MEMORY MANAGEMENT FOR VIRTUAL MACHINES

• US 20150339166A1
• Filed: 03/04/2015
• Published: 11/26/2015
• Est. Priority Date: 05/20/2014
• Status: Active Application

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Hogstrom, Matt R., Shah, Mihir R., Vouk, Nikola, Yadav, Meeta
• Application Number:
  US14/637,428
• Publication Number:
  US 20150339166A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 12/02 Addressing or allocation; Relocation program address sequencing G06F9/00; arrangements for selecting an address in a digital store G11C8/00
• 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

Embodiments of the disclosure relate to managing a memory of a server hosting a plurality of virtual machines. Aspects include receiving a plurality of data pages from each of the plurality of virtual machines to be stored in the memory, filtering each the plurality of data pages into one of a plurality of pools of data pages including a pool of potentially identical data pages, and evaluating the data pages in the pool of potentially identical data pages to identify one or more duplicate data pages and one or more similar data pages. Aspects further include coalescing data pages identified as duplicate data pages and encoding differences for data pages identified as similar pages.

Claims

1. A method for managing a memory of a server hosting a plurality of virtual machines comprising:

receiving a plurality of data pages from each of the plurality of virtual machines to be stored in the memory;

filtering each the plurality of data pages into one of a plurality of pools of data pages including a pool of potentially identical data pages;

evaluating the data pages in the pool of potentially identical data pages to identify one or more duplicate data pages and one or more similar data pages;

coalescing data pages identified as duplicate data pages; and

encoding differences for data pages identified as similar pages.

2. The method of claim 1, wherein the filtering of each of the plurality of data pages comprises performing a Bloom filter and hash function on each of the plurality of data pages.

3. The method of claim 1, wherein the filtering of each of the plurality of data pages comprises computing a similarity factor for each of the plurality of data pages.

4. The method of claim 3, wherein the filtering comprises placing a data page into a pool of similar data pages if the similarity factor for a data page exceeds a minimum threshold value.

5. The method of claim 1, wherein the plurality of pools of data pages further includes a pool of ones data pages, a pool of zeroes data pages and a pool of pre-defined content data pages.

6. The method of claim 5, further comprising coalescing data pages in the pool of ones data pages and coalescing data pages in the pool of zeroes data pages.

7. The method of claim 1, wherein encoding differences for data pages identified as similar pages incudes storing one of the similar data pages is in the memory and calculating and storing the difference between each remaining similar page and the stored page.

8. The method of claim 7, wherein calculating the difference between the remaining similar pages and the stored page comprises delta encoding.