Efficient resource utilization in data centers

Patent metadata

Field	Value
Publication number	US-9823948-B2
Application number	US-201514948426-A
Country	US
Kind code	B2
Filing date	Nov 23, 2015
Priority date	Jan 31, 2014
Publication date	Nov 21, 2017
Grant date	Nov 21, 2017

Abstract

A method includes identifying high-availability jobs and low-availability jobs that demand usage of resources of a distributed system. The method includes determining a first quota of the resources available to low-availability jobs as a quantity of the resources available during normal operations, and determining a second quota of the resources available to high-availability jobs as a quantity of the resources available during normal operations minus a quantity of the resources lost due to a tolerated event. The method includes executing the jobs on the distributed system and constraining a total usage of the resources by both the high-availability jobs and the low-availability jobs to the quantity of the resources available during normal operations.

First claim

What is claimed is:

1. A method comprising:
identifying high-availability jobs and low-availability jobs that demand usage of resources of a distributed system;
determining a first quota of the resources available to low-availability jobs as a quantity of the resources available during normal operations;
determining a second quota of the resources available to high-availability jobs as the quantity of the resources available during normal operations minus a quantity of the resources lost due to a tolerated event, the second quota of the resource available to high-availability jobs being determined as:
wherein Q 2 is the second quota, R N is the quantity of the resource available during normal operations, R L is the quantity of the resource lost due to a tolerated event, and x is a resource multiplier corresponding to the tolerated event;

executing the jobs on the resources of the distributed system; and
constraining a total usage of the resources by both the high-availability jobs and the low-availability jobs to the quantity of the resources available during normal operations.

2. The method of claim 1 , wherein the resources comprise data processing devices, networking systems, power systems, or cooling systems.

3. The method of claim 2 , further comprising migrating or re-executing jobs assigned to resources lost due to the tolerated event to remaining resources.

4. The method of claim 1 , wherein the resources comprise non-transitory memory devices.

5. The method of claim 1 , further comprising determining the second quota of the resources available to high-availability jobs as:
Q 2 =R N −R L −R I
wherein Q 2 is the second quota, R N is the quantity of the resources available during normal operations, R L is the quantity of the resources lost due to a tolerated event, and R I is an increased quantity of the remaining resources needed due to the tolerated event.

6. The method of claim 1 , wherein the resource comprises non-transitory memory devices.

7. The method of claim 1 , further comprising limiting a sum of the first quota and the second quota to a maximum quota.

8. The method of claim 1 , further comprising:
monitoring a usage of the resources by the high-availability jobs; and
when a high-availability job exceeds a threshold usage, downgrading the high-availability job to a low-availability job.

9. The method of claim 1 , further comprising suspending or ending at least some of the low-availability jobs for the period of time before the tolerated event.

10. The method of claim 1 , further comprising determining the quantity of the resources lost due to a tolerated event based on an assignment of the jobs to particular resources and a system hierarchy of the distributed system, the system hierarchy comprising system domains, each having an active state or an inactive state.

11. A method comprising:
identifying high-availability jobs and low-availability jobs that demand usage of resources of a distributed system;
determining a first quota of the resources available to low-availability jobs as a quantity of the resources available during normal operations;
determining a quantity of the resources lost due to a tolerated event based on an assignment of the jobs to particular resources and a system hierarchy of the distributed system, wherein the system hierarchy comprises system levels comprising:
a first system level corresponding to host machines of data processing devices, non-transitory memory devices, or network interface controllers, each host machine having a system domain;
a second system level corresponding to power deliverers, communication deliverers, or cooling deliverers of racks housing the host machines, each power deliverer, communication deliverer, or cooling deliverer of the rack having a system domain; and
a third system level corresponding to power deliverers, communication deliverers, or cooling deliverers of cells having associated racks, each power deliverer, communication deliverer, or cooling deliverer of the cell having a system domain; and
a fourth system level corresponding to a distribution center module of the cells, each distribution center module having a system domain;

determining a second quota of the resources available to high-availability jobs as the quantity of the resources available during normal operations minus the quantity of the resources lost due to the tolerated event;
executing the jobs on the resources of the distributed system; and
constraining a total usage of the resources by both the high-availability jobs and the low-availability jobs to the quantity of the resources available during normal operations.

12. A system comprising:
resources of a distributed system; and
a computer processor in communication with the resources, the computer processor:
identifying high-availability jobs and low-availability jobs that demand usage of the resources;
determining a first quota of the resources available to low-availability jobs as a quantity of the resources available during normal operations;
determining a second quota of the resources available to high-availability jobs as the quantity of the resources available during normal operations minus a quantity of the resources lost due to a tolerated event, the second quota of the resource available to high-availability jobs being determined as:
wherein Q 2 is the second quota, R N is the quantity of the resource available during normal operations, R L is the quantity of the resource lost due to a tolerated event, and x is a resource multiplier corresponding to the tolerated event;
executing the jobs on the resources of the distributed system; and
constraining a total usage of the resources by both the high-availability jobs and the low-availability jobs to the quantity of the resources available during normal operations.

13. The system of claim 12 , wherein the resources comprise data processing devices, networking systems, power systems, or cooling systems.

14. The system of claim 13 , wherein the computer processor migrates or re-executes jobs assigned to resources lost due to the tolerated event to remaining resources.

15. The system of claim 12 , wherein the resources comprises non-transitory memory devices.

16. The system of claim 12 , wherein the computer processor determines the second quota of the resources available to high-availability jobs as:
Q 2 =R N −R L −R I
wherein Q 2 is the second quota, R N is the quantity of the resources available during normal operations, R L is the quantity of the resources lost due to a tolerated event, and R I is an increased quantity of the remaining resources needed due to the tolerated event.

17. The system of claim 12 , wherein the resources comprise non-transitory memory devices.

18. The system of claim 12 , wherein the computer processor limits a sum of the first quota and the second quota to a maximum quota.

19. The system of claim 12 , wherein the computer processor:
monitors a usage of the resources by the high-availability jobs; and
when a high-availability job exceeds a threshold usage, downgrades the high-availability job to a low-availability job.

20. The system of claim 12 , wherein the computer processor suspends or ends at least some of the low-availability jobs for the period of time before the tolerated event.

21. The system of claim 12 , wherein the computer processor determines the quantity of the resources lost due to a tolerated event based o

Assignees

• Google Inc

Inventors

• Cypher Robert
• Dahl Peter
• Schirripa Steven Robert

Classifications

• G06F9/50 Primary

  Allocation of resources, e.g. of the central processing unit [CPU] · CPC title

• G06F9/4881

  Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues · CPC title

• G06F9/5088

  involving task migration · CPC title

• G06F9/5011 Primary

  the resources being hardware resources other than CPUs, Servers and Terminals · CPC title

• G06F2209/503

  Resource availability · CPC title