According to much of what I hear and read, we’re on the precipice of everything moving to the cloud. Why should an IT or business leader care? Lower cost – check, got that, better mobility – got that too. Anything else? There is another element of cloud though that tends to fly under the radar and that is the benefit to a company’s disaster recovery plan.
As a former IT person, one who was heavily involved in disaster planning, I know how challenging it is to put a good disaster recovery plan together. In fact, I often joke about that fact that everyone is an expert in backing up data but restoring it is where the challenge comes from. However, even the best companies with dedicated IT teams struggle to get to a recovery point objective of 24 hours. Think about that. One day of outage before things are brought back up. It doesn’t seem too bad, but who knows what can happen in a day? If you’re in financial services and the market moves quickly, that one day could cost more than the rest of the year. A small business that misses out on delivering a big order could damage its reputation. A school system being unavailable during exams could have significant ramifications as well.
The beauty of the cloud, though, is that all of the infrastructure is located off premise so there’s no local infrastructure to protect, secure, procure and build. I know some of the cloud providers are trying to build resiliency into their offerings. At the April 5th Navicloud event in Los Angeles, NaviSite president Brook Borcherding stated on a panel that he was attempting to build a recovery point objective of zero. So any company using the NaviSite services won’t ever have any downtime, assuming there’s network connectivity.
Companies that want to use multiple cloud providers or build their own can strive for this too but need to make some IT changes. The first step would be to actually embrace the cloud for mission-critical applications. I know a few companies have but the majority have not. The second step would be to align IT to be more cross discipline (check out my previous blog on this). Having IT live in individual stove pipes doesn’t exactly scream seamless IT.
The biggest step, and this will become necessary sooner than later, is to build a dynamic network fabric. Fabrics have, of course, been one of the hottest trends in networking over the past year or so and are the way networks will be built as virtualization and cloud become more prevalent. The network fabric will enable the seamless movement of IT resources across and between data centers and cloud providers. Traditionally, the network is the last thing people think about but it needs to be considered with any cloud deployment.
To have a successful fabric deployment it’s important that companies embrace automation as much as they can. This would include the obvious, like coordinating the movement of storage and network policies with VM migrations, but also automating things like DNS updates and IP address management. Additionally, find tools with robust GUI front ends. Manual changes done through command line interfaces is slow and often error prone. Automation can bring companies closer to the promise of cloud and reduce human latency.
So, if you’re looking for a reason to go cloud, considering making DR part of the justification. If you’re looking for a reason to deploy a network fabric, use cloud and DR as part of the justification. But no matter what the reason or how you start the deployment, remember to automate as much as possible.