METHODS AND SYSTEMS FOR IMPLEMENTING TRANSCENDENT PAGE CACHING

• US 20100186011A1
• Filed: 03/03/2010
• Published: 07/22/2010
• Est. Priority Date: 01/20/2009
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Magenheimer, Daniel
• Granted Patent: US 8,769,206 B2
• Time in Patent Office: Days
• US Class Current: 718/1
• CPC Class Codes: G06F 12/084 with a shared cache
• CPC Class Codes: G06F 2009/4557 Distribution of virtual machine instances; Migration and load balancing
• CPC Class Codes: G06F 2009/45583 Memory management, e.g. access or allocation
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects

Abstract

This disclosure describes, generally, methods and systems for implementing transcendent page caching. The method includes establishing a plurality of virtual machines on a physical machine. Each of the plurality of virtual machines includes a private cache, and a portion of each of the private caches is used to create a shared cache maintained by a hypervisor. The method further includes delaying the removal of the at least one of stored memory pages, storing the at least one of stored memory pages in the shared cache, and requesting, by one of the plurality of virtual machines, the at least one of the stored memory pages from the shared cache. Further, the method includes determining that the at least one of the stored memory pages is stored in the shared cache, and transferring the at least one of the stored shared memory pages to the one of the plurality of virtual machines.

Claims

1. A computer implemented method of implementing transcendent page caching, the method comprising:

establishing a plurality of virtual machines on a physical machine, wherein each of the plurality of virtual machines includes a private cache, and wherein a portion of each of the private caches is used to create a shared cache maintained by a hypervisor, wherein the hypervisor is configured to manage physical resources of the physical machine, divide and assign the physical resources to the plurality of virtual machines as specified by a system administrator, creates an illusion for an operating environment running in each of the plurality of virtual machines that each is running on a separate physical machine, and isolate the virtual machines from each other to ensure no one virtual machine interferes with another virtual machine;

storing memory pages in the plurality of virtual machines'"'"' private caches;

determining that at least one of the stored memory pages has been designated to be removed from one of the plurality of virtual machines'"'"' private caches;

delaying the removal of the at least one of the stored memory pages;

storing the at least one of the stored memory pages in the shared cache;

requesting, by one of the plurality of virtual machines, the at least one of the stored memory pages from the shared cache;

determining that the at least one of the stored memory pages is stored in the shared cache; and

transferring the at least one of the stored shared memory pages to the one of the plurality of virtual machines.

2. The method of claim 1, wherein the shared cache is created using idle portions of the plurality of virtual machines'"'"' private caches.

3. The method of claim 1, further comprising determining that the request to store the at least one memory page in the shared byte cache is denied.

4. The method of claim 3, further comprising dropping the at least one memory page from the one of the plurality of virtual machines'"'"' private caches.

5. The method of claim 1, further comprising determining that the requested at least one of the stored memory pages is not stored in the shared cache.

6. The method of claim 5, accessing a disk storage to retrieve the requested at least one of the stored memory pages in response to the requested at least one of the stored memory pages not being stored in the shared cache.

7. The method of claim 1, wherein the hypervisor is configured to control access by the plurality of virtual machines to the shared cache.

8. The method of claim 7, wherein the hypervisor is further configured to determine the shared cache'"'"'s size allocation.

9. The method of claim 8, further comprising reallocating, by the hypervisor, the shared cache'"'"'s size, based at least in part on the plurality of the virtual machines'"'"' private caches'"'"' level of idleness.

10. The method of claim 1, further comprising determining if a private copy of the requested at least one of the stored memory pages in the shared cache has changed so that it differs from the stored memory pages in the shared cache.

11. The method of claim 10, further comprising purging the requested at least one of the stored memory pages from the shared cache, in response to a private copy of the requested at least one of the stored memory pages being changed.

12. The method of claim 10, further comprising retrieving the requested at least one of the stored memory pages from the shared cache, in response to the requested at least one of the stored memory pages from the shared cache not having a private changed copy.

13. The method of claim 1, wherein the hypervisor performs one or more of the following operations:

a shared cache put, a shared cache get, and a shared cache flush.

14. The method of claim 1, further comprising deleting, by one of the plurality of virtual machines, at least one memory page from the at least one of the virtual machine'"'"'s storage which is also stored in the shared cache.

15. The method of claim 14, further comprising calling the hypervisor to flush the at least one memory page from the shared cache, in response to the one of the plurality of virtual machines deleting the at least one memory page.

16. The method of claim 1, further comprising limiting, by the hypervisor, each of the plurality of virtual machines'"'"' usage of the shared cache.

17. The method of claim 16, wherein the limiting is based at least in part on one or more of the following:

a total amount of idle memory, each of the virtual machines'"'"' shared cache usage history, a number of virtual machines, and a total number of available physical memory.

18. A machine-readable medium for implementing transcendent page caching, the machine-readable medium including sets of instructions stored thereon which, when executed by a machine, cause the machine to:

establish a plurality of virtual machines on a physical machine, wherein each of the plurality of virtual machines include a private cache, and wherein a portion of each of the private caches is used to create a shared cache maintained by a hypervisor wherein the hypervisor is configured to manage physical resources of the physical machine, divide and assign the physical resources to the plurality of virtual machines as specified by a system administrator, creates an illusion for an operating environment running in each of the plurality of virtual machines that each is running on a separate physical machine, and isolate the virtual machines from each other to ensure no one virtual machine interferes with another virtual machine;

store memory pages in the plurality of virtual machines'"'"' private caches;

determine that at least one of the stored memory pages has been designated to be removed from one of the plurality of virtual machines'"'"' private caches;

delay the removal of the at least one of stored memory pages;

store the at least one of stored memory pages in the shared cache;

request, by one of the plurality of virtual machines, the at least one of the stored memory pages from the shared cache;

determine that the at least one of the stored memory pages is stored in the shared cache; and

transfer the at least one of the stored shared memory pages to the one of the plurality of virtual machines.

19. The machine-readable medium of claim 18, wherein the shared cache is created using idle portions of the plurality of virtual machines'"'"' private caches.

20. The machine-readable medium of claim 18, further comprising determining that the request to store the at least one memory page in the shared byte cache is denied.