Caching is probably the most effective optimization tool to speed up the services' response time. When executing a particular request, the GIS server spends almost all of the execution time connecting to the database, indexing, querying, retrieving records, geoprocessing, projecting, and writing the map to an image, which is finally returned. You may notice that all these operations are database related. So if you could eliminate the database factor, you could save a huge amount of processing time. Here is where the concept of caching is introduced. If you could generate tiles of images for certain scales and store them locally on the GIS server's physical hard drive, you can simply index a request to a set of tiles and return them immediately without connecting to the database and doing all this overhead work. Caching could slash a big portion of processing time, minimizing the response time to more than 80 percent and consequently increasing successful requests or throughput. Not...