The Cloud Lets Anyone Recover from Disaster

Data safety is a core business concern these days, even for small to midsize businesses (SMBs). But being prepared to ensure that your business operates even during potentially catastrophic disasters requires more than simply implementing business cloud backup. While such services are certainly part of a healthy disaster recovery (DR) plan, there's a lot more to be considered before you can call your organization safe, especially from an IT standpoint. A "disaster" refers to a complete halt of most or all of your systems simultaneously. If someone asks what your organization would do if all of its systems went dark one day, and you can't answer that question with confidence, then you're setting yourself up for serious problems down the road—and that's "when," not "if" because disaster hits everyone sooner or later. Fortunately, the cloud and the internet combine to make recovering from system disasters easier than ever.

Many SMBs dismiss planning a coherent DR policy because they believe such strategies are only for deep-pocket, enterprise-scale IT budgets. What can an SMB do? In the past, SMB-style preparation usually meant doing frequent backups and storing backup tapes offsite. Effective and cheap. Enterprise DR followed the same rules, but added provisioning for a hot-site, meaning another office that could be provisioned to house data center infrastructure and workers on short notice, or at the very least, have spare "dark" hardware and infrastructure capacity on hand to quickly replace failed systems. These measure do ensure some level of disaster-proofing, but they're both slow by today's standards and exceedingly expensive, which is why many SMBs opt to ignore DR planning entirely.

Fortunately, today's options are not only more effective, they're also cheap enough to be affordable for businesses of any size. That's due to two important IT enhancements: software-based infrastructure and the cloud. Virtualization allows businesses to provision servers as software instances on top of disparate hardware. So, for example, SMB X may have a single physical server in a data closet somewhere, but using Microsoft Hyper-V or VMware ESXi (among other virtualization platforms), they could be running two, three, or more software servers on that hardware with each software instance looking and behaving like any other server as far as the network and users are concerned. By implementing servers as software-based instances, IT managers get the ability to back those servers up by using the same methods they would for data.

But while backing up your virtual servers to a separate disk or tape gives you the ability to restore the server on another server eventually, backing them up to the cloud puts them in another data center—a shared data center that can not only restore your servers, but also initialize them and operate them just like you would and on just a few minutes' notice. That means, instead of paying for another building with an expensive data center you'll only access in emergenices, you can configure a cloud account to continually backup the most recent instances of your servers and simply switch them on if the primary servers at your local site fail. Instant Disaster Recovery-as-a-Service (DRaaS). And, since most of these cloud services operate on a pay-as-you-use basis, this lets even small businesses operate with sophisticated DR options at a fraction of what it would have cost in the past.

In this roundup, we tested five different DRaaS offerings, including Carbonite Server Backup, Microsoft Azure Site Recovery, Quorum onQ Hybrid Cloud Solution, Zerto Virtual Replication, and Zetta Backup and Recovery. All of those solutions are based on the DRaaS model, combined with some client-side software, as well as remote storage options.

What Is DRaaS?

During testing, we discovered that DRaaS vendors may loosely define exactly what DR consists of. Some vendors market their offerings as a simple, all-in-one solution, while others approach the DR market with a variety of options, such as SQL Server or MS Exchange-focused backups or complete server and even data center resurrection in the cloud. Weigh your options carefully as more options usually equals higher cost.

That said, it becomes very clear as to what must be part of a DR solution at a minimum:

• The ability to automatically backup critical systems and data.
• The ability to quickly recover from a disaster, with minimal user interaction.
• Flexible recovery options, such as restoring a single application or the whole infrastructure.
• Easily understood billing structure.
• Backup target options.

While the above outlined must-haves give a basic indication of what a DRaaS solution should do, SMBs should not just stop there to determine what service to choose. With that in mind, we created a checklist that should make the selection process a little easier:

DRaaS Checklist
Armed with the above checklist, SMB owners should be able to make short work of selecting a DRaaS provider that meets their needs.

Backup: (Which backup options/capabilities are offered)

• Will the DRaaS offering protect all of your line-of-business apps and platforms?

• Does the DraaS solution create local backups as well as cloud-based backups?

• Does the DraaS solution incorporate file size management to reduce storage needs?

• What apps, operating systems, and databases are directly supported?

• How frequently is the data backed-up or synchronized?

Recovery: (What happens if there is an on-site failure)

• How long does it take to recover an app? Server? Data?

• Will the DraaS support a self-service model?

• How complex is it to move from a backup to a live state?

• Does the DraaS offer a local appliance and address network change issues?

Failure-State: (How does the temporary recovery solution work)

• What level of performance can be expected when failed apps are hosted in the cloud?

• What architecture is used by the host to guarantee availability?

• What is the maximum number of VMs the DRaaS supports?

Failback: (What happens when on-site capabilities return to normal)

• Are there time limits on how long the provider will host the recovery environment?

• Are there additional charges, penalties, or other costs associated with long-term hosting?

• How does the provider manage restoration/failback?

• Is there a danger of data loss during failback?

• Does the failback process introduce downtime?