| 摘要 |
The present invention relates to a server computer connectable to a network and having a plurality of processors (20, 30, 40) arranged to provide a service to one or more client computers connected to the network. The service involves the provision of blocks of information to a client computer (50), and the processors have access to a data storage (120) in which data for the blocks of information is stored. According to preferred embodiments of the invention, each processor in the server has block retrieval means (80), employed in response to a request from a client computer (50) for a particular block of information from that processor, to retrieve data from the data storage (120) in order to construct a file representing the requested block of information, the file including references to subsequent blocks of information requestable by the client (50). Additionally, load determining means (100) is provided of periodically determining activity data for the processor for inclusion in a load distribution record maintained for all of the processors of the server. Decision means (90) is then used which, for each reference to a subsequent block of information in the file constructed by the block retrieval means (80), is arranged to determine, based on the load distribution record, which processor should service a request from the client computer (50) for that subsequent block of information, and include an address for that processor in the file constructed by the block retrieval means (80). Once the decision means has completed this process, the file is then sent to the client computer (50). In preferred embodiments, the service being provided is the World Wide Web (WWW) service, and the technique is employed on a parallel Web server providing Web pages to clients over the Internet. By employing the technique of the invention, more effective load balancing across the processors of the Web server is possible, thereby enabling problems such as "toasting" of a particular processor to be averted.
|
