| 摘要 |
A method and apparatus that ensures that requests for pages in a particular domain name are routed to the same proxy server by all of a plurality of clients. If, for example, a proxy server has a persistent connection to a server for a domain, all incoming requests for that domain will be sent to the proxy server and will, thus, will be able to take advantage of the persistent connection. Each client contains a proxy table that is periodically updated by one or more of the proxy servers. A proxy table in a client contains an entry corresponding to each proxy server. When a client needs to access a resource through a proxy server, the client truncates the address (e.g., the URL) of the resource. Thus, for example, all addresses in a particular domain name are truncated to the same value. The truncated address is then used to hash into the proxy table in the client and to identify a proxy server. The client sends its request to the identified proxy server. Thus, all requests for a particular domain hash to the same proxy table entry and, hence, to the same proxy server. If the proxy server has opened a persistent connection to the server for the requested domain, the proxy server will be able to take advantage of the persistent connection.
|
