SHAPING VIRTUAL MACHINE COMMUNICATION TRAFFIC

• US 20110231568A1
• Filed: 03/16/2010
• Published: 09/22/2011
• Est. Priority Date: 03/16/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: BANSAL, DEEPAK, WARKHEDE, PRIYANK RAMESH
• Granted Patent: US 8,234,400 B2
• Time in Patent Office: Days
• US Class Current: 709/233
• CPC Class Codes: G06F 9/54 Interprogram communication
• CPC Class Codes: H04L 47/12 Avoiding congestion; Recovering from congestion
• CPC Class Codes: H04L 47/22 Traffic shaping
• CPC Class Codes: H04L 67/147 Signalling methods or messages providing extensions to protocols defined by standardisation
• CPC Class Codes: H04W 4/60 Subscription-based services using application servers or record carriers, e.g. SIM application toolkits

Abstract

Cloud computing platforms having computer-readable media that perform methods to shape virtual machine communication traffic. The cloud computing platform includes virtual machines and a controller. The controller limits the traffic associated with the virtual machines to enable the virtual machines to achieve desired communication rates, especially when a network servicing the virtual machines is congested. The controller may drop communication messages associated with the virtual machines based on a drop probability evaluated for the virtual machines.

Claims

1. A computer-implemented method to manage virtual machine communication traffic in a cloud computing environment having multiple virtual machines, the method comprising:

maintaining an average communication rate for the virtual machines in the cloud computing platform;

computing a drop probability for the virtual machines as a function of the average communication rate and a desired communication rate for the virtual machines; and

randomly dropping a communication message for the virtual machines when a drop probability for the virtual machines is greater than a nonce generated by the cloud computing platform.

2. The computer-implemented method of claim 1, further comprising:

limiting transmission communication messages or reception communication messages on the virtual machines to achieve the desired communication rates for each virtual machine in the cloud computing platform.

3. The computer-implemented method of claim 2, wherein the limiting of transmission communication messages or reception communication messages is triggered when the cloud computing platform detects congestion on a network servicing the virtual machines.

4. The computer-implemented method of claim 3, wherein the limiting of the communication rates for the virtual machines is performed without queues.

5. The computer-implemented method of claim 1, wherein the average communication rate is an average sending rate.

6. The computer-implemented method of claim 5, wherein the average sending rate (ASR) is evaluated for the virtual machines by the cloud computing platform, wherein ASR=α

*previous_ASR+(1−

α

)*current_SR, “

α

”

is set by the cloud computing platform, the “

previous_ASR”

is the ASR previously evaluated for the virtual machines, and “

current_SR”

is the current sending rate detected by the cloud computing platform for the virtual machines.

7. The computer-implemented method of claim 6, wherein the desired communication rate is the desired sending rate.

8. The computer-implemented method of claim 7, wherein the desired sending rate (DSR) is set in communication policies for the virtual machine.

9. The computer-implemented method of claim 8, wherein the drop probability (DP) is evaluated for the virtual machines by the cloud computing platform, wherein

10. The computer-implemented method of claim 1, wherein the nonce (η

) is randomly generated for the virtual machines by the cloud computing platform, wherein

17. The computer-readable media of claim 1, wherein the nonce (η

) is randomly generated for the virtual machines by the cloud computing platform, wherein

11. One or more computer-readable media storing instructions to perform a method to drop virtual machine communication traffic, the method comprising:

calculating an exponentially weighted average for the sending rate of a virtual machine;

identifying the desired sending rate for the virtual machine;

evaluating a drop probability for a communication message sent from the virtual machine; and

dropping a communication message when a nonce generated for the virtual machines is greater than the drop probability.

12. The computer-readable media of claim 11, wherein the communication message is a packet.

13. The computer-readable media of claim 12, wherein the virtual machine communication traffic is governed by a Transmission Control Protocol (TCP).

14. The computer-readable media of claim 11, wherein the exponentially weighted average sending rate (ASR) is evaluated for the virtual machine by a cloud computing platform, wherein ASR=α

*previous_ASR+(1−

α

)*current_SR, “

α

”

is set by the cloud computing platform, the “

previous_ASR”

is the ASR previously evaluated for the virtual machines, and “

current_SR”

is the current sending rate detected by the cloud computing platform for the virtual machines.

15. The computer-readable media of claim 14, wherein the desired sending rate (DSR) is set in a service level agreement for the virtual machine.

16. The computer-readable media of claim 15, wherein the drop probability (DP) is evaluated for the virtual machines by the cloud computing platform, wherein

18. A cloud computing platform configured to shape communication traffic for virtual machines, the cloud computing platform comprising:

virtual machines that communicate over a network; and

a controller configured to drop communication messages associated with the virtual machines when a nonce generated for the virtual machines is greater than the drop probability.

19. The cloud computing platform of claim 18, wherein the controller evaluates the average sending rate (ASR) for the virtual machine, wherein ASR=α

*previous_ASR+(1−

α

)*current_SR, “

α

”

is set by the controller, the “

previous_ASR”

is the ASR previously evaluated for the virtual machines, and “

current_SR”

is the current sending rate detected by the controller for the virtual machines, and the controller evaluates the drop probability (DP) for the virtual machines, wherein

20. The cloud computing platform of claim 19, wherein the controller randomly generates nonce (η

), wherein