Method and system for load balancing and management

A method for balancing load among a plurality of servers on a computer system comprising the steps of:

• obtaining a value for current resource usage for each server;
• obtaining a value for expected future resource usage for each server;
• for each server, determining a load balancing factor from the value for current resource usage and the value for future resource usage;
• and
• assigning sessions to the plurality of servers according to the relative values of the load balancing factors for each server;
• wherein the step of determining a load balancing factor includes determining a load balancing factor from the value for current resource usage, the value for future resource usage, and at least one prior value for current resource usage, and further includes obtaining a first probability factor for each server from the value for current resource usage and at least one prior value for current resource usage, obtaining a second probability factor for each server from the value for future resource usage, and determining a combined probability factor for each server from the first probability factor and the second probability factor.

A method for balancing load among a plurality of servers on a computer system comprising the steps of:

• obtaining a value for current resource usage for each server;
• obtaining a value for expected future resource usage for each server;
• for each server, determining a load balancing factor from the value for current resource usage and the value for future resource usage;
• and
• assigning sessions to the plurality of servers according to the relative values of the load balancing factors for each server;
• wherein the step of obtaining a value for current resource usage for each server includes obtaining a measure of current resource usage from each server;
• and if the measure of current resource usage is less than a first minimum usage value, assigning as the value for current resource usage the first minimum usage value, and otherwise assigning as the value for current resource usage the measure of current resource usage.

A method for balancing load among a plurality of servers on a computer system comprising the steps of:

• obtaining a value for current resource usage for each server;
• obtaining a value for expected future resource usage for each server;
• for each server, determining a load balancing factor from the value for current resource usage and the value for future resource usage;
• and
• assigning sessions to the plurality of servers according to the relative values of the load balancing factors for each server;
• wherein the step of determining for each server a load balancing factor includes assigning a load balancing factor for a server so as to preclude assigning sessions to the server if the value for expected future resource usage for the server exceeds a future maximum usage value.

A method for balancing load among a plurality of servers on a computer system comprising the steps of:

• obtaining a value for current resource usage for each server;
• obtaining a value for expected future resource usage for each server;
• for each server, determining a load balancing factor from the value for current resource usage and the value for future resource usage;
• and
• assigning sessions to the plurality of servers according to the relative values of the load balancing factors for each server;
• wherein the step of determining for each server a load balancing factor includes assigning a load balancing factor for a server so as to preclude assigning sessions to the server if all of the threads available to the server are handling requests and the server has not received or sent a packet within a configured time interval.