Google has taken a lead on cloud rivals by granting its virtual machines immortality via a new live migration feature.

This means that when Google shuts down cloud data centers for scheduled maintenance in the future, client VMs will stay up while being migrated to another online bit barn.

The announcement was made in a post to a Google Compute Engine user group on Wednesday, alongside a bevy of cloud announcements from arch-rival Amazon Web Services at the web bazaar's re:Invent conference in Las Vegas.

"Now, we will be able to automatically move the affected VMs out of the way for you, and the zone itself will remain up and usable through the scheduled maintenance events," Google wrote.

"We will, of course, continue to perform scheduled maintenance: patching our systems with the latest software, performing routine tests and preventative maintenance, and generally ensuring that our infrastructure is as fast and efficient as we know how to make it."

Cloud admins will be able to specify whether they want their virtual machines to be live migrated at a small performance cost, or terminated and restarted on new infrastructure.

"The exact guest performance impact and duration depends on many factors, but we expect most applications and workloads won't be adversely affected," the company wrote.

Live VM migration has been seen as something of a holy grail for many cloud providers, and it is a feature so far unavailable in other clouds, though VMware has said it plans to do the tech as well.

Though many vendors such as VMware (vMotion), Oracle (Oracle VM), and Xen have live migration capabilities, cloud providers have so far not implemented the technology.

Google did not explain exactly how it has been able to achieve the feat, but we are certain that it will come from the advanced capabilities granted to the Chocolate Factory by its "living" Borg and Omega resource managers and schedulers.

The migration options should be visible to admins "in a few weeks", Google said. There are a couple of caveats, notably that to be capable of live migration a machine must either use persistent disks exclusively for block storage, or a scratch disk for boot and persistent disk for additional storage.

So far, Google has added the features to its US-central1-a and US-central1-b data center zones, but is sure to expand over time. ®

Related stories

• AMD: Why we had to evacuate 276TB from Oracle DB to Hadoop (24 March 2014)

  https://www.theregister.com/2014/03/24/amd_hadoop_migration/

• Stanford academics unleash Quasar cluster juggler on mega bit barns (28 February 2014)

  https://www.theregister.com/2014/02/28/quasar_launch/

• Inside Microsoft's Autopilot: Nadella's secret cloud weapon (7 February 2014)

  https://www.theregister.com/2014/02/07/microsoft_autopilot_feature/

• HPCaaS experiment becomes commercial service (4 February 2014)

  https://www.theregister.com/2014/02/04/hpcaas_experiment_becomes_commercial_service/

• So you want to build the next Google. Who ya gonna call? Er, Big Blue? (9 January 2014)

  https://www.theregister.com/2014/01/09/ibm_resource_scheduler_upgrade/

• Google's new cloud CRUSHES Amazon in RAM battle (3 December 2013)

  https://www.theregister.com/2013/12/03/google_cloud_launch_analysis/

• Google makes Compute Engine official (3 December 2013)

  https://www.theregister.com/2013/12/03/google_compute_engine_launches/

• Is that you, HAL? No, it's NEIL: Google, US Navy pour money into 'associative' AI brain (26 November 2013)

  https://www.theregister.com/2013/11/26/neil_associative_ai/

• Half-blind Glassholes rejoice! 'Prescription' Google goggles a-comin' (23 November 2013)

  https://www.theregister.com/2013/11/23/blind_glassholes_rejoice_google_working_on_prescription_headsets/

• If this doesn't terrify you... Google's computers OUTWIT their humans (15 November 2013)

  https://www.theregister.com/2013/11/15/google_thinking_machines/

• Amazon forms THE VIRTY DOZEN to assassinate rival flash cloud servers (14 November 2013)

  https://www.theregister.com/2013/11/14/amazon_cloud_server_refresh/

• Bigger, fiddly to manage: The second coming of Windows Server 2012 (22 October 2013)

  https://www.theregister.com/2013/10/22/win_server_2012_r2_review/

• Sysadmins hail Windows Server 2012 R2's killer ... clipboard? (5 September 2013)

  https://www.theregister.com/2013/09/05/sysadmins_hail_server_2012_r2s_killer_app/

• VMware goes after biz critical apps with vSphere 5.5 (26 August 2013)

  https://www.theregister.com/2013/08/26/vmware_esxi_vsphere_nsx_vsan_sddc/

• HDS gives HUS a flashy boost, polishes NAS platters too (10 July 2013)

  https://www.theregister.com/2013/07/10/hds_hustles_hus/

• The good and the bad in Hyper-V's PowerShell (30 May 2013)

  https://www.theregister.com/2013/05/30/hyper_v_powershell_hands_on/