Topology aware grid services scheduler architecture

A computer system for processing client requests, said system comprising:

• a grid services scheduler connected to a plurality of grid container arrays, wherein each grid container array comprises:
• a plurality of persistent containers, each container residing in a computer host;
• and
• one micro scheduler, wherein, within each grid container array, said micro scheduler is topologically positioned in a container that has the highest performance of all containers in said grid container array,
• wherein each of said containers includes a plurality of objects,
• wherein objects within containers that make up a grid container array comprise a grid object array,
• wherein said grid services scheduler divides a client request into a plurality of tasks and assigns groups of said tasks to each micro scheduler,
• wherein each of said microschedulers assigns individual tasks from a group of tasks received from said grid services scheduler to objects within a corresponding grid object array,
• wherein each said grid container array resides in a local area network, such that communications between objects within each said grid container array comprises local communications, and
• wherein said containers comprise persistent service containers and are used to process multiple client requests of multiple clients over a time period that spans multiple client requests.

A method of processing client requests over a computer network of hosts, said method comprising:

• creating persistent containers in said hosts;
• creating objects in said containers;
• grouping said containers into grid container arrays, wherein each grid container array resides in a local area network, such that communications between objects within each said grid container array comprises local communications;
• grouping objects within containers that are within a single grid container array into grid object arrays;
• creating one micro scheduler for each grid container array;
• dividing each of said requests into a plurality of tasks;
• assigning groups of said tasks to said microschedulers, wherein said microschedulers assign individual tasks to said object, and
• wherein said containers comprise service containers and are used to process multiple client requests of multiple client over a time period that spans multiple client requests.

A method of processing client requests over a computer network of hosts, said method comprising:

• creating persistent containers in said hosts;
• creating objects in said containers;
• grouping said containers into grid container arrays, wherein each grid container array resides in a local area network, such that communications between objects within each said grid container array comprises local communications;
• grouping objects within containers that are within a single grid container array into grid object arrays;
• creating one micro scheduler for each grid container array;
• topologically positioning said micro scheduler in a container that has the highest performance of tall containers in said grid container array;
• dividing each of said requests into a plurality of tasks;
• assigning groups of said tasks to said microschedulers, wherein said microschedulers assign individual tasks to said object, and
• wherein said containers comprise persistent service containers and are used to process multiple client requests of multiple client over a time period that spans multiple client requests.