VMware on AWS – How this will look from a Veeam perspective

Announced at the end of August at VMworld in Las Vegas was VMware on AWS, nothing new here as this was spoken about a year ago, but things are now moving towards being available.

From the small bit of exposure, I have seen it really does look like they have done a great job in making things seamless between an on premises vSphere deployment and the new vSphere offering on AWS.

Let’s firstly look at what VMC looks like.

What do you get if you want to spin this up?

  • Minimum 4 ESXi Hosts
    • 36 Cores,
    • 512GB RAM
    • 8x2TB NVMe SSD for vSAN
  • HA, vMotion, DRS, Elastic DRS Support
  • vCenter deployed
  • vSAN preconfigured. (1 single Datastore)
  • NSX access limited (FW, VPN, NAT).
  • Default connection to Amazon AWS VPC to consume AWS directly.

Veeam should work straight out the box, then right?

From a Veeam deployment point of view and initially yes Veeam will just work the way you expect or as it does on your on-premises vSphere environment today. The two key noticeable differences are:

HotAdd only mode available, with this release there is a specific new VDDK that has been released that only allows for the Veeam virtual appliance transport to work here.

 

Also note:

There is no ability to leverage the NFS based features within Veeam, such things as Instant VM Recovery (IVMR), Virtual Labs, SureBackup, SureReplica.

Moving forward this could be changed from VMware.

What will work then?

All other features such as Backup and Replication including application aware processing in guest. From a recovery perspective, the ability to restore the entire VM, Windows Fie Level Recovery as well as all Veeam Explorers.

Another key thing to mention again is the replication engine will work, this means you can leverage Veeam to migrate an on premises workloads to VMC.

  • Backup incl. In Guest Application processing
  • Restore (Original VM, to new name)
  • Windows File Level Recovery
  • Veeam Explorers for Exchange, SQL, SharePoint, Oracle, AD
  • VM Replication

Backup Scenarios

Simple backup

A short-term retention backup, that 7,14,30-day retention, using a native AWS EC2 instance running windows with some disk attached for a repository. This could act as your primary backup repository residing as close to the primary data for that failure scenario and fast recovery. Traffic from VMC to AWS is free.

Backup Copy to Glacier (tape)

If tape is still a requirement then Veeam has the supported capability of using the Amazon VTL to push data today into Glacier.

Backup up to S3/Glacier via Archive Tier (v10)

Coming in v10 this will extend that functionality further by making it possible natively to send backup data to either S3/Glacier via a policy defined on a scale out backup policy.

The above covers the short-term backup retention, but to adhere to the 3-2-1 rule we should look at the other options we have available to us.

Backup Copy to another AWS region

Even leaving an AWS region will cost you an egress charge! Be warned. AWS traffic to other geolocation will be charged by AWS. Veeam WAN Accelerators reduce the traffic. This would simulate an offsite copy of your data.

Backup Copy to On Premises

I believe this could be the most popular use case, flipping your on premises equipment to now your offsite or your DR location. Sending that offsite backup copy down to your existing sweating assets on premises.

Backup Copy to Cloud Connect

A nice elegant solution if looking to remove all on premises hardware would be to consider Veeam Cloud Connect for that offsite backup or DR location.

Disaster Recovery Scenarios

There is also the possibility to replicate from VMware on AWS region to another for disaster recovery scenarios.

And again, if you are looking to sweat those on premises resources you could also replicate your virtual machines to and from your VMware on AWS instances.

The other and most common scenario and use case would be migrating workloads from on premises into the VMware on AWS platform using the Veeam replication. I also want to emphasise here again that you should consider your data movement when using AWS anything that egresses out of the AWS network or even to a different AWS region is going to cost you money, be sure to calculate accordingly and choose the right protection model for your business requirements.

I have to add here that the lovely visio Veeam diagrams in this post were not created by me, as soon as I know who the author was I will give them full credit for them here

I also had the pleasure of speaking with vmblog.com at VMworld US 2017, where I walked through a demo of Veeam protecting workloads in VMware on AWS.

Leave a Reply

Your email address will not be published. Required fields are marked *