DYNAMIC MIGRATION OF VIRTUAL MACHINES BASED ON WORKLOAD CACHE DEMAND PROFILING

• US 20120226866A1
• Filed: 03/02/2011
• Published: 09/06/2012
• Est. Priority Date: 03/02/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bozek, James J., Hansson, Nils Peter Joachim, Suffern, Edward S., Wooldridge, James L.
• Granted Patent: US 8,478,961 B2
• Time in Patent Office: Days
• US Class Current: 711/122
• CPC Class Codes: G06F 12/0897 with two or more cache hierarchy levels with multilevel cache hierarchies G06F12/0811
• CPC Class Codes: G06F 2009/4557 Distribution of virtual machine instances; Migration and load balancing
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects
• CPC Class Codes: G06F 9/4856 resumption being on a different machine, e.g. task migration, virtual machine migration G06F9/5088 takes precedence

Abstract

A computer-implemented method comprises obtaining a cache hit ratio for each of a plurality of virtual machines, and identifying, from among the plurality of virtual machines, a first virtual machine having a cache hit ratio that is less than a threshold ratio. The identified first virtual machine is then migrated from the first physical server having a first cache size to a second physical server having a second cache size that is greater than the first cache size. Optionally, a virtual machine having a cache hit ratio that is less than a threshold ratio is identified on a class-specific basis, such as for L1 cache, L2 cache and L3 cache.

Claims

1. A computer-implemented method, comprising:

obtaining a cache hit ratio for each of a plurality of virtual machines;

identifying, from among the plurality of virtual machines, a first virtual machine having a cache hit ratio that is less than a threshold ratio, wherein the first virtual machine is running on a first physical server; and

migrating the first virtual machine from the first physical server having a first cache size to a second physical server having a second cache size that is greater than the first cache size.

2. The computer-implemented method of claim 1, wherein identifying a first virtual machine having a cache hit ratio that is less than a threshold ratio, further comprises:

obtaining a class-specific cache hit ratio for each of a plurality of cache classes available to the first virtual machine on the first physical server; and

comparing, for each of the plurality of cache classes, the class-specific cache hit ratio to the threshold ratio, wherein the first virtual machine is identified as having a cache hit ratio that exceeds the threshold ratio in response to determining that one or more of the class-specific cache hit ratios is less than the threshold ratio.

3. The computer-implemented method of claim 2, wherein the class-specific cache hit ratio for the first virtual machine is the percentage of memory accesses that are available in the particular cache class excluding memory accesses satisfied by any subordinate cache.

4. The computer-implemented method of claim 2, wherein the plurality of cache classes includes L1 cache and L2 cache.

5. The computer-implemented method of claim 2, wherein the plurality of cache classes includes L1 cache, L2 cache and L3 cache.

6. The computer-implemented method of claim 2, wherein comparing, for each of the plurality of cache classes, the class-specific cache hit ratio to the threshold ratio, further comprises:

comparing, for each of the plurality of cache classes, the class-specific cache hit ratio to a class-specific threshold ratio for the same cache class, wherein the first virtual machine is identified as having a cache hit ratio that exceeds the threshold ratio in response to determining that one or more of the class-specific cache hit ratios is less than the class-specific threshold ratio for the same cache class.

7. The computer-implemented method of claim 2, further comprising:

obtaining a cache hit ratio for each of a plurality of virtual machines running on the first physical server.

8. The computer-implemented method of claim 2, further comprising:

selecting the second physical server to receive the first virtual machine from the first physical server, wherein the second physical server is determined to have a greater amount of cache than the first physical server in the cache class having a class-specific cache hit ratio less than the threshold ratio.

9. The computer-implemented method of claim 8, wherein the second physical server is determined to have an equal or greater amount of cache than the first physical server for each cache class that is subordinate to the cache class having a class-specific cache hit ratio less than the threshold ratio.

10. The computer-implemented method of claim 9, wherein the second physical server is determined to have sufficient memory and CPU bandwidth available to run the virtual machine.

11. The computer-implemented method of claim 1, further comprising:

determining an average cache hit ratio for all virtual machines; and

calculating the threshold ratio as being a predetermined percentage below the average cache hit ratio.

12. The computer-implemented method of claim 6, further comprising:

determining a class-specific average cache hit ratio for all virtual machines within a migration domain; and

calculating the class-specific threshold ratio as being a predetermined percentage below the class-specific average cache hit ratio.

13. The computer-implemented method of claim 1, further comprising:

obtaining cache size data for each of the physical servers.

14. The computer-implemented method of claim 13, wherein the cache size data for each physical server is obtained as vital product data of each physical server.

15. The computer-implemented method of claim 1, further comprising:

migrating a second virtual machine off of the second physical server.

16. The computer-implemented method of claim 15, wherein the second virtual machine is determined to have the lowest rate of memory accesses among all the virtual machines running on the second physical server.

17. A computer program product including computer usable program code embodied on a computer usable storage medium, the computer program product including:

computer usable program code for obtaining a cache hit ratio for each of a plurality of virtual machines;

computer usable program code for identifying, from among the plurality of virtual machines, a first virtual machine having a cache hit ratio that is less than a threshold ratio, wherein the first virtual machine is running on a first physical server; and

computer usable program code for migrating the first virtual machine from the first physical server having a first cache size to a second physical server having a second cache size that is greater than the first cache size.

18. The computer program product of claim 17, further comprising:

computer usable program code for obtaining a class-specific cache hit ratio for each of a plurality of cache classes available to the first virtual machine on the first physical server; and

computer usable program code for comparing, for each of the plurality of cache classes, the class-specific cache hit ratio to the threshold ratio, wherein the first virtual machine is identified as having a cache hit ratio that exceeds the threshold ratio in response to determining that one or more of the class-specific cache hit ratios is less than the threshold ratio.

19. The computer program product of claim 18, wherein the class-specific cache hit ratio for the first virtual machine is the percentage of memory accesses that are available in the particular cache class excluding memory accesses satisfied by any subordinate cache.

20. The computer program product of claim 18, wherein the computer usable program code for comparing, for each of the plurality of cache classes, the class-specific cache hit ratio to the threshold ratio, further comprises:

computer usable program code for comparing, for each of the plurality of cache classes, the class-specific cache hit ratio to a class-specific threshold ratio for the same cache class, wherein the first virtual machine is identified as having a cache hit ratio that exceeds the threshold ratio in response to determining that one or more of the class-specific cache hit ratios is less than the class-specific threshold ratio for the same cache class.

21. The computer program product of claim 18, further comprising:

computer usable program code for selecting the second physical server to receive the first virtual machine from the first physical server, wherein the second physical server is determined to have a greater amount of cache than the first physical server in the cache class having a class-specific cache hit ratio less than the threshold ratio.

22. The computer program product of claim 17, further comprising:

computer usable program code for determining that the second physical server has an equal or greater amount of cache than the first physical server for each cache class that is subordinate to the cache class having a class-specific cache hit ratio less than the threshold ratio.

23. The computer program product of claim 17, further comprising:

computer usable program code for determining an average cache hit ratio for all virtual machines; and

computer usable program code for calculating the threshold ratio as being a predetermined percentage below the average cache hit ratio.

24. The computer program product of claim 20, further comprising:

computer usable program code for determining a class-specific average cache hit ratio for all virtual machines within a migration domain; and

computer usable program code for calculating the class-specific threshold ratio as being a predetermined percentage below the class-specific average cache hit ratio.

25. The computer program product of claim 17, further comprising:

computer usable program code for obtaining cache size data for each of the physical servers.