vCloud Air DRaaS – The Good, Bad & Ugly

I was recently given the task of looking into the vCloud Air DRaaS offering from VMware to understand how this could meet customers requirements to provide a disaster recovery platform for their production virtual machines.

This blog contains my understanding of the components, fit together, the good, the bad and the downright ugly parts of the vCloud Air DRaaS offering.

vCloud Air Key Benefits

The key benefits of the vCloud Air Disaster Recovery service are:

  • Restore point objective settings per virtual machine of 60 minutes or less
  • Subscription based ‘Recovery as a Service’ reducing capital outlay
  • Simple and secure asynchronous replication for virtual machines
  • Self-service disaster recovery testing of virtual machine
  • Guaranteed resource availability
  • On premise monitoring and management with the vSphere Web Client
  • Support for initial data seeding using vCloud Connector Offline Data Transfer
  • Supports the same guest operating systems and applications as vSphere
  • Application consistency for virtual machines running a Microsoft Guest Operating System
  • Integrates with your existing vSphere environment

vCloud Air Overview

vCloud Air Disaster Recovery is based on the following technologies:

  • Local VMware vSphere
  • Local and remote VMware vSphere Replication
  • Local and remote vCloud Connector
  • Remote VMware vCloud Air Disaster Recovery

A logical overview is shown below.

Logical Overview

vCloud Air Disaster Recovery is comprised of a number of components which enable the management, replication and access to virtual machines.

  • vSphere Replication – Provides asynchronous replication to copy virtual machines to an alternate location. Making the Makes the virtual machine copy available for recovery
  • vCloud Connector – Provides support for initial data seeding and failback of virtual machines to source site using Offline Data Transfer
  • vCloud Air – Infrastructure as a Service cloud owned and operated by VMware

vSphere Replication Appliance

A special version of vSphere Replication is used to copy virtual machines between location, and make that copy available for restoration using the vSphere Web Client.

  • vSphere Replication protects the virtual machine on an ongoing basis, replicating only changes that are made
  • Integrates with Microsoft Volume Shadow Copy Services (VSS) ensure that applications such as SQL Server databases are quiescent and consistent when replica data is being generated
  • Application consistency is achieved using Microsoft Windows Operation System with VMware Tools installed
  • vCloud Tunnelling Agent in the vSphere Replication appliance creates a tunnel to secure the transfer of replication data to vCloud Air Disaster Recovery service

vCloud Connector

vCloud Connector provides a single user interface for viewing private and public clouds and transferring content between them, as shown in the diagram below.

vCloud Connector

  • Connectivity between on premises and vCloud Air Disaster Recovery is achieved using SSL
  • Offline Data Transfer enables transfer of large volumes of data to the vCloud Air infrastructure
    • Data is securely encrypted onto a storage device supplied by VMware
  • vCloud Connector facilities failback of virtual machines

Test Recovery

vCloud Air Disaster Recovery service provides the ability to perform two test recoveries per twelve month period.

  • Testing a recovery does not affect production virtual machine workloads
  • Virtual machines in the test environment can be powered on for up to 7 days
  • While a test recovery runs, vSphere Replication continues to replicate data from the virtual machines at your source site
  • Test invoked at the source site have the ability to synchronise recent changes
  • Test recovery is per virtual machine

The diagram on the next page shows a logical overview of a test recovery.

Test Recovery

Failover

When you recover a virtual machine from your source site to vCloud Air, the production state of the virtual machine represents a point in time before the outage. Data accumulated after the last replication to vCloud Air and before the recovery is not available.

  • Recovering a virtual machine to vCloud Air stops replication from the source site
  • Virtual machines can be ran for up to 30 days in a failover scenario without additional cost.
    • If the source site is unavailable for a prolonged period, virtual machines can be transitioned to vCloud Air Private or Dedicated Cloud
  • Service Level Agreement of 4 hours or less of dedicated compute in a failover event, powered on and remotely accessible
  • Failover is per virtual machine

Failback

Failback of virtual machines that have been recovered in vCloud Air is performed using vCloud Connector.

  • Virtual machines are copied back to the source site using vCloud Connector using SSL
    • Default SSL Certificates are required to be replaced with an internal certificate authority
  • Failback should be within a planned outage windows as virtual machines are required to be powered off before being copied back to the source site

Consideration should be given to the amount of data to be transferred back in relation to the bandwidth available during failback

The Ugly

Having read this blog post you will have seen that a number of items would have jumped out at you.  The key considerations are shown below:

  • Failover to vCloud Air Disaster Recovery is per virtual machine.
    • Orchestration is possible, however this is currently at API level
  • DHCP is the default port group network setting. On failover each virtual machine receives an APIPA address.  Manual selection of the correct port group along with verification of network settings for each VM is required.
    • Note this is likely to change as vCloud Air Disaster Recovery Service is updated
  • vCloud Air Disaster Recovery does not include backup of virtual machines. Operational consideration should be given to the impact this may have to the recoverability of in guest data or failure of a VM
  • Failback from vCloud Air Disaster Recovery to on premises requires the virtual machines to be shutdown and copied back via vCloud Connector

Final Word

vCloud Air Disaster Recovery is a relatively new service introduced by VMware to UK datacentres in April 2014.  Even though vCloud Air Disaster Recovery is built on the proven technologies vCloud Director and vSphere Replication, a number of considerations exist,  Make sure that you make sure the service meets your requirements!

4 thoughts on “vCloud Air DRaaS – The Good, Bad & Ugly

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s