Disaster Recovery to the cloud - Scenario 1: Customer’s on-premise VMware environment to the cloud
It seems disaster recovery (DR) is mostly about planning, putting those plans in place, and then hoping they work in a disaster. At least that is what studies done by the Disaster Recovery Preparedness (DRP) Council show. They discovered only 13% of companies actually test their DR plans. And 25% of those that test, don’t retest after a failure.
Another study by the Aberdeen Group revealed the average cost of downtime for large businesses is £413,709 per hour. For small to medium businesses the average cost is a distressing £67,586 per hour.
The growth of the cloud and infrastructure as a service (IaaS) means businesses have many more options for disaster recovery and business continuity. Many cloud providers offer DR-as-a-service providing a secure offsite facility for critical business applications and data. But not all services or cloud providers are equal.
In this two part series on disaster recovery, we’ll cover how you can use the cloud for DR in a way that eliminates many of the difficulties of ‘traditional’ recovery options from an IT perspective.
This first post will look at using the process of taking an in-house IT solution and setting up a disaster recovery site in the cloud. The second post will examine implementing a disaster recovery plan from a cloud to a cloud whether they are private, public or hybrid.
Issues with traditional disaster recovery ‘solutions’
Organisations hosting their own IT infrastructure often try and build and manage their own DR IT infrastructure too. While this approach ensures some level of recovery and physical security, it introduces issues as well:
Double the capital investment – Recreating the exact hardware and software environment offsite means twice the price, and often more when everything is taken into account. This means many companies are able to protect only the most critical of systems and applications which still results in a break in business continuity even if the recovery is a success.
Focus on data and systems – DR plans are often written and implemented by IT staff or management. Their priority is backing up the data, rebuilding the servers and networks at the DR site, and then turning it all back on. Basically the nuts and bolts of how to get you back up and running. This isn’t an issue really, except that it’s the business applications and the ability to use them that matters for business continuity. It’s a mindset difference with tremendous impact on how quickly your services are restored.
Untestable – Traditional DR plans are not always testable, or they can’t be fully tested without interruption to the production systems. This is part of the reason the DRP Council study found companies simply didn’t test their DR processes.
Human resource dependent – There is usually a large manual component to the recovery process stressing both human assets and budgets, especially in cases where only part of an organisation’s application or data is protected.
Using the cloud to make it easier and quicker
We’ve developed disaster recovery options for a number of various sized companies over the last few years. And consistently we’ve helped them create and test DR plans to address all of the issues we listed above. As a certified VMware Enterprise Service Provider, historically we’ve often relied on VMware vCenter Site Recovery Manager (SRM) which simplifies, automates, and provides the ability to test the implementation of a DR plan between private VMware clusters.
But more recently we have been very impressed with a technology company called ‘Zerto’. Zerto is a virtual appliance based DR solution solely built for VMware environments and offers a totally new concept on disaster recovery that builds on VMware’s capabilities. So much so that Zerto won the best in show at VMworld in 2011.
Zerto’s technology performs replication at the hypervisor layer versus the storage layer as in many traditional recovery options. This approach offers many benefits to companies managing disaster recovery from an in-house solution to a cloud DR site:
Hardware agnostic – Because Zerto works at the hypervisor layer, you no longer need to replicate the exact hardware at the DR site. You can use EMC storage arrays in house, and seamlessly recover on a cloud infrastructure using NetApp. Whatever brand and configuration of hardware our clients have, they can recover on our cloud infrastructure without worrying about compatibility.
Focus on applications – Ultimately you want your business applications up and running so you can keep working without any downtime should a disaster happen. Zerto is also designed to ensure you can do that by using ‘virtual protection groups’. No matter how many virtual machines your applications need to function, you group them together and recover them together all at the same point in time. Since the VMs are in sync, even the most delicate apps recover safely (Microsoft SharePoint for example). So the focus has moved from replicating servers, networks and storage to recovering your applications to keep you up and running.
Built for testing – One the most critical aspects of a DR plan is testing, and testing, then testing again. You need to know it will work when you need it. Zerto is designed so you can test as frequently as you want without affecting your production systems by creating a ‘bubble’ network. And instead of the standard recovery point objective (RPO), Zerto uses point in time, journal based recovery. They refer to it as Tivo for your environment. This aids in testing, but also is a main reason you can ensure business continuity from whichever recovery point you to choose.
Entirely automated – Failover, failback, and recovery are all automated with Zerto, similar to VMware’s SRM. This removes much of the human element in disaster recovery plans meaning your staff is able to focus on your business and less on servers and storage. Added benefits over SRM include support for multiple sites and failback once the production site is up and running again.
Using the cloud and this type of recovery technology for your disaster recovery plans can help you limit the impact of any disaster affecting your in-house environment. In our next post we’ll discuss how organisations already using the cloud for their production environment can leverage this same technology for their disaster recovery and business continuity solutions.