DYNAMICALLY PROVISIONING VIRTUAL MACHINES

• US 20110134761A1
• Filed: 12/03/2009
• Published: 06/09/2011
• Est. Priority Date: 12/03/2009
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Suffern, Edward S., Smith, Bruce A., Wooldridge, James L.
• Granted Patent: US 8,311,032 B2
• Time in Patent Office: Days
• US Class Current: 370/252
• CPC Class Codes: H04L 43/0852 Delays
• CPC Class Codes: H04L 43/106 using time related information in packets, e.g. by adding timestamps

Abstract

Method and computer program for managing the network response times experienced by virtual machines. Traffic is routed within an Ethernet network through a network switch according to IP addresses identified within Ethernet frames. A plurality of compute nodes are each coupled to the network switch via a separate Ethernet link. A response time is determined for each virtual machine running on the compute nodes, wherein the response time for a particular virtual machine is the difference between a time stamp in a TCP/IP Request message and a time stamp in a TCP/IP Response message for an IP address assigned to the particular virtual machine. The particular virtual machine may then be migrated to a target compute node in response to a particular one of the virtual machines on a particular one of the compute nodes having a response time that exceeds a response time setpoint.

Claims

1. A method, comprising:

routing traffic within an Ethernet network through a network switch according to IP addresses identified within Ethernet frames, wherein the Ethernet network includes a plurality of compute nodes, and each compute node is coupled to the network switch via a separate Ethernet link;

determining a response time for each of a plurality of virtual machines running on the plurality of compute nodes, wherein the response time for a particular one of the virtual machines is calculated as the difference between a time stamp in a TCP/IP Request message and a time stamp in a TCP/IP Response message for an IP address (media access control address) assigned to the particular one of the virtual machines; and

in response to a particular one of the virtual machines on a particular one of the compute nodes having a response time that exceeds a response time setpoint, migrating the particular virtual machine from the particular compute node to a target one of the compute nodes that is determined to be currently providing a lower response time than the specified compute node.

2. The method of claim 1, further comprising:

the network switch maintaining a management information database identifying the IP addresses associated with each of the Ethernet links.

3. The method of claim 2, further comprising:

reading the management information database of the network switch to determine the response time provided (amount of network bandwidth available) by each compute node; and

selecting the target compute node as one of the compute nodes currently providing a lower response time than the specified compute node.

4. The method of claim 1, wherein the response time is determined by a network switch.

5. The method of claim 1, further comprising:

identifying the TCP/IP Response message as having the next sequence number following a sequence number in the TCP/IP Request message for the IP address.

6. The method of claim 1, further comprising:

enforcing a time delay after migrating the specified virtual machine before any further virtual machine may be migrated.

7. The method of claim 1, further comprising:

prior to migrating the particular virtual machine, determining that the response time being provided on the second Ethernet link is lower than or equal to the response time required by the particular virtual machine.

8. The method of claim 1, further comprising:

prior to migrating the particular virtual machine, determining that each of the virtual machines on the target compute node are experiencing a lower response time than the response time of the particular virtual machine on the particular compute node.

9. A computer program product including computer usable program code embodied on a computer usable storage medium, the computer program product comprising:

computer usable program code for routing traffic within an Ethernet network through a network switch according to IP addresses identified within Ethernet frames, wherein the Ethernet network includes a plurality of compute nodes, and each compute node is coupled to the network switch via a separate Ethernet link;

computer usable program code for determining a response time for each of a plurality of virtual machines running on the plurality of compute nodes, wherein the response time for a particular one of the virtual machines is calculated as the difference between a time stamp in a TCP/IP Request message and a time stamp in a TCP/IP Response message for an IP address (media access control address) assigned to the particular one of the virtual machines; and

computer usable program code for migrating, in response to a particular one of the virtual machines on a particular one of the compute nodes having a response time that exceeds a response time setpoint, the particular virtual machine from the particular compute node to a target one of the compute nodes that is determined to be currently providing a lower response time than the specified compute node.

10. The computer program product of claim 9, further comprising:

computer usable program code for maintaining a management information database identifying the IP addresses associated with each of the Ethernet links.

11. The computer program product of claim 10, further comprising:

computer usable program code for reading the management information database of the network switch to determine the response time provided (amount of network bandwidth available) by each compute node; and

computer usable program code for selecting the target compute node as one of the compute nodes currently providing a lower response time than the specified compute node.

12. The computer program product of claim 9, further comprising:

computer usable program code for identifying the TCP/IP Response message as having the next sequence number following a sequence number in the TCP/IP Request message for the IP address.

13. The computer program product of claim 9, further comprising:

computer usable program code for enforcing a time delay after migrating the specified virtual machine before any further virtual machine may be migrated.

14. The computer program product of claim 9, further comprising:

computer usable program code for determining, prior to migrating the particular virtual machine, that the response time being provided on the second Ethernet link is lower than or equal to the response time required by the particular virtual machine.

15. The computer program product of claim 9, further comprising:

computer usable program code for determining, prior to migrating the particular virtual machine, that each of the virtual machines on the target compute node are experiencing a lower response time than the response time of the particular virtual machine on the particular compute node.

16. A method, comprising:

routing traffic within an Ethernet network through a network switch according to IP addresses identified within Ethernet frames, wherein the Ethernet network includes a plurality of compute nodes, and each compute node is coupled to the network switch via a separate Ethernet link;

determining a response time for each of a plurality of virtual machines running on the plurality of compute nodes, wherein the response time for a particular one of the virtual machines is calculated as the difference between a time stamp in a TCP/IP Request message and a time stamp in a TCP/IP Response message for an IP address (media access control address) assigned to the particular one of the virtual machines; and

provisioning an additional virtual machine to a particular one of the compute nodes, wherein the particular compute node is determined as providing the lowest response time to the virtual machines running on the particular compute node.

17. The method of claim 16, further comprising:

prior to migrating the particular virtual machine, determining that the response time being provided on the second Ethernet link is lower than or equal to the response time required by the particular virtual machine.

18. A computer program product including computer usable program code embodied on a computer usable storage medium, the computer program product comprising:

computer usable program code for routing traffic within an Ethernet network through a network switch according to IP addresses identified within Ethernet frames, wherein the Ethernet network includes a plurality of compute nodes, and each compute node is coupled to the network switch via a separate Ethernet link;

computer usable program code for determining a response time for each of a plurality of virtual machines running on the plurality of compute nodes, wherein the response time for a particular one of the virtual machines is calculated as the difference between a time stamp in a TCP/IP Request message and a time stamp in a TCP/IP Response message for an IP address (media access control address) assigned to the particular one of the virtual machines; and

computer usable program code for provisioning an additional virtual machine to a particular one of the compute nodes, wherein the particular compute node is determined as providing the lowest response time to the virtual machines running on the particular compute node.

19. The computer program product of claim 18, further comprising:

computer usable program code for determining, prior to migrating the particular virtual machine, that the response time being provided on the second Ethernet link is lower than or equal to the response time required by the particular virtual machine.