Allocation of tasks in large scale computing systems

摘要

Aspects of the invention may be used to allocate tasks among computing machines in large scale computing systems. In one aspect, the method includes executing a first task in the plurality of tasks on a first computing machine and determining a performance degradation threshold for the first task. The method further includes calculating a predicted performance degradation of the first task when a second task is executed on the first computing machine, wherein the predicted performance degradation is determined by comparing a performance interference score of the second task with a performance sensitivity curve of the first task. The method further includes executing the second task on the first computing machine when the predicted performance degradation of the first task is below the performance degradation threshold.

主权项

1. A method of modeling performance degradation of a first application, the method comprising:
executing the first application on a computing machine, wherein the first application consumes a first varying portion of common machine resources of the computing machine; executing a memory expandable application on the computing machine, wherein the memory expandable application consumes a second varying portion of the common machine resources of the computing machine defined according to a variable working set size, wherein the first varying portion of the common machine resources is separate from the second varying portion of the common machine resources, and wherein the memory expandable application is an application that is programmed to incrementally vary its own working size set by varying a number of operations executed by the memory expandable application; causing changes to a size of the first varying portion of the common machine resources available to the first application by enlarging or reducing the second varying portion of resources in the common machine resources consumed by the memory expandable application; measuring a performance metric of the first application for each working set size; and plotting the performance metric of the first application as a function of the working set size of the memory expandable application.