A Predictive Model for Frequently Viewed Tiles in a Web Map |
| |
| Authors: | Sterling Quinn Mark Gahegan |
| |
| Institution: | 1. Department of Geography Pennsylvania State University and ESRI, Redlands, California;2. School of Environment and Centre for eResearch, The University of Auckland, New Zealand |
| |
| Abstract: | Distributing pre‐generated image tiles from a server‐side cache is a fast and popular way of serving maps on the Web. However, these caches can grow unmanageably large for maps covering many scales over a broad extent. This research presents a model for determining high‐priority geographic areas for tile caching. The model considers variables found in previous research to be of interest to Web map users, such as populated places, major roads, coastlines, and tourist attractions. The proportion of area, disk space, and time saved when creating a partial cache (consisting only of high‐interest areas) and a cache of the entire map extent are compared to establish that marked savings are possible. Several applications of the model beyond tile caching are briefly discussed. |
| |
| Keywords: | |
|
|
