Now’s the time to review your DR plan.
Much has changed with disaster recovery in the 10-plus years since we got a firsthand lesson from Hurricane Katrina. Server virtualization now makes recovery of many systems much easier and faster than it was, and the rise of the cloud puts more options at our disposal than we could have previously imagined.
Hurricane season is officially here, making it a good time to review the current state of disaster recovery solutions.
What is Disaster Recovery?
First, a bit of terminology; This article is about disaster recovery, which for our purposes deals with the continuation or recovery of technology systems important to a business following a disaster — as opposed to business continuity, which covers all aspects of keeping a business running.
Any disaster recovery discussion involves the terms Recovery Time Objective (RTO), which is how long it takes to get things up and running, and Recovery Point Objective (RPO), which refers to the time period of lost data. Disaster recovery typically measures RTO and RPO in terms of hours. High availability is yet another topic that deals with RTO and RPO measured in seconds or minutes.
Finally, keep in mind that disaster recovery is not the same as backup. While disaster recovery focuses on restoring whole systems, backup focuses on restoring data. If I need to find a deleted email from last month, I need backup. If I need to get my whole email system up and running after a fire in my building, I need disaster recovery. With that said, there are some products that are designed and able to perform both functions well.
Leveraging the Cloud
My favorite approach to accomplishing disaster recovery is to offload the problem by using applications that are completely managed and hosted by cloud Software as a Service (SAAS) providers. When the right cloud-based application exists for your needs at the right price (and with its own documented disaster recovery plan), I say use it. This approach works very well for email, documents, certain line of business apps, and soon, if not today, phone systems.
Of course, there is not yet a SAAS application for every need, and the next best option is to move self-managed legacy applications from servers in your office to resilient data centers or cloud Infrastructure as a Service (IAAS) providers. As before, it’s still important to ensure that your provider and design meet your disaster recovery expectations.
The problem here is that many applications just don’t perform very well when hosted remotely. Sometimes using Remote Desktop or Citrix is a fix, but when it’s not, the next approach is to keep applications local and replicate them to the cloud or a branch office. There are many ways to accomplish this, ranging from low-cost solutions that handle one application or server at a time to enterprise solutions that can synch hundreds of servers with automated failover. Assuming your Internet connection is sized appropriately, replication can work well, but it definitely requires good documentation and regular testing to ensure that everything goes as smoothly as possible.
Other approaches might include using a completely different application for a particular need during outages, resorting to a single user on a laptop rather than a fully networked application, or forgoing disaster recovery and relying on backup specifically for those applications whose acceptable RTO is high.
The good news is that it has never been easier to keep your business going no matter what happens, and if you haven’t reviewed your plan in the last year or two, now is the time.