System for on-line service in which gateway computer uses service map which includes loading condition of servers broadcasted by application servers for load balancing

In a computer network which comprises a gateway computer that connects a wide area network to a local area network, a method of processing service requests from users so as to balance a processing load among application servers of the network, the application servers interconnected by the local area network, the method comprising the steps of:

• receiving a service request at a host data center, said service request transmitted from client computer to said host data center over said wide area network, said service request corresponding to a service;
• identifying a group of application servers that are assigned to said service, said step of identifying comprising accessing a service map stored by said gateway computer, said service map automatically updated by said application servers;
• determining processing loads of each application server in said group;
• based on said processing loads, selecting an application server from said group to process said service request;
• and
• routing said service request to a selected application server.

A gateway computer, comprising:

• a first network interface for communicating with client computers;
• second network interface for communicating with a plurality of servers;
• a memory, said memory containing load data for said plurality of servers, said load data updated periodically such that said load data indicates current loads for said plurality of servers, said load data indicating, for each server of said plurality of servers, an available processing power for processing requests from client computers;
• and
• a CPU, said CPU programmed to each of:
• (a) receive a request from a client computer,
• (b) in response to said request, access said load data to identify a server of said plurality of servers that is lightly load, and
• (c) route said request to said server identified in step (b).

A data center for providing on-line services to end users who connect to the data center from client computers that run client applications, comprising:

• a plurality of application servers interconnected by a local area network, each application server running at least one service application that implements an on-line service, different groups of said application servers running different service applications to provide different on-line services, application servers which run like service applications forming service groups;
• and
• a plurality of gateway computers connected to said application servers by said local area network and connected to the client computers by a wide area network, each gateway computer programmed to receive service requests from client computers and to route said service requests to application servers running corresponding service applications, the gateway computers further programmed to establish and maintain bi-directional data connections between the client computers and said application servers over the local area network based on entries within the service map, the data connections corresponding to user service sessions, each gateway computer further programmed to, upon a service request from a client computer to open a user-specified service, doing each of:
• (a) identifying all application servers that fall within a service group which correspond to said user-specified service,
• (b) determining respective processing loads for each application server identified in step (a), the step of determining comprising accessing a service map, said service map indicating processing loads for substantially all of said application servers, including application servers of multiple different service groups, said service map automatically updated by said application servers to reflect changes in processing loads, and
• (c) based on processing loads determined in step (b), selecting an application server from said service group to process said request to open said user-specified service.