Dynamic resource scheduling
View Patent ↗Embodiments of the invention relate to a system and method for dynamically scheduling resources using policies to self-optimize resource workloads in a data center. The object of the invention is to allocate resources in the data center dynamically corresponding to a set of policies that are configured by an administrator. Operational parametrics that correlate to the cost of ownership of the data center are monitored and compared to the set of policies configured by the administrator. When the operational parametrics approach or exceed levels that correspond to the set of policies, workloads in the data center are adjusted with the goal of minimizing the cost of ownership of the data center. Such parametrics include yet are not limited to those that relate to resiliency, power balancing, power consumption, power management, error rate, maintenance, and performance.
1. A method for dynamically distributing workloads across a plurality of computers in a data center, the method comprising:
identifying one or more policy rules, wherein the one or more policy rules correspond to one or more parametrics in the data center, the one or more parametrics including an identification of a resource platform type and a power performance ratio of a resource in the data center;
selecting one or more workload profiles from a set of established workload profiles;
monitoring power consumption of the resource and a measure of a workload performed by the resource, wherein the workload corresponds to a workload type and the workload type corresponds to operations associated with the platform type;
identifying a current measure of the power performance ratio;
identifying that power consumed by the resource has increased geometrically when the operating frequency of the resource has not increased geometrically according to the current measure of the power performance ratio; and
adjusting the workloads in the data center based on the one or more policy rules and the one or more workload profiles of the set of established workload profiles, wherein at least one of the one or more parametrics corresponds to one or more data center power consumption levels, and at least one of the one or more policy rules correspond to one or more data center power consumption thresholds.
2. The method of claim 1 , wherein the one or more policy rules correspond to at least one of the one or more selected workload profiles.
3. The method of claim 1 , wherein total data center workload is decreased when the total data center power consumption is above a maximum power consumption threshold.
4. The method of claim 1 , wherein the at least one of the one or more parametrics corresponding to the data center power consumption levels includes the power consumption of one or more individual computing resources in the data center.
5. The method of claim 4 , wherein the workload running on a particular resource is reduced when the power consumption of that particular resource is above or approaches a maximum power consumption threshold for that particular resource.
6. The method of claim 1 , wherein at least one of the one or more parametrics corresponds to one or more data center resiliency levels.
7. The method of claim 6 , wherein at least one of the one or more policy rules correspond to one or more data center resiliency thresholds.
8. The method of claim 7 , wherein at least one of the one or more parametrics corresponding to the data center resiliency levels includes error rate, temperature, fan speed, or latency.
9. The method of claim 8 , wherein the workload running on a particular resource is reduced when the resiliency of that particular resource is above a maximum resiliency threshold for that particular resource.
10. The method of claim 1 , wherein the at least one of the one or more parametrics corresponds to one or more data center performance levels.
11. The method of claim 10 , wherein the at least one of the one or more policy rules corresponds to one or more data center performance thresholds.
12. The method of claim 11 , further comprising: identifying an error rate associated with the resource; and
identifying, that a corrective action should be performed based on the identified error rate, wherein:
the corrective action includes at least one of:
reducing a number of operations per unit time that are allocated to the resource,
migrating at least a portion of the operations allocated to the resource to another resource, and
adjusting a fan speed at the resource, and the corrective action is performed.
13. The method of claim 11 , wherein the workload performed by the resource is increased when the power performance ratio is below a minimum power performance threshold, and the workload performed by the resource is decreased when the power performance ratio is above the maximum power performance threshold.
14. A non-transitory computer readable storage medium having embodied thereon a program executable by a processor to perform a method, the method comprising:
identifying one or more policy rules, wherein the one or more policy rules correspond to one or more parametrics in the data center, the one or more parametrics including an identification of a resource platform type and a power performance ratio of a resource in the data center;
selecting one or more workload profiles from a set of established workload profiles;
monitoring power consumption of the resource and a measure of a workload performed by the resource, wherein the workload corresponds to a workload type and the workload type corresponds to operations associated with the platform type;
identifying a current measure of the power performance ratio; identifying that power consumed by the resource has increased geometrically when the operating frequency of the resource has not increased geometrically according to the current measure of the power performance ratio; and
adjusting the workloads in the data center based on the one or more policy rules and the one or more workload profiles of the set of established workload profiles, wherein at least one of the one or more parametrics corresponds to one or more data center power consumption levels, and at least one of the one or more policy rules correspond to one or more data center power consumption thresholds.
15. The non-transitory computer readable storage medium method of claim 14 , wherein the one or more policy rules correspond to at least one of the one or more selected workload profiles.
16. The non-transitory computer readable storage medium method of claim 14 , wherein the at least one of the one or more parametrics corresponding to the data center power consumption levels includes the power consumption of one or more individual resources in the data center, and wherein the workload running on a particular resource is reduced when the power consumption of that particular resource is above or approaches a maximum power consumption threshold for that particular resource.
17. The non-transitory computer readable storage medium method of claim 14 , wherein at least one of the one or more parametrics corresponds to one or more data center resiliency levels, and wherein at least one of the one or more policy rules correspond to one or more data center resiliency thresholds.
18. The non-transitory computer readable storage medium method of claim 17 , wherein at least one of the one or more parametrics corresponding to the data center resiliency levels includes error rate, temperature, fan speed, or latency, and wherein the workload running on a particular resource is reduced when the resiliency of that particular resource is above or approaches a maximum resiliency threshold for that particular resource.
19. A system comprising:
a plurality of compute nodes, wherein a compute node of the plurality of compute nodes is associated with a platform type and the platform type is associated with a resource;
a memory in each of the compute nodes;
a processor in each of the computer nodes;
one or more policy rules stored in the memory, wherein the one or more policy rules correspond to one or more parametrics in a data center; and
one or more modules stored in memory and executable by the processor to select workload profiles from a set of established workload profiles, the one or more modules monitoring the one or more parametrics in the data center and adjusting the workloads in the data center based on the one or more policy rules and the one or more workload profiles, wherein:
power consumption of the resource and a measure of a workload performed by the resource are monitored,
the workload performed by the resource corresponds to a workload type,
the workload type corresponds to operations associated with the platform type, and
a current measure of power performance ratio is identified,
at least one of the one or more parametrics corresponding to one or more data center power consumption levels, the processor identifies that power consumed by the resource has increased geometrically when the operating frequency of the resource has not increased geometrically according to the current measure of the power performance ratio, and at least one of the one or more policy rules corresponding to one or more data center power consumption thresholds.
20. The system of claim 19 , wherein at least one of the one or more policy rules correspond to at least one of the one or more selected workload profiles.