Photobank kiev - Fotolia


Are you getting data movement for disaster recovery wrong?

George Crump discusses best practices for data movement, including the most effective methods for small companies, SMBs and larger enterprises.

The first step in any disaster recovery plan is to get a secondary copy of data off site. The way copy is transported and stored is a critical factor in determining if a disaster recovery plan will be successful. The range of options for data movement is almost limitless and will vary depending on the size of the organization and the demands being placed on IT.

Finding an SMB data movement strategy

Small businesses can easily be exposed to data movement failures, even though they likely have far less data to move than a larger business or enterprise. Many small organizations simply copy data to a series of external hard drives for transport off site. It's fairly common that a small business can fit all of its data onto a single drive and these drives are inexpensive. But they are not designed to be portable and there are no guarantees on shelf life. In other words, the chance of data loss is high during transport and the chances of a recovery failure are even higher when the drive is finally plugged back in.

Careful planning is required to ensure both the success of a disaster recovery plan and the recovery of decades-old data.

Small businesses often can't justify the investment in a tape infrastructure. While the media is affordable, the cost of a tape drive can be upwards of $2,000. As a result, I recommend using a ruggedized, hard drive technology like RDX or cloud storage.

RDX allows the SMB data center to use a similar strategy to the hard drive transport described above but with a device designed to be ejected, transported and able to sit on a shelf, unused, for long periods of time.

Cloud backup providers like Infrascale, Datto and Dakota Cloud Recovery often provide an on-site backup appliance to meet local recovery demands. The appliance also replicates to a cloud storage target for use in case of a disaster. Many of these providers also include disaster recovery as a service capabilities.

Building an enterprise strategy for data movement

For businesses that have more than one location -- especially if they have multiple locations with IT personnel on site -- using the cloud or RDX becomes less appealing. The cloud is often redundant to what they already have in place and a portable type of product like RDX is too small for their needs.

These businesses need to be concerned about two types of data transport: data transported to meet an immediate disaster recovery (DR) crisis, typically the latest copy of data, and data transported to meet a long-term retention demand.

Data to restore applications in the event of a disaster should be electronically transferred to the secondary location, since this type of transfer is secure and reliable. It also can be performed in a variety of ways, ranging from primary storage replication to backup appliance replication. This allows the organization to match the RTO/RPO with the most cost-effective device.

These secondary storage systems should only be used to store the most recent copy of data and a small amount of retained data. While disks can be constantly scanned for data integrity, the cost to power, cool and house enough systems to store a typical enterprise's decade-long retention demands can break the IT budget.

Older data should move to tape or the cloud

Older data that must be retained for compliance or business reasons should be moved to either tape media or the cloud for long-term archiving. Moving data to the cloud is typically just another electronic transfer and many products have native cloud extensions. Of course, the cost of storing data in the cloud for long periods of time needs to be considered.

The second option, tape, is still the least-expensive way to store data long term. Tape does present a problem with transport, though, as most tape errors are caused by improper transport.

One option is to eliminate tape transport altogether by designing a "touchless" tape environment. In this design, a tape library is implemented at both the primary and secondary data centers. Data is backed up to tape at the primary data center as its being replicated to the secondary data center. Then, once in the secondary data center, data is backed up to the tape library there. Both libraries are sized so that cartridges never have to be removed from either system. Modern tape libraries can store hundreds of petabytes of information so never, or at least rarely, having to interact with the tape device is a realistic goal.

Over time, and as retained data capacity requirements increase, a dual library setup will become less expensive than cloud storage. But for many organizations, upfront costs put a secondary tape library out of reach. For these environments, careful transport and storage is a more economical approach, meaning the use of shipping containers designed specifically to transport tapes. The selection of a qualified tape storage facility, such as Iron Mountain, that can control environmental conditions is also important.

Transport and storage of secondary copies of data are often an afterthought for many organizations. After all, the likelihood of a data center failure is rare. But when these secondary copies are needed it is often a very dire situation, and the wrong time to find out that data was not properly transported or stored. Careful planning is required to ensure both the success of a DR plan and the recovery of decades-old data.

About the author:
George Crump is president of Storage Switzerland, an IT analyst firm focused on storage and virtualization.

Next Steps

Automated data movement at forefront of storage tiering innovations

Are file virtualization, data movement different from HSM, ILM?

Dig Deeper on Disaster recovery storage