Many companies are asking two questions after moving some or all of their workloads into cloud vendors, such as Amazon, Google, or Azure. Do I need DR for the cloud? Where do you failover to if you’re running in the cloud?
Do I Need DR in The Cloud?
The short answer to this is yes. Remember, there is no such thing as “the cloud;” there is only someone else’s data center. Yes, the cloud is resilient. It’s more resilient than most people’s data centers. But it is not magic. Things can happen in the cloud, and when they do your workloads will crash or cease to exist.
Mistakes still happen in the cloud, and there are multiple examples of mistakes committed by cloud vendors. Microsoft forgot to renew its SSL certificates on Azure. Amazon allowed a race condition to take AWS down. Google accidentally allowed the deletion of millions of gmail accounts.
Your on-site personnel who manage your cloud infrastructure can also make mistakes. Finally, outside forces like viruses, malware, ransomware, and other attacks, can take out your entire environment in a single move. The cloud doesn’t solve all known problems.
Where Do You Failover to From The Cloud?
Where do you failover to? You could failover to another environment inside the same provider. That may be the least expensive option, but it’s the least secure one. You could also failover to another cloud provider. This is possible – depending on the apps you are running – but it can also get quite expensive to store all your data in two cloud vendors.
One final approach would be to failover to a data center that you own. If you’d like to do that, you’ll need two things. You need to be able to continually replicate your data back to your data center, and you’d need to manage the failover process.
Zerto Virtual Replication (ZVR) v5.5 Adds Failback from Microsoft Azure
Zerto is a well-known disaster recovery software that can run on VMware and Hyper-V, and is quickly spreading its roots as they evolve into an enterprise IT resilience cloud solution. It has been able to failover to another system running Hyper-V or VMware for some time, and announced a year ago at VMworld 2016 that ZVR v5.0 could failover – though not failback – from a virtualized data center to Azure. This recently changed in Zerto Virtual Replication v5.5 with the product now able to failover to/from a Hyper-V/VMware data center to/from Azure. This means a customer can now use Azure as their DR site for their data center, or their data center as their DR site for anything running in Azure.
When asked why Zerto is not yet doing this for Amazon – clearly the larger player in the cloud infrastructure market, Zerto explained that it has always been a hypervisor-centric system. Unlike competing products that run clients in VMs, it runs at the hypervisor level. They need to speak to a hypervisor to be able to failover and failback. Microsoft has open APIs that Zerto can use for this level of access, but Amazon and Google do not as of yet. Zerto is not the only company being held back by this limitation.
In addition to the Azure failover and failback feature discussed above, Zerto also added other features in 5.5. They include recovery speed enhancements for AWS with up to 6x faster RTO and on-the-fly conversions from Hyper-V and VMware, an analytics platform that is able to see all clouds managed and provides real-time and historical analysis capabilities, an automated upgrade process, and an extension to many of the Zerto APIs for increased automation and integration with other cloud environments.
Being able to failover to and from Azure using Hyper-V or VMware – without requiring agents inside VMs – is very compelling for enterprises looking to take an easy first step in their cloud journey and build on it. It also increases the value of Azure to those wishing to use it as their cloud computing or DR platform. Time will tell whether this becomes a differentiator for Azure, or if Amazon and Google simply follow suit with a set of APIs that would give vendors like Zerto hypervisor-level access.