Disaster Recovery and the Cloud
Migrating to the cloud can be a daunting task. First you have to go through an exhaustive process deciding what to migrate, then you have to do plenty of scripting and re-architecting to get all that to cloud, all while keeping everything secure. If you’re curious about the cloud but hesitant to take on all the challenges of a large migration, you should consider a gradual first step: disaster recovery.
Cloud Disaster Recovery vs. Traditional DR
Disaster recovery is a good fit for the cloud for several reasons. Traditional disaster recovery is expensive and cumbersome, and infrequently tested. That means should an actual outage or disaster happen, a recovery could take days or even weeks, since the DR site needs to be a close match to the protected site, and for most environments that requires extensive and complex manual processes. In between (infrequent) tests or dry runs these steps can be inadvertently missed. A duplicate site is equipped, powered and operated for the prospect of use less than 5% of the time, if at all. That is incredibly wasteful. Cloud DR is a superior operating model to backup and traditional warm or cold standby models: it offers better protection and often much lower costs.
A Good Way to get a Feel for the Cloud
DR is an easy and rewarding way to start getting a feel for the cloud. For one, you’ll still have your applications and data in your primary datacenter, so you won’t have to worry about committing any important workloads completely to the cloud. Second, once you’ve successfully set up DR in the cloud, you can eliminate your secondary datacenter or co-location facility. It’s a quick way to save costs, not just experiment.
Another financial factor to consider is the fact that there are no dependencies in the cloud. You can do DR for month or year without having to get locked into any contract. This is possible because, in the cloud, you’re not using native services.
Compute is expensive in the cloud. There are plenty of stories of IT departments who end up with hefty bills after leaving compute on 24/7 in AWS or Azure. But you only use DR solutions on occasion – basically when testing or in an actual emergency. Thus, compute costs typically remain low for cloud DR, making it a less expensive way to try the cloud.
There are a number of organizations who don’t have any DR strategy in place to begin with. This is a risky proposition. However, if that’s the case, you can get started fresh with DR in the cloud without having to worry about migrating anything. You’ll quickly see that this is cheaper than opening a secondary datacenter or doing co-location, both of which incur greater expenses from replicating servers.
Moreover, there are a lot of people who mistake backup for disaster recovery. These folks simply back up their data rather than developing a real DR strategy. So when something happens and their datacenter goes down, they’ll have the files, but rebuilding from just that will take days or even weeks. The issue is in reconstructing the applications, which are much more complicated than simple data. The truth is that if you have a solid DR strategy, you don’t need to invest much in backup. (If you’re new to DR, make sure to take an application-centric approach rather than focusing primarily on the data).
Testing, Testing, 1,2,3
DR is only reliable if you test it regularly. As mentioned earlier, DR testing is complex to do on premises, which is why many enterprises that deploy traditional DR don’t do it nearly as often as they should. This isn’t the case in the cloud, though. Testing can be done with the touch of button, allowing you to failover immediately and begin running apps in real time. With traditional disaster recovery, that process would require days of preparation and resources to coordinate the actual testing across multiple functions and IT resources.
The cloud opens the door for much greater automation, and that’s what makes testing so much easier there. It’s difficult to manually failover to a secondary datacenter, create a report, then get things running back in your primary datacenter. But all that can be done within hours using cloud automation.
Dipping your Toe in
Too many businesses fail to take advantage of useful technology due to the difficulty and scope of implementing it. When it comes to the cloud, don’t feel like you have to dive in head first. You can start by tipping your toe in with disaster recovery. Once you’ve done DR there, you’ll get a greater sense of what other workloads will be a good fit for the cloud. Or if you’re not happy with the DR experience, you won’t be locked into a contract and can move back to a physical datacenter with relative ease. There’s little risk, cost and effort involved.
There are many enterprises that plan to only do DR in the cloud in order to save on secondary datacenters or co-location facilities, then become so enamored that they move other workloads as well. The key benefits – cost and convenience – still apply for most all applications. The cloud isn’t an either/or proposition, and more and more enterprises are embracing a hybrid approach. DR is an easy way to get started.
By James Kessinger