A method and apparatus for a server that includes a file processor that interprets each requested data file, such as a web page, requested by a client in a process analogous to that of a browser application or other requesting application. The file processor initiates the loading of each referenced data item within the requested document in anticipation that the client will make the same requests upon receiving the requested data file. Each referenced data item is loaded into the server cache. When the client browser application requests these referenced data items they can be returned to the client browser application without accessing a slower persistent data storage. The requested data items are loaded from the server cache, which has a faster access time than the persistent data storage.