporated in RadioTool and describes its use in relation to the exemplar Ricochet radio network.