Method for optimizing end-to-end processing performance by selecting optimal values after running benchmarks repeatedly with different values

摘要

A method for benchmarking and optimizing the end-to-end processing performance of a computer network system that identifies performance variables that affect system performance, runs an initial performance test to establish a baseline value for each performance variable and a baseline performance value for the system, pins all performance variables to their constant baseline values except for one floating variable, incrementally changes the value of the floating variable, conducts a new performance test for each changed value, records the values of the system performance and the associated performance variables in a matrix, restores the floating variable to its baseline value, and repeats the prior steps for a new floating variable. The method proceeds until all performance variables have been tested as the floating variable and the matrix is complete. Using the matrix, a computer system designer can understand the performance variables that most affect system performance and can optimize hardware and software configurations.