Hitachi Data Systems is announcing a global reseller agreement today with Diligent Technologies whose 25x Backup Data Compression is, IMHO, conceptually the best in the industry. I’m sure the HDS folks have thoroughly checked the technology and found that it actually meets the expectations set by the marketing.
So What’s With Deduping?
Marketing folks have long been blamed for coming up with pointless new terms when the old terms were good enough. Deduping, on the other hand, seems to have been invented by technologists uncomfortable with calling it what it really is: compression. Neville Yates, CTO of Diligent, made a point of telling me they don’t call it compression because techno-purists have strong opinions on the topic. But techies don’t sign the checks.
Double Whammy: Slower Adoption, Higher Sales Costs
Ideally, good marketing helps get ideas to market faster and makes it easy for customers to “get” the advantage. Which is why the choice of “deduping” to describe a compression technology is so blindingly stupid.
Quick, Choose One: 25x Compression or 25x Deduplication
I think this is very cool technology and an important feature that enables disk-based backup by making it cost competitive with tape libraries. So why make it harder to explain by inventing a new word with two variants: file-based and block-based? After all, it is only a feature – not a product.
Congrats to the Diligent team
for snagging HDS, whose unmatched OEM relationships will help them grow the Backup Compression market much faster than they could on their own. I wouldn’t be surprised to see EMC buying Data Domain, whose hash-based indexes Diligent and HDS will spend too much time putting down. Although why HDS’s big customers, like HP and Sun, wouldn’t just go direct to Diligent isn’t clear. But what the hey! Its cool technology offering real benefits – that is the important thing.
Your comments and hypotheses always welcome.
Update
Bill Andrews, CEO of another backup compression company, Exagrid, wrote a note explaining the differences between the vendors in this space. It seems evenhanded, so I’m excerpting it in the comments. If d2d backup is something you are interested in it is well worth your time.
These comments are courtesy of Bill Andrews, CEO of Exagrid. Anyone with amplification or amendment is welcome to comment.
There are four basic methods of data reduction.
At the client server level with an agent – take out the duplicate data at the block level – Avamar
Behind the existing backup server – take out the duplicate data at the block level – Data Domain, Quantum/ADIC/RockSoft
Behind the existing backup server – take out near duplicate block level data and then do byte-level delta – Diligent
Behind the existing backup server – take out duplicate data at the byte-level using byte-level delta technology – ExaGrid, Sepaton
To truly use the existing backup application and backup server you must be a VTL or NAS based target. In this way the backup servers are simply pointed at the NAS or VTL.
The market breaks out into two markets:
Mid Market / Small Enterprise
300 to 3,000 employees
Ethernet environment
Customers prefers simplicity and cost of a NAS target behind their backup server with data reduction
They do not want the cost or complexity of VTL
ExaGrid and Data Domain
Enterprise Market
Fortune 2000 type companies
Fiber SAN environment
Want VTL with data reduction for a variety of reasons
Quantum, Data Domain, Diligent, Sepaton, Falconstor
Vendors who are shipping NAS based disk based backup products with data reduction that sit behind the existing backup server
ExaGrid
Data Domain
Vendors shipping VTL based disk based backup products with data reduction that sit behind the existing backup server
Data Domain
Diligent
Vendors who have announced VTL disk based backup products with data reduction but are NOT shipping as of yet
Falconstor
Quantum
Sepaton
Avamar is not a disk-based backup product but is a true competitor to the existing backup products since you must take out the existing backup agents and replace them with Avamar. They compete with NetBackup, Galaxy, TSM, BackupExec, etc. They require agents on the all the application servers in place of the current backup application agents. In other words they do not sit behind the existing backup server.
Symantec is not a disk-based backup product, with PureDisk, as they applied data reduction to remote site backup which brings the data back to a central site to be backed up. In other words they do not sit behind the existing backup server.