![]() ![]() ![]() |
![]() |
|
|
![]() ![]() ![]() ![]() ![]() |
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Return to Distributed, Web-based GIS The research of Internet distribution of GIS contents is still in its infancy. This paper reports an implementation of a systematic approach to optimize Internet distribution of GIS datasets. The goal is to reduce user-perceived response time and improve users' navigation efficiency.On the server side, a large GIS dataset associated with a map is decomposed into small blocks and organized into an overview guiding hierarchy. On the client side, locality-based caching and pre-fetching techniques are devised at the application level. To prevent the GIS application server from being overwhelmed by the excessive pre-fetching requests, it is necessary to enforce a priority queue for the web server: user requests always have high priority than pre-fetching requests. Experiments have shown significant improvement on user-perceived response time. ![]() DiSC'02 © 2003 Association for Computing Machinery |