Managing a workload in a cluster of computing systems with multi-type operational resources

1. A method for managing a workload in a plurality of computing systems, each of said plurality of computing systems comprising one or more general resources and one or more specialized resources, the method comprising:

determining, by a computer, parameters characteristic of a performance of said each of said plurality of computer systems, said parameters comprising a crossover factor for each of said one or more specialized resources and each of said one or more general resources;

calculating, by said computer, an equivalent capacity (ECP), said ECP being a measure of a processing capacity of a computing system, said calculating comprising:

based on said parameters and a test workload, exhausting said one or more specialized resources, and assigning a remaining workload of said test workload to said one or more general resources to determine utilization of said one or more general resources; and

wherein said calculating of said ECP is based on said utilization of said one or more general resources, and is based on said crossover factor for said each of said one or more specialized resources and said each of said one or more general resources, wherein said crossover factor is based on a cost of said assigning of said remaining workload to said one or more general resources; and

based on said ECP for said each of said plurality of computing systems, assigning, by a computer, said workload to an appropriate computing system selected from said plurality of computing systems.

2. The method according to claim 1 , wherein said parameters comprises a displaceable capacity for said each of said one or more specialized resources and said each of said one or more general resources.

3. The method according to claim 2 , wherein the displaceable capacity is determined at a given importance level.

4. The method according to claim 1 , wherein said parameter comprises a relative usage for said each of said one or more specialized resources and said each of said one or more general resources, wherein the relative usage is determined with respect to a service class of work executing on the computer system.

5. The method according to claim 4 , wherein the relative usage is periodically measured.

6. The method according to claim 1 , wherein said assigning said remaining workload of said test workload to said one or more general resources further comprises:

terminating said remaining workload assignment once said one or more general resources is exhausted.

7. The method according to claim 6 further comprising assigning a said crossover factor whenever work is assigned to said one or more general resources instead of said one or more specialized.

8. The method according to claim 7 , wherein the crossover factor is specific to a type of said one or more specialized resources.

9. The method according to claim 1 further comprising calculating another equivalent capacity (AECP) of a computing system in a heterogeneous architecture environment by multiplying said system's ECP with a factor depending on said system's relative architecture cost (AC).

10. A system for managing a workload in a plurality of computing systems, each of said plurality of computing systems comprising one or more general resources and one or more specialized resources, comprising:

a memory;

a processor in communications with said memory, said processor capable of performing a method comprising:

determining parameters characteristic of a performance of said each of said plurality of computer systems, said parameters comprising a crossover factor for each of said one or more specialized resources and each of said one or more general resources;

calculating an equivalent capacity (ECP), said ECP being a measure of a processing capacity of a computing system, said calculating comprising:

based on said parameters and a test workload, exhausting said one or more specialized resources, and assigning a remaining workload of said test workload to said one or more general resources to determine utilization of said one or more general resources; and

wherein said calculating of said ECP is based on said utilization of said one or more general resources, and is based on said crossover factor for said each of said one or more specialized resources and said each of said one or more general resources, wherein said crossover factor is based on a cost of said assigning of said remaining workload to said one or more general resources; and

based on said ECP for said each of said plurality of computing systems, assigning, by a computer, said workload to an appropriate computing system selected from said plurality of computing systems.

11. The system according to claim 10 , wherein said parameters comprises a displaceable capacity of said each of said one or more specialized resources and said each of said one or more general resources.

12. The system according to claim 11 , wherein the displaceable capacity is determined at a given importance level.

13. The system according to claim 10 , wherein said parameter comprises a relative usage for said each of said one or more specialized resources and said each of said one or more general resources, wherein the relative usage is determined with respect to a service class of work executing on the computer system.

14. The system according to claim 13 , wherein the relative usage is periodically measured.

15. The system according to claim 10 , wherein said assigning said remaining workload of said test workload to said one or more general resources further comprises:

terminating said remaining workload assignment once said one or more general resources is exhausted.

16. The system according to claim 15 further comprising assigning said crossover factor whenever work is assigned to said one or more general resources instead of a said one or more specialized.

17. The system according to claim 16 , wherein the crossover factor is specific to a type of said one or more specialized resources.