| 摘要 |
An application service provider provides to remote networked computers service sessions using one of a plurality of similarly functioning software applications residing on different servers with different unique network addresses. To request a service session, a remote computer transmits a "probe request" to a unique network address associated with the service provider. A load balancer then assigns one of the servers to provide the service. The service provider then sends back to the remote computer a "probe response" providing the unique network address of the assigned server. Thereafter, for packet-based messages sent by the remote computer that are part of the service session, the messages are addressed to the unique network address of the assigned server. "Probe request" and "probe response" messages may also be used during the course of a service session for fault tolerance.
|
