Bandwidth-Efficient Virtual Machine Image Delivery

• US 20140325507A1
• Filed: 08/01/2013
• Published: 10/30/2014
• Est. Priority Date: 04/30/2013
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Freimuth, Douglas M., Wong, Ho Yin Starsky, Zafer, Murtaza
• Granted Patent: US 9,311,128 B2
• Time in Patent Office: Days
• US Class Current: 718/1
• CPC Class Codes: G06F 11/14 Error detection or correction of the data by redundancy in operation G06F11/16 takes precedence
• CPC Class Codes: G06F 16/00 Information retrieval; Database structures therefor; File system structures therefor
• CPC Class Codes: G06F 2009/45562 Creating, deleting, cloning virtual machine instances
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects

Abstract

A mechanism is provided for bandwidth-efficient virtual machine image delivery. Responsive to a request to generate a virtual machine (VM) in a node using an existing virtual machine image (VMI) file, a set of file chunks that constitute the VMI file is identified. The set of file chunks are retrieved from within a set of distributed nodes by establishing an optimized plan for retrieving the set of file chunks in a bandwidth-efficient manner. Responsive to retrieving the set of file chunks from within the distributed nodes, the set of file chunks are reassembled into the VMI file for generation of the VM.

Claims

1. A method, in a set of distributed nodes, for bandwidth-efficient virtual machine image delivery, the method comprising:

responsive to a request to generate a virtual machine (VM) in a node using an existing virtual machine image (VMI) file, identifying a set of file chunks that constitute the VMI file;

retrieving the set of file chunks from within the set of distributed nodes, wherein retrieving the set of file chunks from within the set of distributed nodes comprises establishing an optimized plan for retrieving the set of file chunks in a bandwidth-efficient manner; and

responsive to retrieving the set of file chunks from within the set of distributed nodes, reassembling the set of file chunks into the VMI file for generation of the VM.

2. The method of claim 1, wherein the optimized plan is established by the method comprising:

for each file chunk in the set of file chunks, determining whether the file chunk exists in a local storage device;

responsive to the file chunk existing in the local storage device, retrieving the file chunk from the local storage device; and

responsive to the file chunk failing to exist in the local storage device;

identifying one or more locations of the file chunk in one or more other nodes within the set of distributed nodes;

responsive to the file chunk existing in more than one node of the one or more other nodes, choosing an optimal node to retrieve the file chunk from based on one or more of a set of criteria associated with each of the more than one other node; and

responsive to identifying an optimal node, retrieving the file chunk from a storage device associated with the optimal node.

3. The method of claim 2, wherein the set of criteria comprises one or more of bandwidth access to the node, historical data associated with accessing the node, priority of accessing the node, grouping of the file chunk with other file chunks to be retrieved from the node, how the file chunk is stored with other file chunks in the node, or access-control policies associated with accessing file chunks in the one or more other nodes.

4. The method of claim 2, wherein a location of each file chunk failing to exist in the local storage device is identified from a metadata table, wherein the metadata table comprises information about each file chunk, and wherein the information comprises one or more of a chunk-hash, a size, a start-offset in the VMI file, or original file-name where the file chunk belongs.

5. The method of claim 2, wherein a location of each file chunk failing to exist in the local storage device is identified from metadata that is aggregated into the metadata table from the one or more other nodes.

6. The method of claim 5, wherein the metadata is aggregated into the metadata table when a new VMI file is generated, at periodic intervals, or at scheduled intervals.

7. The method of claim 1, wherein the set of file chunks that constitute the VMI file are identified from a manifest file associated with the VMI file.

8. The method of claim 7, wherein the manifest file is generated in at least one of the node where the VM is being generated or another node of the one or more other nodes and wherein, responsive to the manifest file being generated in the other node of the one or more other nodes then the manifest file is aggregated into the node where the VM is being generated.

9-20. -20. (canceled)