Design of Client-side Caching and Region Prefetching Algorithms for Internet GIS

One of the critical problems of Internet GIS is latency time of displaying the map requested by clients due to network traffic. To enhance the user's response time, it is required to minimize the number of interactions between a client and a server. The minimization of network traffic can be achieved by data caching and object prefetching. For Internet GIS applications, the operators we consider are browsing, panning, zooming-in/out, and displaying a user's requested map. From an implementation point of view, they do have certain similarities. because in every case it needs to browse the map on the basis of some proximity of spatial data. The main idea of this work is to use spatial-temporal locality for designing the replacement scheme. This paper presents new replacement strategies in client-side caching and prefetching, named proximity-based replacement strategy and region prefetch policy respectively. Cache replacement strategy is on the basis of agingconcept that the oldest cell must be replaced when local cache is full. Proximity-based replacement strategy increases the age of all cells according to their proximity-distances. The aging of all cached cells (a cell: a unit of cached data in memory), increases in proportion to the distance between each cell and the region displayed on the current window. The unit of accessing spatial objects in Internet GIS clients is not a data page, but a region. The unit of prefetching spatial objects for browsing the map should be a region in order to reduce the number of interactions between a client and a server.