Data Back-up vs. Replication vs. Disaster Recovery

Featured article by Ken Donoghue, Peak 10 Contributing Writer 

We often hear the three concepts expressed in the title above used interchangeably. For example, it’s not uncommon for someone to say, “I have a disaster recovery (DR) plan. We back up our files once a day.” Or, “We replicate our data to tape and store it offsite.” Each concept has its place protecting data and retrieving data, but how and why each applies is different.

Let’s start by saying that neither data backup nor data replication constitutes a DR plan. Second, the purpose of data backup has to do with recovery but not disaster recovery. Third, replication has everything to do with disaster recovery, but there are many more components to a DR plan. Got it?

Let’s take a look deeper.

Data backup up is just that ─ making a backup copy of your data files by taking periodic snapshots. It could be hours between snapshots or a day. The backup is often to physical tape (the least expensive storage medium but also the least flexible in terms of retrieving information) or to a virtual tape library (VTL) and kept offsite. Retrieving the media and then pulling the information off is a time-consuming task and not in line with any sort of meaningful DR requirements.

The use case for data backup is to have a copy or copies of everything from the least important information to the most critical. This can be for the purpose of compliance and/or to pinpoint data recovery of, let’s say, an employee’s emails for e-discovery, or a single transaction from five years ago or a deleted file from yesterday.

Many organizations traditionally have done their own backup. The cloud provides an attractive alternative, especially with volumes of data constantly on the rise. Security and compliance must be maintained as always. Many Peak 10 colocation customers rely on the company to back up their data according to predefined time windows and to provide ongoing retention.

Data replication is another matter. Replication means copying and moving data to another location, typically in real time or near-real time. Doing this for all of an organization’s data would be prohibitively expensive, which is why replication is reserved for only the most essential applications, processes and data that are required for keeping a business up and running in the event of disaster.

A key component of a data replication strategy is to define recovery time objectives (RTO) and recovery point objectives (RPO); that is, how long can you be without critical workloads and how much data loss can you reasonably absorb? The less flexible the RTO/RPO requirements are, the more complex and costly the solution will be.

Peak 10 provides customers with solutions according to their needs, tailoring the most cost-effective and efficient means possible to achieve their objectives. This includes asynchronous, near-synchronous and synchronous replication. Using continuous data protection (CDP) versus scheduled replication happening a few times a day will result in less data loss in the event of a failure.

Data replication and RTO/RPO are essential to an effective DR plan, but are not all that is required. Replication means only that you have a copy of company-critical information available to you, that you have the raw material for declaring an emergency and consequently enacting a comprehensive DR plan. In turn, a DR plan will be a component of a larger business continuity (BC) plan, which is becoming a regulatory compliance necessity in some industries such as healthcare.

Disaster Recovery as a Service (DRaaS) is becoming one of Peak 10’s high-demand services. A huge benefit of DRaaS is that it affords the ability to regularly test your DR plan, which is one of the most ignored aspects of self-provided DR planning. Once you have defined your RTO and RTO, tested your plan, and clearly defined the parameters necessary to declare an emergency, the rest happens automatically with DRaaS. This includes one of the most overlooked aspect of DR … failing back to the production site once the emergency has passed.

Understanding that the nature and purpose of backup, replication and DR are complementary but not interchangeable IT functions will help you design and implement the right combination of each to the advantage of your entire organization. It’s the needs of your entire organization that define what is required of each separately, as well as collectively.

Ken Donoghue, Peak 10 Contributing Writer 

Ken’s public relations and business writing career includes more than 25 years in high technology. The lavish launch of the first true portable computer, the Data General One, in 1984 and the first fault-tolerant server for Windows computing in 2001 are among his favorite career highlights. Ken makes his home in mid-coast Maine, working as an independent business and technology writer. He has been a contributing writer to Peak 10 Inc., a national IT infrastructure and cloud services provider based in Charlotte, N.C., since October 2013.