| 摘要 |
A distributed multiprocessor server system and a set of steps performed by the server system are disclosed that facilitate efficient handling of a potentially large workload arising from request messages received from users via a communicatively coupled network link. Workload requests are initially received by a default processor. The default processor redirects at least a portion of the workload request, based upon the type of request, to a specialized processor in a set of special purpose handle processors. Once delegated by a default processor, the request (or a portion thereof) is completed by the specialized handler processor to which the request was delegated. Communications thereafter proceed between the handler processor and a network interface without passing through the default processor.
|
