Efficient provisioning and deployment of virtual machines

Patent metadata

Field	Value
Publication number	US-9929931-B2
Application number	US-201314050365-A
Country	US
Kind code	B2
Filing date	Oct 10, 2013
Priority date	Mar 16, 2011
Publication date	Mar 27, 2018
Grant date	Mar 27, 2018

Abstract

Machines, systems and methods for managing quality of service in a virtualized computing environment, the method comprising: provisioning one or more active virtual machines (VMs) over one or more hosts in a virtualized computing network, wherein one or more resources are allocated to the active VMs before the active VMs service one or more requests; monitoring information associated with quality of service defined for servicing of the requests; and designating at least an active VM as a shadow VMs, in response to results of the monitoring, wherein at least one resource remains allocated to the shadow VM, while the shadow VM enters a dormant state and no longer services any requests.

First claim

What is claimed is:

1. A method comprising:
creating a first active virtual machine (VM) on one or more hosts in a virtualized computing network, wherein one or more resources are allocated to the first active VM before the first active VM services one or more requests, the one or more resources including at least one of virtual memory space, CPU resources and physical storage space, and wherein the first active VM is located in one or more active pools of VMs;
monitoring information associated with quality of service defined for servicing of the requests, wherein the quality of service is based on a number of simultaneous user sessions;
determining a number of active VMs needed to provide a service;
designating the first active VM as a first shadow VM in response to the determining the number of active VMs needed to provide the service, wherein at least one resource of the one or more resources remains allocated to the first shadow VM, and wherein the first shadow VM is maintained in a dormant state;
transferring the first shadow VM to an inactive pool of VMs in response to designating the first active VM as a first shadow VM, wherein the first shadow VM is configured to be activated and deployed upon request, wherein one or more shadow VMs are in one or more inactive pools of VMs located on the one or more hosts and one or more active VMs are in one or more active pools of VMs located on the one or more hosts;
scoring the one or more shadow VMs according to the resources of each of the one or more shadow VMs;
calculating a placement plan that places one or more shadow VMs having the highest scores and the one or more active VMs on the one or more hosts;
determining the calculated placement plan satisfies one or more placement objectives, the one or more placement objectives including at least one of maximizing a number of shadow VMs, minimizing the number of hosts, and minimizing costs associated with migrating shadow VMs;
reallocating resources to the one or more shadow VMs having the highest scores based on the calculated placement plan; and
placing the one or more shadow VMs having the highest scores and the one or more active VMs on the one or more hosts according to the calculated placement plan.

2. The method of claim 1 , wherein the inactive pool of VMs is located on a host and an active pool of VMs of the one or more active pools of VMs is located on the host, and wherein the inactive pool of VMs and the active pool of VMs are logically separated on the host, the method further comprising:
determining that there is a need for a specific application to add additional VMs to service a request;
determining that at least one shadow VM associated with the specific application is present within the inactive pool of VMs;
reactivating the at least one shadow VM in response to determining that at least one shadow VM associated with the specific application is present in the inactive pool of VMs;
transferring the at least one shadow VM from the inactive pool of VMs to the active pool of VMs and registering the at least one shadow VM with the active pool of VMs as a second active VM in response to reactivating the at least one shadow VM; and servicing the request using the second active VM.

3. The method of claim 1 , further comprising:
receiving a length of time the first shadow VM has been in the dormant state;
determining that the length of time the first shadow VM has been in the dormant state satisfies a threshold time period; and
deallocating the at least one resource from the first shadow VM and removing the first shadow VM from the inactive pool of VMs in response to determining that the length of time the first shadow VM has been in the dormant state satisfies the threshold time period.

4. The method of claim 1 , further comprising: determining that there is a need for a specific application to add additional VMs to service a request;
determining that no shadow VMs associated with the specific application are present within the inactive pool of VMs;
determining that there are insufficient resources available to create one or more new VMs;
fully deactivating one or more shadow VMs having the lowest scores;
provisioning the one or more new VMs;
registering the one or more new VMs; and
reallocating resources from the one or more deactivated shadow VMs to the one or more registered new VMs.

5. The method of claim 1 , the method further comprising:
storing, by a service provider, in a network storage, at least one VM master image;
specifying, by the service provider, the at least one VM master image in a service definition manifest; and
transmitting, by the service provider, the service definition manifest to a site management.

6. The method of claim 5 , further comprising:
utilizing the at least one VM master image to provision a VM on the one or more hosts.

7. The method of claim 5 , further comprising:
creating, by the site management, one or more VMs at the one or more hosts in response to transmitting the service definition manifest to a site management; and
hosting one or more of the created VMs on a host of the one or more hosts.

8. The method of claim 7 , further comprising:
downloading, by the host, a particular VM master image to a local storage device from a network storage device based on the service definition manifest;
caching the particular VM master image in the local storage device; and
creating a new VM by cloning the particular VM master image.

9. A system comprising:
a hardware processor; and
a non-transitory computer-readable storage medium having program instructions embodied therewith, the program instructions executable by the hardware processor to cause the system to:
create a first active virtual machine (VM) on one or more hosts in a virtualized computing network, wherein one or more resources are allocated to the first active VM before the first active VM services one or more requests, the one or more resources including at least one of virtual memory space, CPU resources and physical storage space, and wherein the VMs are located in one or more active pool of VMs;
monitor information associated with quality of service defined for servicing of the requests, wherein the quality of service is based on a number of simultaneous user sessions;
determine a number of active VMs needed to provide a service;
designate the first active VM as a first shadow VM in response to the determining the number of active VMs needed to provide the service, wherein at least one resource of the one or more resources remains allocated to the first shadow VM, and wherein the first shadow VM is maintained in a dormant state;
transfer the first shadow VM to an inactive pool of VMs in response to designating the first active VM as a first shadow VM, wherein the first shadow VM is configured to be activated and deployed upon request, wherein one or more shadow VMs are in one or more inactive pools of VMs located on the one or more hosts and one or more active VMs are in one or more active pools of VMs located on the one or more hosts;
score the one or more shadow VMs according to the resources of each of the one or more shadow VMs;
calculate a placement plan that places one or more shadow VMs having the highest scores and the one or more active VMs on the one or more hosts;
determine the calculated placement plan satisfies one or more placement objectives, the one or more placement objectives including at least one of maximizing a number of shadow VMs, minimizing the number of hosts, and minimizing costs associated with migrating shadow VMs;
reallocate resources to the one or more shadow VMs having the highest scores based on the calculated placement plan; and
place the one or

Assignees

• IBM

Inventors

• Breitgand David
• Epstein Amir
• Nagin Kenneth

Classifications

• G06F2009/4557

  Distribution of virtual machine instances; Migration and load balancing · CPC title

• G06F9/5005

  to service a request · CPC title

• G06F2009/45562

  Creating, deleting, cloning virtual machine instances · CPC title

• G06F2009/45575

  Starting, stopping, suspending or resuming virtual machine instances · CPC title

• H04L43/16 Primary

  Threshold monitoring · CPC title