Optimizing virtual machine synchronization for application software

• US 10,261,818 B2
• Filed: 02/10/2015
• Issued: 04/16/2019
• Est. Priority Date: 03/10/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Blythe, Christopher J., Dawson, Michael H., Hrischuk, Curtis E., Quirk, Aaron J.
• Primary Examiner(s): Wu, Benjamin C
• Application Number:
  US14/618,011
• Publication Number:
  US 20150154044A1
• Time in Patent Office: 1,526 Days
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3419 by assessing time
• CPC Class Codes: G06F 11/3442 for planning or managing the needed capacity
• CPC Class Codes: G06F 2009/45591 Monitoring or debugging support
• CPC Class Codes: G06F 2201/815 Virtual middleware or OS functionality using virtual machines to implement generic software techniques for error detection or fault masking G06F11/1484
• CPC Class Codes: G06F 9/45533 Hypervisors; Virtual machine monitors
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects
• CPC Class Codes: G06F 9/50 Allocation of resources, e.g. of the central processing unit [CPU]

Abstract

Real-time application metrics of an application executed by a virtual machine are dynamically monitored by a controlling agent and analyzed to determine an optimal configuration of the virtual machine for executing the application. Based on the measured metrics, tunable parameters of the virtual machine may be adjusted to achieve desired application performance.

Claims

1. A method of optimizing performance of a virtual machine executed by a system that includes a plurality of physical processing units, the method comprising:

measuring, by a controlling agent of the system running outside of the virtual machine that controls the virtual machine, a metric of an application executed by the virtual machine, the metric comprising a hardware metric of the system upon which the virtual machine is executed that is unavailable for measurement by the virtual machine, wherein the hardware metric comprises a utilization rate of physical processing units among the plurality of physical processing units according to a processor affinity of the virtual machine allocated to the virtual machine to which servicing of the application is restricted; and

adjusting a plurality of tunable execution parameters of the virtual machine for optimizing execution of the application based on the measured metric, wherein the plurality of tunable execution parameters comprises an allocation of the plurality of physical processing units according to the processor affinity of the virtual machine, to which servicing of the application is restricted, allocated to the virtual machine and a spin strategy that controls spinning in locks used for synchronization within the virtual machine.

2. The method according to claim 1, wherein the adjusting comprises:

transmitting, by the controlling agent, a request to the virtual machine to adjust the plurality of tunable execution parameters; and

adjusting, by the virtual machine, the plurality tunable execution parameters in response to the transmitted request.

3. The method according to claim 1, further comprising:

restricting the virtual machine to an initial allocation of processing units among the plurality of physical processing units according to the processor affinity of the virtual machine and setting the spin strategy to an initial spin strategy,

wherein the measuring comprises measuring the metric of the virtual machine allocated the initial allocation of processing units according to the processor affinity of the virtual machine and set to the initial spin strategy.

4. The method according to claim 1, wherein the adjusting comprises:

increasing the allocation of the plurality of processing units according to the processor affinity of the virtual machine among the plurality of physical processing units to which servicing of the virtual machine is restricted in response to measuring an increased utilization rate; and

decreasing the allocation of the plurality of processing units according to the processor affinity of the virtual machine among the plurality of physical processing units to which the servicing of the virtual machine is restricted in response to measuring a decreased utilization rate.

5. The method according to claim 1, wherein the spin strategy comprises disabling spinning while waiting for the synchronization lock to be freed or continuous spinning while waiting for the synchronization lock to be freed.

6. The method according to claim 1, wherein the measuring comprises intermittently measuring the metric over a period of time, and

wherein the adjusting comprises adjusting the plurality of tunable execution parameters, based on the intermittently measured metric.

7. A system for optimizing performance of an application executed by a virtual machine, the system comprising:

at least one processor configured to implement a controlling agent of the system running outside of the virtual machine that controls the virtual machine, the controlling agent comprising;

a measuring unit that measures a metric of an application executed by the virtual machine, the metric comprising a hardware metric of the system upon which the virtual machine is executed that is unavailable for measurement by the virtual machine, wherein the hardware metric comprises a utilization rate of physical processing units according to a processor affinity of the virtual machine among the plurality of physical processing units allocated to the virtual machine to which servicing of the application is restricted; and

an adjusting unit that adjusts a plurality of tunable execution parameters of the virtual machine for optimizing execution of the application based on the measured metric, wherein the tunable execution parameter comprises an allocation of the plurality of physical processing units according to the processor affinity of the virtual machine, to which servicing of the application is restricted, allocated to the virtual machine and a spin strategy that controls spinning in locks used for synchronization within the virtual machine.

8. The system according to claim 7, wherein measuring unit transmits a request to the virtual machine to adjust the plurality of tunable execution parameters, and

wherein the adjusting unit adjusts the plurality of tunable execution parameters in response to the transmitted request.

9. The system according to claim 7, wherein the virtual machine is restricted to an initial allocation of processing units among the plurality of physical processing units according to the processor affinity of the virtual machine and the spin strategy is set to an initial spin strategy, and

wherein the measuring unit measures the metric of the virtual machine allocated the initial allocation of processing units according to the processor affinity of the virtual machine and set to the initial spin strategy.

10. The system according to claim 7, wherein the adjusting unit increases the allocation of the plurality of processing units according to the processor affinity of the virtual machine among the plurality of physical processing units to which servicing of the virtual machine is restricted in response to measuring an increased utilization rate, and

decreases the allocation of the plurality of processing units according to the processor affinity of the virtual machine among the plurality of physical processing units to which the servicing of the virtual machine is restricted in response to measuring a decreased utilization rate.

11. The system according to claim 7, wherein the spin strategy comprises disabling spinning while waiting for the synchronization lock to be freed or continuous spinning while waiting for the synchronization lock to be freed.

12. The system according to claim 7, wherein the measuring unit intermittently measures the metric over a period of time, and

wherein the adjusting unit adjusts the plurality of tunable parameters based on the intermittently measured metric.