Dynamic allocation of physical computing resources amongst virtual machines

• US 9,003,037 B2
• Filed: 07/25/2012
• Issued: 04/07/2015
• Est. Priority Date: 07/25/2012
• Status: Active Grant

• Current Assignee: Vmware LLC (Broadcom, Inc.)
• Original Assignee: VMware, Inc. (Broadcom, Inc.)
• Inventors: Beveridge, Daniel James
• Primary Examiner(s): NGUYEN, THU HA T
• Application Number:
  US13/558,211
• Publication Number:
  US 20140032761A1
• Time in Patent Office: 986 Days
• Field of Search: 709/223, 709/224, 709/226
• US Class Current: 709/226
• CPC Class Codes: G06F 9/45533 Hypervisors; Virtual machine monitors
• CPC Class Codes: G06F 9/5011 the resources being hardware resources other than CPUs, Servers and Terminals
• CPC Class Codes: G06F 9/542 Event management; Broadcasting; Multicasting; Notifications

Abstract

Techniques for resource allocation are described. Some embodiments provide a resource allocation manager configured to dynamically allocate physical computing resources amongst multiple virtual machines hosted on a physical computing system. The resource allocation manager receives information about the importance of activities that can occur on the multiple virtual machines. Activities may include or be based on virtual machine states, virtual desktop states, business activities or processes performed by users, or the like. Then, the resource allocation manager determines that an activity is occurring on one of the virtual machines. In response, the resource allocation manager reallocates physical computing resources of the physical computing system that are assigned or available to the one virtual machine, based on the occurring activity and its corresponding importance. The resource allocation manager continually and dynamically reallocates physical computing resources assigned to corresponding virtual machines, in order to match resource allocation to activity importance.

Claims

1. A method for resource allocation in a virtualized computing environment comprising at least one physical computing system hosting multiple virtual machines, the method comprising:

dynamically allocating physical computing resources of the at least one physical computing system amongst the multiple virtual machines, each virtual machine configured to execute a virtual desktop that is remotely accessible by a user, the dynamic allocation including;

receiving indications of importance corresponding to activities triggered by an active user session with the virtual desktop that can occur on the multiple virtual machines;

monitoring a state of the virtual desktop of each virtual machine to detect the activities triggered by the active user session;

determining, based on the monitoring of the state of the virtual desktop, that one of the activities has been triggered by the active user session and is occurring on one of the multiple virtual machines; and

in response to the occurring activity, automatically modifying, based on the occurring activity and its corresponding importance, an allocation of physical computing resources of the at least one physical computing system that are assigned to the one virtual machine.

2. The method of claim 1, wherein the activities that can occur on the multiple virtual machines include multiple virtual desktop states associated with a virtual machine, including at least one of:

whether or not a user is logged into the virtual machine, whether a user is logged in but currently disconnected from the virtual machine, whether a user is logged and currently connected to the virtual machine, whether a user has been idle for longer than a predetermined time interval, an indication of an application being executed on the virtual machine, and/or an indication of a level or type of network activity performed by the virtual machine.

3. The method of claim 2, wherein determining that one of the activities is occurring on one of the multiple virtual machines includes:

receiving an indication that the one virtual machine is currently associated with one of the multiple virtual desktop states.

4. The method of claim 1, wherein the activities that can occur on the multiple virtual machines include multiple business tasks performed by users with respect to one or more applications executing on a virtual machine, and wherein determining that one of the activities is occurring on one of the multiple virtual machines includes:

receiving an indication that one of the multiple business tasks is being performed by a user on the one virtual machine.

5. The method of claim 4, wherein determining that one of the activities is occurring on one of the multiple virtual machines includes:

determining that the one business task is being performed by the user on the one virtual machine, based on indications of actions and/or events occurring within an application that is executing on the one virtual machine and that is being operated by the user, the indications of actions and/or events received from a monitoring module executing on the one virtual machine.

6. The method of claim 1, further comprising:

monitoring and recording information about desktop states associated with virtual machines;

monitoring and recording information about application actions and events occurring within applications executing on virtual machines; and

determining activities occurring on the multiple virtual machines based on the information about desktop states and/or the information about application actions and events.

7. The method of claim 1, wherein receiving the indications of importance corresponding to activities that can occur on the multiple virtual machines includes:

receiving from a user an assignment of a business value of each of the activities.

8. The method of claim 7, wherein receiving the assignment of the business value of each of the activities includes:

assigning each of the activities to a corresponding resource level, priority, and/or category.

9. The method of claim 1, wherein determining that one of the activities is occurring on one of the multiple virtual machines includes:

receiving performance information related to the one activity, the performance information indicating at least one of response time of the activity, execution time of the activity, and/or throughput related to the activity.

10. The method of claim 9, wherein automatically modifying the allocation of physical computing resources includes:

when the received performance information indicates that performance of the one activity has degraded outside of a specified range, increasing the allocation of physical computing resources.

11. The method of claim 1, wherein automatically modifying the allocation of physical computing resources includes:

increasing a share of the physical computing resources of the at least one physical computing system, wherein the amount of the physical computing resources available to the one virtual machine is based on a ratio of the increased share over a sum of all shares assigned to each of the multiple virtual machines hosted on the physical computing system.

12. The method of claim 1, wherein automatically modifying the allocation of physical computing resources includes:

increasing or decreasing one or more of CPU resources, memory resources, I/O resources, and/or network bandwidth resources available to the one virtual machine.

13. The method of claim 1, wherein automatically modifying the allocation of physical computing resources includes:

increasing or decreasing an amount of CPU resources, memory resources, I/O resources, and/or network bandwidth resources that are reserved for use by the one virtual machine, wherein reserved resources are guaranteed to be available to the one virtual machine.

