Network performance monitor for virtual machines

Patent metadata

Field	Value
Publication number	US-9063769-B2
Application number	US-201013575643-A
Country	US
Kind code	B2
Filing date	Feb 4, 2010
Priority date	Feb 4, 2010
Publication date	Jun 23, 2015
Grant date	Jun 23, 2015

Abstract

Embodiments of the present invention relate to a method and system to monitor network performance in a computer including at least one physical server having a hardware machine attached via a hypervisor to a set of virtual machines. According to an embodiment of the present invention, the system measures network performance parameters for selected packet flows in transition between the hardware machine and the virtual machines, analyzes the measured performance parameters, and identifies a source of network degradation.

First claim

The invention claimed is:

1. A method to monitor performance of a computer comprising at least one physical server where each physical server comprises a hardware machine attached via a hypervisor to a plurality of virtual machines, wherein the method in each physical server comprises:
measuring server performance parameters for a selected packet flow in transition between the hardware machine and the virtual machines, where the server performance parameters are measured by utilizing time stamps obtained in the hardware machine and the virtual machines for at least one of incoming packets and outgoing packets;
estimating the measured server performance parameters by observing parameters time stamped at the hardware machine and corresponding parameters time stamped at the virtual machines;
identifying at least one of the virtual machines as a source of server degradation by establishing degradation of the observed parameters time stamped at the at least one of the virtual machines; and
wherein each physical server has a physical hardware clock which is used to obtain the time stamps in the hardware machine and the virtual machines.

2. The method to monitor performance according to claim 1 , wherein the method further comprises:
migrating the at least one of the virtual machines to improve resource allocation.

3. The method to monitor performance according to claim 2 , wherein the improvement is performed by migration across physical servers of virtual machines contributing to or experiencing resource depletion in the computer.

4. The method to monitor performance according to claim 2 , wherein the computer further comprising a Service Level Agreements database, and whereby the improvement is performed in accordance to Service Level Agreements or application requirements obtained from the Service Level Agreements database.

5. The method to monitor performance according to claim 1 , wherein the source of server degradation is determined by observing at least one of jitter and delay.

6. The method to monitor performance according to claim 2 , wherein a decision to migrate at least one virtual machine comprises:
estimating a baseline for delay by calculating an average of the delay for the selected packet flow in transition between the hardware machine and the virtual machines;
estimating for the selected packet flow an average delay value for a specified time window; and,
triggering a virtual machine migration if average delay value > baseline.

7. The method to monitor performance according to claim 2 , wherein a decision to migrate at least one virtual machine comprises:
estimating for the selected packet flow an average jitter value for a specified time window; and,
triggering a virtual machine migration if at least one of an average jitter value > a service level agreement value and if the average jitter value for the specified time window is increasing.

8. A computer comprising:
at least one physical server, where each physical server comprises a hardware machine attached via a hypervisor to a plurality of virtual machines, wherein each physical server further comprises a performance monitor that interfaces with the hardware machine and the plurality of virtual machines, and wherein the performance monitor is configured to:
measure server performance parameters for a selected packet flow in transition between the hardware machine and the virtual machines, where the server performance parameters are measured by utilizing time stamps obtained in the hardware machine and the virtual machines for at least one of incoming packets and outgoing packets;
estimate the measured server performance parameters by observing parameters time stamped at the hardware machine and corresponding parameters time stamped at the virtual machines; and,
identify at least one of the virtual machines as a source of server degradation by establishing degradation of the observed parameters time stamped at the at least one of the virtual machines; and,

wherein each physical server further comprises a physical hardware clock which is used to obtain the time stamps in the hardware machine and the virtual machines.

9. The computer according to claim 8 , wherein the performance monitor is further configured to:
migrate the at least one of the virtual machines to improve resource allocation.

10. The computer according to claim 9 , wherein the improvement is performed by migration across physical servers of virtual machines within the computer which are contributing to or experiencing resource depletion in the computer.

11. The computer according to claim 9 , wherein the computer further comprising a Service Level Agreements database, and wherein the improvement is performed in accordance to Service Level Agreements or application requirements obtained from the Service Level Agreements database.

12. The computer according to claim 8 , wherein the performance monitor is further configured to determine the source of server degradation by observing at least one of jitter and delay.

13. The computer according to claim 9 , wherein the performance monitor is configured to make a decision to migrate at least one virtual machine by:
estimating a baseline for delay by calculating an average of the delay for the selected packet flow in transition between the hardware machine and the virtual machines;
estimating for the selected packet flow an average delay value for a specified time window; and,
triggering a virtual machine migration if average delay value > baseline.

14. The computer according to claim 9 , wherein the performance monitor is configured to make a decision to migrate at least one virtual machine by:
estimating for the selected packet flow an average jitter value for a specified time window; and,
triggering a virtual machine migration if at least one of an average jitter value > a service level agreement value and if the average jitter value for the specified time window is increasing.

15. The computer according to claim 8 , further comprising:
a Service Level Agreements database interfaced with the performance monitor; and
a network server interfaced with the hardware machine.

16. A performance monitor in a physical server, wherein the physical server comprises a hardware machine attached via a hypervisor to a plurality of virtual machines, wherein the physical server further comprises a physical hardware clock, wherein the hardware machine comprises a hardware network interface card, wherein each virtual machine comprises a virtual machine network interface card, and wherein the performance monitor comprises:
a processor; and,
a memory that stores processor-executable instructions, wherein the processor interfaces with the memory to execute the processor-executable instructions, whereby the performance monitor is operable to:
select a packet flow to monitor which is in transition between the hardware machine and the virtual machines;
receive time stamps from the hardware network interface card and the virtual network interface card for at least one of incoming packets and outgoing packets associated with the selected packet flow, wherein the time stamps were obtained in the hardware machine and the virtual machines by using the physical hardware clock;
estimate server performance parameters by utilizing the obtained time stamps; and,
identify at least one of the virtual machines as a source of server degradation by utilizing at least the estimated server performance parameters.

17. The performance monitor according to claim 16 , wherein the performance monitor is opera

Assignees

• Damola Ayodele
• Johnsson Andreas
• Ericsson Telefon Ab L M

Inventors

• Damola Ayodele
• Johnsson Andreas

Classifications

• G06F11/349

  for interfaces, buses · CPC title

• G06F11/3419

  by assessing time · CPC title

• G06F11/3433

  for load management (allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029) · CPC title

• G06F2201/815

  Virtual · CPC title

• G06F11/3466

  Performance evaluation by tracing or monitoring · CPC title