SYSTEM AND METHOD FOR SYNCHRONIZING TRANSIENT RESOURCE USAGE BETWEEN VIRTUAL MACHINES IN A HYPERVISOR ENVIRONMENT

• US 20110072427A1
• Filed: 09/21/2009
• Published: 03/24/2011
• Est. Priority Date: 09/21/2009
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Garmark, Sten
• Granted Patent: US 8,261,269 B2
• Time in Patent Office: Days
• US Class Current: 718/1
• CPC Class Codes: G06F 12/0253 Garbage collection, i.e. reclamation of unreferenced memory
• CPC Class Codes: G06F 2009/45583 Memory management, e.g. access or allocation
• CPC Class Codes: G06F 8/30 Creation or generation of source code
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects

Abstract

A system and method for synchronizing transient resource usage between virtual machines, e.g. Java Virtual Machines (JVMs), running within a hypervisor. In accordance with an embodiment, the system allows for synchronizing garbage collection and code optimization to reduce transient processor (cpu) and memory usage. In accordance with an embodiment, the system comprises a computer; a hypervisor for executing virtual servers running JVMs; a communication channel between the virtual servers; wherein each virtual server further comprises non-transient and transient memory and a synchronization module. In accordance with an embodiment the synchronization modules schedule garbage collects and code optimizations to minimize overlaps, thereby reducing the peak transient memory and cpu usage and the resulting volatility of transient resource usage within a computer. In accordance with another embodiment, a cloud manager can move virtual servers between computers to optimize computer volatility within a cloud.

Claims

1. A system for synchronizing transient resource usage between virtual servers in a hypervisor environment, comprising:

a computer which includes

a hypervisor running thereon, for executing virtual servers, and

a communication channel, that enables synchronization between the virtual servers;

a plurality of virtual servers executing on the hypervisor, each of which virtual servers includes

a non-transient memory space or heap, for storage and execution of relatively longer-lived processes;

a transient memory space or heap, for storage and execution of relatively shorter-lived processes; and

wherein the virtual servers can advance or postpone execution of those processes that make use of the transient memory space or heap, to synchronize transient resource usage between the plurality of virtual servers.

2. The system of claim 1, wherein each virtual server runs one or more Java Virtual Machines (JVMs).

3. The system of claim 1, wherein one or more of the virtual servers has an operating system tailored to only run a single JVM, and to reduce volatility of the single JVM.

4. The system of claim 1, wherein the non-transient memory includes internal JVM memory and Java Heap memory.

5. The system of claim 1, wherein the transient memory is memory used for garbage collection processes.

6. The system of claim 1, wherein the transient memory is memory used for code optimization processes.

7. The system of claim 1, wherein each of the plurality of virtual servers further comprises a synchronization module which determines predictions of times of garbage collects, code optimizations, or other processes at that virtual server.

8. The system of claim 7, wherein the transient memory is memory used for garbage collection processes, and wherein the synchronization module advances garbage collects to avoid overlaps with other garbage collects.

9. The system of claim 7, wherein the transient memory is memory used for code optimization processes, and wherein the synchronization module postpones code optimizations to avoid overlaps with garbage collects and other code optimizations.

10. The system of claim 1, wherein the number of actual overlaps is used to determine if a computer can host more virtual servers.

11. The system of claim 7, wherein each of the plurality of virtual servers further comprises a synchronization module, and wherein one of the synchronization modules is selected as the master synchronization module, wherein the master synchronization module requests new predictions from the synchronization modules at the other virtual servers, and uses the predictions to update a schedule of garbage collects, code optimizations, or other processes, from an original schedule to a modified schedule.

12. The system of claim 7, wherein the synchronization modules uses the communication channel to communicate information between the plurality of virtual servers, and to synchronize transient resource usage between the virtual servers.

13. The system of claim 1, wherein the communication channel is provided as part of the hypervisor.

14. The system of claim 7, wherein each of the synchronization modules performs the steps of:

waiting for a notification from a virtual server,

requesting garbage collection predictions from all other virtual servers,

detecting and minimizing the number of garbage collection overlaps by advancing garbage collects as necessary,

detecting and minimizing the number of code optimization overlaps by postponing code optimizations as necessary, and

scheduling garbage collects, code optimizations, or other processes to synchronize transient resource usage between the plurality of virtual servers.

15. The system of claim 1, wherein the system is used as a component for controlling the volatility of a cloud of virtual servers, and wherein the system further comprises:

a cloud volatility manager that monitors the volatility of computers within the cloud;

a plurality of computers participating in the cloud, each of which one or more computers further includes

virtual servers with measured resource usage volatility, and configured to synchronize transient resource usage between the plurality of virtual servers; and

wherein the cloud volatility manager can switch or move virtual servers from a first of the plurality of computers participating in the cloud, to a second of the plurality of computers participating in the cloud, to improve volatility in the cloud.

17. The method of claim 1, wherein each of the plurality of virtual servers further comprises a synchronization module which determines predictions of times of garbage collects, code optimizations, or other processes at that virtual server, and wherein each of the synchronization modules performs the steps of:

waiting for a notification from a virtual server,

requesting garbage collection predictions from all other virtual servers,

detecting and minimizing the number of garbage collection overlaps by advancing garbage collects as necessary,

detecting and minimizing the number of code optimization overlaps by postponing code optimizations as necessary, and

scheduling garbage collects, code optimizations, or other processes to synchronize transient resource usage between the plurality of virtual servers.

18. The method of claim 1, wherein the method is used as a component for controlling the volatility of a cloud of virtual servers, and wherein the method further comprises:

providing a cloud volatility manager that monitors the volatility of computers within the cloud;

providing a plurality of computers participating in the cloud, each of which one or more computers further includes virtual servers with measured resource usage volatility, and configured to synchronize transient resource usage between the plurality of virtual servers; and

wherein the cloud volatility manager can switch or move virtual servers from a first of the plurality of computers participating in the cloud, to a second of the plurality of computers participating in the cloud, to improve volatility in the cloud.

16. A method of synchronizing transient resource usage between virtual servers running in a hypervisor environment, comprising the steps of:

providing a computer which includes

a hypervisor running thereon, for executing virtual servers, and

a communication channel, that enables synchronization between the virtual servers;

providing a plurality of virtual servers executing on the hypervisor, each of which virtual servers includes

a non-transient memory space or heap, for storage and execution of relatively longer-lived processes;

a transient memory space or heap, for storage and execution of relatively shorter-lived processes; and

wherein the virtual servers can advance or postpone execution of those processes that make use of the transient memory space or heap, to synchronize transient resource usage between the plurality of virtual servers.

19. A computer readable medium, including instructions stored thereon which, when read and executed by one or more computers, causes the one or more computers to perform the method comprising the steps of:

providing a computer which includes

a hypervisor running thereon, for executing virtual servers, and

a communication channel, that enables synchronization between the virtual servers;

providing a plurality of virtual servers executing on the hypervisor, each of which virtual servers includes

a non-transient memory space or heap, for storage and execution of relatively longer-lived processes;

a transient memory space or heap, for storage and execution of relatively shorter-lived processes; and

wherein the virtual servers can advance or postpone execution of those processes that make use of the transient memory space or heap, to synchronize transient resource usage between the plurality of virtual servers.