14. The method of claim 1, wherein automatically modifying, based on the occurring activity and its corresponding importance, an allocation of physical computing resources, further comprises:

modifying a number of shares of a central processing unit (CPU) that was previously allocated to a virtual machine based on the occurring activity and its corresponding importance; and

independently modifying a number of shares of physical memory that was previously allocated to the virtual machine based on the occurring activity and its corresponding importance, wherein the modification of the number of shares of the physical memory is performed independently of the modification of the number of shares of the CPU.

15. A non-transitory computer-readable storage medium including instructions that are configured, when executed by a computing system, to perform a method for resource allocation in a virtualized computing environment comprising at least one physical computing system hosting multiple virtual machines, the method comprising:

dynamically allocating physical computing resources of the at least one physical computing system amongst the multiple virtual machines, each virtual machine configured to execute a virtual desktop that is remotely accessible by a user, the dynamic allocation including;

receiving indications of importance corresponding to activities triggered by an active user session with the virtual desktop that can occur on the multiple virtual machines;

monitoring a state of the virtual desktop of each virtual machine to detect the activities triggered by the active user session;

determining, based on the monitoring of the state of the virtual desktop, that one of the activities has been triggered by the active user session and is occurring on one of the multiple virtual machines; and

in response to the occurring activity, automatically modifying, based on the occurring activity and its corresponding importance, an allocation of physical computing resources of the at least one physical computing system that are assigned to the one virtual machine.

16. The computer-readable storage medium of claim 15, further comprising:

determining historical average performance of the multiple activities;

comparing current performance of the occurring activity to is historical average performance to determine whether to automatically modify the allocation of physical computing resources.

17. The computer-readable storage medium of claim 16, further comprising:

determining current performance of the occurring activity by averaging a predetermined number of recently occurring execution times or response times of the occurring activity.

18. The computer-readable storage medium of claim 15, wherein receiving the indications of importance corresponding to activities that can occur on the multiple virtual machines includes:

for each of the activities that can occur on the multiple virtual machines, receiving separate indications of a CPU resource level, a memory resource level, an I/O resource level, a network resource level, wherein the indicated resource levels indicate resource needs of the corresponding activity.

19. The computer-readable storage medium of claim 18, wherein automatically modifying the allocation of physical computing resources includes:

separately modifying one or more of CPU resources, memory resources, I/O resources, and network resources available to the one virtual machine based on the received separate indications and on the importance of the occurring activity.

20. The computer-readable storage medium of claim 15, wherein automatically modifying the allocation of physical computing resources includes:

determining resource utilization of the one virtual machine due to the occurring activity; and

increasing or decreasing at least one of CPU resources, memory resources, I/O resources, and network resources available to the one virtual machine based on the determined resource utilization and on the importance of the occurring activity.

21. The computer-readable medium of claim 20, wherein determining resource utilization of the one virtual machine includes receiving execution profile information collected with respect to an application associated with the occurring activity.

22. The computer-readable medium of claim 20, wherein increasing at least one of CPU resources, memory resources, I/O resources, and network resources available to the one virtual machine includes:

increasing one of the CPU resources, memory resources, I/O resources, and network resources while maintaining the other of the CPU resources, memory resources, I/O resources, and network resources at a previous level.

23. The computer-readable medium of claim 15, wherein determining that one of the activities is occurring on one of the multiple virtual machines includes determining that a user session is idle, the user session associated with a virtual desktop hosted by the one virtual machine, wherein the one virtual machine is not idle and performing one or more non-interactive tasks, and further comprising:

reducing, based on the idle user session, the allocation of physical computing resources of the at least one physical computing system that are assigned to the one virtual machine.

24. The computer-readable medium of claim 15, wherein automatically modifying the allocation of physical computing resources is further based on whether a current time is associated with a period of low or high resource utilization of the multiple virtual machines.

25. A computing system for resource allocation in a virtualized computing environment comprising at least one physical computing system hosting multiple virtual machines, the system comprising:

a memory;

a processor;

a module that is stored on the memory and that is configured, when executed by the processor, to dynamically allocate physical computing resources of the at least one physical computing system amongst the multiple virtual machines, each virtual machine configured to execute a virtual desktop that is remotely accessible by a user, the dynamic allocation performed by;

receiving indications of importance corresponding to activities triggered by an active user session with the virtual desktop that can occur on the multiple virtual machines;

monitoring a state of the virtual desktop of each virtual machine to detect the activities triggered by the active user session;

determining, based on the monitoring of the state of the virtual desktop, that one of the activities has been triggered by the active user session and is occurring on one of the multiple virtual machines; and

in response to the occurring activity, automatically modifying, based on the occurring activity and its corresponding importance, an allocation of physical computing resources of the at least one physical computing system that are assigned to the one virtual machine.

26. The computing system of claim 25, wherein the computing system is a management server that manages the allocation, distribution, and provisioning of multiple virtual desktops for multiple client users.

27. The computing system of claim 26, wherein the management server includes a module configured to provide state information related to each of the multiple virtual desktops.

28. The computing system of claim 25, wherein the multiple virtual machines are system virtual machines managed by a hypervisor and each having a distinct associated guest operating system.

29. The computing system of claim 25, wherein each of the multiple virtual machines executes an associated monitoring module that is configured to transmit indications of actions and events occurring within applications that are being executed by the associated virtual machine and operated by a user.

30. The computing system of claim 29, wherein the module is further configured to:

record the indications of actions and events transmitted by the monitoring modules;

determine that the one activity is occurring based on sequences of actions and events represented by the recorded indications of actions and events.