The cloud may be one of the most hyped technologies in years, but that doesn’t mean it’s not useful. This is especially true when it comes to building a disaster recovery (DR) plan. Companies can now create cloud based recovery sites that are available if the normal data center is unavailable for whatever reason. There are many ways to design a recovery plan, and we’ll go over one of the most popular below.
Evaluate Your Needs
Before you get into the nitty gritty of designing anything, you need to first know what you actually require This determines the configuration and infrastructure that will work best for you. Some companies use cloud storage as a part of disk-to-disk-to-cloud backup. This means the primary backups are onsite, but they are also available on a cloud server where they’re protected if something disables the data center. Other companies replicate virtual machines to the cloud so they can be hosted if something goes wrong.
Choose a Provider
After you know what you need, you can start looking for cloud providers that offer what you need. Some providers may allow replication of a virtual machine but won’t host it. Some do little more than offer storage. Make sure to dig deep and find a provider that offers exactly what you need. Don’t pick the first one you see, either. Come up with a handful and compare costs and levels of service.
Cost Estimation
So you have a provider, now you need to determine what the costs will be using cloud DR. Every provider has a pricing model, but most of them have some combination of fees that might include: monthly subscription, amount of Internet bandwidth used, amount of storage space used, number of virtual machines.
It’s also smart to check whether a provider requires payment for a VM that isn’t on. Some will, but some will not. Of course, the latter is best, especially if you don’t expect to use the virtual machine often.
Develop a Bandwidth Strategy
It’s essential to decide on a strategy for managing the bandwidth. There are a few reasons for this. First, most cloud providers out there charge for consumption of bandwidth. Secondly, your ISP may have monthly caps or charge for excess bandwidth. Third, you need enough bandwidth available for timely backups. Finally, you want to be sure the backups don’t consume tons of bandwidth to the point where another Internet usage point suffers.
Determine Logistics
Logistical planning might not take long if you plan to use the cloud only for storage. However, if you plan to perform failovers on the cloud, you’ll have many more considerations. The logistics will vary based on your cloud service and infrastructure, but there are some common things to consider.
You’ll want to determine how you copy data from the data center to the cloud. This might be software based if you’re using a public cloud. You’ll need something supported by your onsite resources and the cloud, as well.
You will also want to consider Active Directory synchronization, as clustering solutions may require nodes to be in a common Active Directory domain. As such, if you plan to extend a cluster to the cloud for DR, you will also need to extend an onsite AD domain to the cloud.
Machine Replication
Finally, if you plan to use replication of virtual machines, you will need to determine what you hope to gain from it. You might want image based recovery or the ability to mount a cloud copy of the VM to extract data. In either case, you’ll want to determine this before you set things up, so you can ensure the option is available.