Despite the tremendous growth in the importance and reach of the Web, users unfortunately have little recourse when a Web page fails to load. Web browsers currently provide little feedback on such failures, and typically suggest re-checking the URL or the machine’s network settings. As a result, users are often unable to accurately diagnose Web access problems, and resort to heuristic attempts at resolution, such as haphazardly modifying their settings or simply trying again later. Our system, WebProfiler, leverages end-host cooperation to pool together observations on the success or failure of individual Web accesses from multiple vantage points. These are fed into a simple, collaborative blame attribution algorithm which, as our evaluation on a controlled testbed shows, can accurately diagnose 3.6 times as many failures than would be possible from a single client’s perspective. We present the design of WebProfiler and our prototype implementation, which targets Web failure diagnosis in an enterprise network. We also discuss the additional challenges involved in extending WebProfiler to accommodate clients spread in the wide-area Internet.
Source: Microsoft Research
This entry was posted
on Friday, May 16th, 2008 and is filed under Search News, Technology and Internet, Web Search.
You can follow any responses to this entry through the RSS 2.0 feed.
Responses are currently closed, but you can trackback from your own site.