Effective remote backup with data deduplication

Service provider takeaway: Service providers can use data deduplication technologies to create an effective remote data backup strategy for customers who employ many remote offices.

You thought the backup system at Bailey Building & Loan's Elmira office was working just right last year. Every night the backup application ran automatically and ejected the tape. Every morning, it was Uncle Billy's job to insert a fresh tape and mail the prior night's tape to the home office in Bedford Falls. Knowing Uncle Billy isn't an IT pro -- and frankly isn't the most reliable of execs -- you even set the system up to email you if the backup job failed.

This morning, you listened to a voicemail from Uncle Billy in which he explained that someone broke into the Bailey Building & Loan's Elmira office and stole its NAS appliance. Even worse, after some investigation, you learn that months ago, Uncle Billy had decided not to mail the tapes to the home office after reading about other companies losing tapes and CDs in transit. Of course, since the thieves took everything but the kitchen sink, they're gone too, and the latest one in Bedford Falls is months old.

How can you prevent this problem from repeating itself? Start by giving tape drives the boot. You can help your customers (whether Uncle Billy, his more reliable nephew George Bailey, or even nasty Mr. Potter, for that matter) with remote data backup by switching to disk-to-disk backup across the Internet using data deduplication, ensuring that your customers' backup data gets offsite. By breaking the data on a client's NAS into blocks and only sending unique blocks across the Internet, deduplication technology reduces the amount of data that needs to be sent by 20- to 100-fold. That reduction means that the typical remote office's DSL or T-1 line can handle the nightly remote data backup load, without a costly infrastructure upgrade.

More on data deduplication Data deduplication in the storage channel Storage compression and data deduplication tools Leveraging data deduplication for VARs

Navigating new storage technologies isn't always easy for customers; service providers can help them make sense of new developments and choose the right one. New-generation backup software (such as EMC's Avamar, Symantec's NetBackup Puredisk or Asigra's Televaulting) not only deduplicates data at remote offices before sending it across the Internet, it also can perform global deduplication across multiple offices so all the remote data and programs that are stored on multiple NASes in multiple offices are only backed up once. This means, for example, that only the first instance of an email attachment sent to multiple remote offices is stored on the device when the backup takes place.

While all of these products can protect Windows-powered NASes, you'll have to do a little homework to find a product that can protect NASes that run their own proprietary operating systems.

Service providers with a data center and an eye toward being a managed service provider (MSP) can use Asigra's Televaulting to build a single repository to store data from multiple clients while ensuring that data is protected.

VTL-based deduplication

For customers that just don't want to give up tape, service providers can still help take advantage of deduplication technology and remote data backup by installing a virtual tape library (VTL) that supports replication, such as Quantum's DXi series or FalconStor's VTL software, in remote offices. Your clients can continue to use their favorite data backup software, and their shiny new VTL will deduplicate the data and then replicate it to another VTL in the main data center.

VTL-based deduplication and replication, however, aren't a panacea. While they allow clients to continue using the remote data backup software they know and love, they are somewhat less efficient. VTL-based deduplication will only remove data that is duplicated at each remote office. On the other hand, customers who choose to use technologies like those offered by Avamar et al. can take advantage of disk-to-disk backup and replication technologies that allow data to be deduplicated globally. In addition, with the VTL setup, your client's backup software will be a bit confused, as the replicated data will make it look like one tape is in two places at once.

The bottom line is that remote office tape backup is so 20th century; at one time, it might have meant you'd have a wonderful life, but not anymore. Have your clients deduplicate data and use the Net. You and Uncle Billy will sleep better, and you might have the opportunity to pick up some services revenue to boot.