Rebalancing virtual resources for virtual machines based on multiple resource capacities

• US 9,600,345 B1
• Filed: 03/11/2016
• Issued: 03/21/2017
• Est. Priority Date: 01/29/2016
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Cropper, Joseph W., Tenner, Jeffrey W.
• Primary Examiner(s): Kawsar, Abdullah Al
• Assistant Examiner(s): CHU JOY, JORGE A
• Application Number:
  US15/067,626
• Time in Patent Office: 375 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 2009/4557 Distribution of virtual machine instances; Migration and load balancing
• CPC Class Codes: G06F 2209/5011 Pool
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects
• CPC Class Codes: G06F 9/5011 the resources being hardware resources other than CPUs, Servers and Terminals
• CPC Class Codes: G06F 9/5077 Logical partitioning of resources; Management or configuration of virtualized resources specific details on emulation or internal functioning of virtual machines G06F9/455
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system

Abstract

A workload of a virtualized computing environment is rebalanced by resizing and/or moving one or more virtual machines allocated resources from a shared pool of resources in the virtualized computing environment based upon multiple resource capacities specified for each virtual machine and representing different portions of the resources from the shared pool of resources, e.g., minimum and desired resource capacities.

Claims

1. A computer-implemented method of rebalancing a workload of a virtualized computing environment, the virtualized computing environment of a type including a shared pool of resources disposed in a plurality of hosts and a plurality of virtual machines resident on the plurality of hosts and configured to share the shared pool of resources, wherein the virtualized computing environment stores, for each virtual machine among the plurality of virtual machines, a respective first resource capacity that represents a first portion of the shared pool of resources for the respective virtual machine and a respective second resource capacity that represents a second portion of the shared pool of resources for the respective virtual machine that is greater than that of the first resource capacity, the method comprising:

identifying a virtual machine among the plurality of virtual machines that is allocated a smaller portion of the shared pool of resources than that of the respective second resource capacity for the respective second portion;

determining using at least one processor if a home host for the identified virtual machine has sufficient unused capacity to support allocating to the identified virtual machine the respective second resource capacity for the identified virtual machine on the home host;

in response to determining that the home host for the identified virtual machine has sufficient unused capacity to support allocating to the identified virtual machine the respective second resource capacity for the identified virtual machine on the home host, resizing the identified virtual machine on the home host to allocate at least the respective second resource capacity to the identified virtual machine;

in response to determining that the home host for the identified virtual machine has insufficient unused capacity to support allocating to the identified virtual machine the respective second resource capacity for the identified virtual machine on the home host, attempting to identify a first candidate host among the plurality of hosts that has sufficient unused capacity to support allocating to the identified virtual machine the respective second resource capacity for the identified virtual machine on the first candidate host after accounting for the respective second resource capacity for each virtual machine among the plurality of virtual machines that is already resident on the first candidate host;

in response to identifying the first candidate host, moving the identified virtual machine to the candidate host and allocating to the identified virtual machine the respective second resource capacity for the identified virtual machine on the first candidate host;

in response to failing to identify the first candidate host, attempting to identify a second candidate host among the plurality of hosts that has sufficient unused capacity to support allocating to the identified virtual machine the respective first resource capacity for the identified virtual machine on the second candidate host after accounting for the respective second resource capacity for each virtual machine among the plurality of virtual machines that is already resident on the second candidate host; and

in response to identifying the second candidate host, moving the identified virtual machine to the second candidate host and allocating to the identified virtual machine the respective first resource capacity for the identified virtual machine on the second candidate host;

wherein attempting to identify the second candidate host includes identifying multiple hosts among the plurality of hosts having sufficient unused capacity to support allocating the respective first resource capacity to the identified virtual machine after accounting for the respective second resource capacities for virtual machines already resident thereon and selecting the second candidate host from among the multiple hosts based upon the second candidate host having a greatest remaining resource capacity from among the multiple hosts, wherein a respective remaining resource capacity for the second candidate host is equal to that of another host among the multiple hosts, wherein selecting the second candidate host further includes selecting the second candidate host over the other host to minimize resource contention, wherein the virtualized computing environment further stores, for each virtual machine among the plurality of virtual machines, a respective shared weight representing a ratio to which the respective virtual machine receives unused resource capacity relative to other virtual machines, and wherein selecting the second candidate host over the other host to minimize resource contention includes selecting the second candidate host based upon the second candidate host having a lower sum of shared weights for virtual machines resident on the second candidate host than that of the other host.

2. The method of claim 1, further comprising, in response to failing to identify the second candidate host, leaving the identified virtual machine on the home host.

3. The method of claim 1, wherein attempting to identify the second candidate host includes determining a remaining resource capacity for the second candidate host by dividing a sum of the respective second resource capacities for virtual machines resident on the second candidate host and that of the identified virtual machine by a total resource capacity for the second candidate host.

4. The method of claim 1, wherein attempting to identify the first candidate host includes:

identifying multiple hosts among the plurality of hosts having sufficient unused capacity to support allocating the respective second resource capacity to the identified virtual machine after accounting for the respective second resource capacities for virtual machines already resident thereon; and

selecting the first candidate host from among the multiple hosts based upon the first candidate host having a greatest remaining resource capacity from among the multiple hosts.

5. The method of claim 4, wherein a respective remaining resource capacity for the first candidate host is equal to that of another host among the multiple hosts, and wherein selecting the first candidate host further includes selecting the first candidate host over the other host to minimize resource contention.

6. The method of claim 5, wherein selecting the first candidate host over the other host to minimize resource contention includes selecting the first candidate host based upon the first candidate host having a lower sum of shared weights for virtual machines resident on the first candidate host than that of the other host.

7. The method of claim 1, wherein the first resource capacity for each respective virtual machine is a minimum resource capacity that represents an assured portion of the shared pool of resources for the respective virtual machine, and wherein the second resource capacity for each respective virtual machine is a desired resource capacity that represents a desired portion of the shared pool of resources for the respective virtual machine that is greater than the assured portion represented by the minimum resource capacity.

8. The method of claim 1, wherein the shared pool of resources includes a shared pool of processor resources, and wherein each first and second resource capacity represents a portion of the shared pool of processor resources.

9. The method of claim 1, wherein the shared pool of resources includes a shared pool of memory resources, and wherein each first and second resource capacity represents a portion of the shared pool of memory resources.

10. The method of claim 1, wherein identifying the virtual machine among the plurality of virtual machines includes identifying multiple virtual machines allocated respective smaller portions of the shared pool of resources than that of the respective second resource capacities for the respective second portions, the method further comprising, for each such identified virtual machine among the multiple identified virtual machines:

resizing such identified virtual machine on a respective home host therefor in response to determining that the respective home host has sufficient unused capacity to support allocating to such identified virtual machine the respective second resource capacity;

moving such identified virtual machine to a respective first candidate host and allocating to such identified virtual machine the respective second resource capacity for such identified virtual machine on the respective first candidate host in response to identifying the respective first candidate host as having sufficient unused capacity to support allocating to such identified virtual machine the respective second resource capacity after accounting for the respective second resource capacity for each virtual machine among the plurality of virtual machines that is already resident on the respective first candidate host;

moving such identified virtual machine to a respective second candidate host and allocating to such identified virtual machine the respective first resource capacity for such identified virtual machine on the respective second candidate host in response to identifying the respective second candidate host as having sufficient unused capacity to support allocating to such identified virtual machine the respective first resource capacity after accounting for the respective second resource capacity for each virtual machine among the plurality of virtual machines that is already resident on the respective second candidate host;

or

leaving such identified virtual machine on the home host in response to a failure to identify the respective first or second candidate host.