Generate VMs based on actual load in a Resource Pool

During the last VMworld in Barcelona I was in a session (INF-VSP2033 – Auto Scaling and Cloud Bursting in the Hybrid IaaS Cloud) from Christophe Decanini and Chris Knowles who showed a solution with two sites located in Europe and the USA with vSphere and vCO environments on both sites. The sites were connected with F5 Loadbalancer and the load could be migrated from one Country to the other Country……sincerely the session is not online in the VMworld online session folder so you cannot see what the showed…..

To give you a hint what is possible with the vCO, I created my “own” solution based on the load in a given resource pool, see following video:

Some background notes what I did with my Workflow:

  • I monitor a resource pool for the load in the RP. It is also possible to do so in a folder
  • I can define a minimum amount of VMs with must be run in the RP. A deletion below this minimum point is not possible from the Workflow
  • I can define a maximum amount of running virtual machines in the RP. The workflow cannot provision VMs above the maximum amount.
  • My virtual machines witch were provisioned are all Linux based (I use a CentOS). The data for the machines (for example a Website) must be located outside the VMs. I connect them via NFS or iSCSI. The machines are an ideal example for vApps
  • The load is captured over a time interval of 5 minutes. Short peaks will not result in any action (create new VMs or delete existing VMs)

So have fun with the Video and get a look what is possible with the vCO. If you have questions feel free to ask ([email protected]).

Load based generation of virtual machines from Christian Strijbos on Vimeo.