HASHING STORAGE IMAGES OF A VIRTUAL MACHINE

• US 20100138827A1
• Filed: 11/30/2008
• Published: 06/03/2010
• Est. Priority Date: 11/30/2008
• Status: Active Grant

• Current Assignee: Red Hat Israel Ltd. (International Business Machines Corporation)
• Original Assignee: Red Hat Israel Ltd. (International Business Machines Corporation)
• Inventors: Baron, Ayal, Frank, Shahar
• Granted Patent: US 8,701,106 B2
• Time in Patent Office: Days
• US Class Current: 718/1
• CPC Class Codes: G06F 2009/4557 Distribution of virtual machine instances; Migration and load balancing
• CPC Class Codes: G06F 2009/45579 I/O management, e.g. providing access to device drivers or storage
• 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

Virtual machine images are transferred from a source storage location to a target storage location over a network. In one embodiment, a host at the source storage location computes signature values of a plurality of disk blocks that contain a plurality of virtual machine images. Each computed signature value corresponds to one of the disk blocks. A subset of the disk blocks, all of which have different signature values, is then transferred from the source storage location to the target storage location. Only one copy of duplicate disk blocks is transferred.

Claims

1. A computer-implemented method comprising:

computing signature values, at a source storage location, of a plurality of disk blocks that contain a plurality of virtual machine images, each signature value corresponding to one of the disk blocks; and

transferring a subset of the disk blocks from the source storage location to a target storage location over a network, all of the disk blocks in the subset having different signature values.

2. The method of claim 1, wherein computing signature values further comprises:

identifying disk blocks that have identical signature values from the plurality of disk blocks; and

transferring only one of the disk blocks that have the identical signature values.

3. The method of claim 1, wherein computing signature values further comprises:

hashing each of the plurality of disk blocks to obtain the signature value for the disk block.

4. The method of claim 1, wherein transferring a subset of the disk blocks further comprises:

comparing the signature values computed from the plurality of the disk blocks with base signature values that are computed from a base image of a virtual machine; and

determining the disk blocks that have the signature values different from the base signature values.

5. The method of claim 1, further comprising:

receiving a list of signature values corresponding to shared blocks in a remote pool that resides at the target location; and

sending an update of the remote pool to the target location, the update including data of a shared block that is not currently in the remote pool and an offset at which the shared block is to be placed in the remote pool, the shared block being shared by the virtual machine images.

6. The method of claim 1, wherein computing the signature values further comprises:

determining a differential between a current version of the virtual machine images and a previous version of the virtual machine images; and

computing the signature values for the disk blocks that contain the differential.

7. A system comprising:

data storage to store a plurality of disk blocks that contain a plurality of virtual machine images; and

a host coupled to the data storage to compute a signature value for each of the disk blocks and to transfer a subset of the disk blocks, all of which have different signature values, to a target storage location over a network.

8. The system of claim 7, wherein each of the virtual machine images is a stored image of a virtual machine that runs on the host.

9. The system of claim 7, wherein the host further comprises:

a hash unit to hash each disk block to obtain the signature value for the disk block.

10. The system of claim 7, wherein the host further comprises:

a differential unit to compare the signature values of the disk blocks with base signature values that are computed from a base image of a virtual machine.

11. The system of claim 7, wherein the host further comprises:

a deduplication unit to identify the disk blocks that have identical signature values.

12. A computer readable storage medium including instructions that, when executed by a processing system, cause the processing system to perform a method comprising:

computing signature values, at a source storage location, of a plurality of disk blocks that contain a plurality of virtual machine images, each signature value corresponding to one of the disk blocks; and

transferring a subset of the disk blocks from the source storage location to a target storage location over a network, all of the disk blocks in the subset having different signature values.

13. The computer readable storage medium of claim 12, wherein computing signature values further comprises:

identifying disk blocks that have identical signature values from the plurality of disk blocks; and

transferring only one of the disk blocks that have the identical signature values.

14. The computer readable storage medium of claim 12, wherein computing signature values further comprises:

hashing each of the plurality of disk blocks to obtain the signature value for the disk block.

15. The computer readable storage medium of claim 12, wherein transferring a subset of the disk blocks further comprises:

comparing the signature values computed from the plurality of the disk blocks with base signature values that are computed from a base image of a virtual machine; and

determining the disk blocks that have the signature values different from the base signature values.

16. The computer readable storage medium of claim 12, wherein transferring a subset of the disk blocks further comprises:

receiving a list of base signature values from the target storage location, the base signature values computed from disk blocks that contain a base image of a virtual machine; and

determining the disk blocks that have the signature values different from the base signature values.

17. The computer readable storage medium of claim 12, wherein the method further comprises:

receiving a list of signature values corresponding to shared blocks in a remote pool that resides at the target location; and

sending an update of the remote pool to the target location, the update including data of a shared block that is not currently in the remote pool and an offset at which the shared block is to be placed in the remote pool, the shared block being shared by the virtual machine images.

18. A computer-implemented method, comprising:

receiving, from a source storage location, a reference to a shared block and an off-set at which the shared block is located within a shared pool at a target storage location, the shared block shared by multiple virtual machine images to be transferred from the source storage location to the target storage location; and

receiving blocks of the multiple virtual machine images from the source storage location, all of the blocks being different.

19. The method of claim 18, further comprising:

sending, to the source storage location, a list of signature values of the shared blocks in the shared pool before a next synchronization of the virtual machine images.

20. The method of claim 18, further comprising:

reconstructing each virtual machine image using the received blocks and the shared blocks in the remote pool.