초록 ▼

A load sharing system which minimizes overall costs by assigning segments of a divisible load to distributed processors based on the monetary cost of each processor. The distributed processors are connected to a network such as a local area network or the internet. A controller divides a divisible l

A load sharing system which minimizes overall costs by assigning segments of a divisible load to distributed processors based on the monetary cost of each processor. The distributed processors are connected to a network such as a local area network or the internet. A controller divides a divisible load and assigns each segment of the load to a processor based on the processor's monetary cost which is a function of its operating cost and its processing speed. The allocation is then optimized by transferring portions of the load from the expensive processors to the cheaper processor while maintaining an acceptable finish time for all processing. The optimization can also be performed for minimizing the finish time of the processors within a cost constraint.

대표청구항 ▼

[ We claim:] [1.] A system for distributing an incoming computing job comprising:a plurality of processors, each with an associated monetary cost;a queue for receiving said incoming computing job;a controller for dividing said received computing job into segments and allocating said segments to at l

[ We claim:] [1.] A system for distributing an incoming computing job comprising:a plurality of processors, each with an associated monetary cost;a queue for receiving said incoming computing job;a controller for dividing said received computing job into segments and allocating said segments to at least two of said processors; anda network for connecting said processors and said controller;wherein said controller retrieves associated monetary cost data for said plurality of processors and identifies at least one processor with a low said monetary cost, wherein said allocation of said segments is responsive to said retrieved associated monetary cost data of each said processor and wherein said controller further reallocates said segments to said processors in order to minimize said computing job's overall monetary cost.