DRaaS solutions offer a cost-effective means to protect their business systems and data, and to return to operation in the event of a system failure or larger, more catastrophic outage. But, choosing the right one will depend on your organization’s needs and budget. Scanning DRaaS providers websites can be a little confusing since many make the same claims making it difficult to compare and contrast their capabilities.
| Criteria |  |  |  |  |  |
|---|---|---|---|---|---|
| Pricing Model |
$$$Appliance Cost (upfront) + monthly subscription (cloud) |
$$Appliance Cost (upfront) + monthly subscription (cloud) |
$$$Monthly subscription (cloud) |
$$$$Appliance Cost (upfront) + monthly subscription (cloud) |
$$$Annual software license |
| 2017 Gartner MQ for DRaaS |
Visionary |
Visionary |
Leader |
Visionary |
Did not qualify |
| Protected Environments | |||||
| —- Virtual |  | |  |  |  |
| —- Physical | | | | | |
| Orchestration | |  | | | |
| DR Testing | | | | | |
| Speed of Recovery | | | | | |
| Cloud Choice | | | | | |
| Support | | | | | |
| Failback | | | | | |
LEGEND
Pricing Model
We took a page out of Yelp’s book and used a five dollar ($$$$$) scale to compare the pricing of the different vendors. It’s important that you do your own price comparison as pricing can be highly variable, but make sure you do a full apples-to-apples comparison. We looked at the total cost of the solution that included the appliance cost (if applicable), cloud storage, software/licensing costs, support costs, and costs to perform DR tests or declare disasters.
Protected Environments
According to IDC, while over 75% of workloads will be virtualized in 2016, 79% of servers deployed are still dedicated to physical, non-virtualized roles. This is why it’s important to protect physical and virtual servers. For virtual protection, complete coverage (full circle) went to companies that protected VMware vSphere, Microsoft Hyper-V, Citrix XenServer, KVM, and Oracle KVM environments – the more that were supported, the more the circle was shaded. For physical servers, we evaluated whether the vendor protected Windows and Linux environments.
Orchestration
We evaluated DRaaS providers on the ease and completeness of their orchestration functionality. Some providers have very limited orchestration built in, while others make good use of runbooks and customized scripting to simplify the orderly failover of inter-connected systems.
DR Testing
Given the importance of testing, we looked at each vendor’s process for DR testing. Do you have to formally declare a disaster? Do you have to pay an additional cost every time you want to perform a test? Are you required to engage with the vendor to spin up VMs on the appliance or within the cloud? The easier the process, the more complete the circle.
Speed of Recovery
When disaster strikes, speed matters. So, we looked at how long it takes a solution to spin up a VM. This will depend on where the VMs are being spun up from – one could expect faster times when they are brought online from a local appliance vs. a public cloud. If your IT department has to formally declare a disaster this will naturally slow down the recovery time vs. a self-service model. We also looked at vendors who had formal SLAs around recovery time.
Cloud Choice
In order to get full credit on this criterion, the DRaaS providers had to support replication to private clouds, public clouds (e.g., Amazon AWS, Microsoft Azure, and/or Google Cloud Platform), and their own cloud options. The more options, the better.
Support
This is often a factor that is not truly appreciated until your organization experiences an actual disaster and needs real-time support and guidance. In evaluating the players in this space, we looked at online review sites, their support options, cost of support, and hours of operation.
Failback
When you evaluate a vendor’s failback procedures, you really have to understand how they’re failing back. The key criteria to look at is how fast and how much downtime is inflected with the vendor’s failback process. With some vendors, they leverage a vmotion type failback which results in zero downtime. Most vendors rely on a power down, restore data (back to production), and then a power up sequence. This process can take time and cause downtime.