The Enterprise Game Changer
The rapid emergence and acceptance of deduplication into the mainstream of enterprise storage in the last 18 - 24 months has been nothing short of phenomenal. Enterprise storage is a segment of the computer industry that typically measures change in years, not months. While changes in specific media formats occur, the idea that a totally new technology can gain such widespread acceptance and adoption in so short a time is as revolutionary as the technology itself.
The problem that emerges when a technology goes from the drawing board to corporate acceptance so quickly is that it is easy for individuals to start to assume that every product that supports deduplication implements deduplication the same way or delivers the same level of benefits. While it might be true that different products implement deduplication technology using similar underlying algorithms for data reduction, they do not all scale in the same way.
Most VTLs or disk-based backup appliances that embed deduplication technology are typically limited in two ways: the number of front end controllers that can deduplicate the data and the amount of back end storage. These limitations affect how fast they can deduplicate backed up data and how much data they can store. This inability to scale forces companies to purchase another VTL or disk-based backup appliance - either a larger unit and migrate data from the old VTL or disk-based backup appliance to the new unit or purchase a second unit.
Neither option is particularly desirable since in both cases they still have the same inherent performance and capacity limitations as the initial disk library. However using a second deduplicating VTL or disk-based backup appliance requires it to start the deduplication process from scratch so much of the new performance and capacity in the second unit is wasted re-duplicating data that the first unit previously reduced. This pattern holds true for each succeeding deduplicating VTL or disk-based backup appliance that companies buy - none of the benefits are shared between units so companies end up with lots of silos of deduplicated data.
I bring out these points not to denigrate deduplication. Deduplication is better than just using dumb disk libraries with no deduplication at all. But there is more to consider when bringing deduplication in house than just buying and bringing any product that supports deduplication. Companies need to buy a product with scalable architectures that support the technology of deduplication.
Deduplication is a game-changing technology as it changes how enterprises manage their data. But it does more than just change the game for how data is stored, it changes the architecture of the products that companies can and should use to store it. The sooner companies understand this architectural change and the need for grid storage in enterprise deduplication, the sooner they can stop spending as much time managing their backup and storage infrastructure and start spending more time managing their business.
The problem that emerges when a technology goes from the drawing board to corporate acceptance so quickly is that it is easy for individuals to start to assume that every product that supports deduplication implements deduplication the same way or delivers the same level of benefits. While it might be true that different products implement deduplication technology using similar underlying algorithms for data reduction, they do not all scale in the same way.
Most VTLs or disk-based backup appliances that embed deduplication technology are typically limited in two ways: the number of front end controllers that can deduplicate the data and the amount of back end storage. These limitations affect how fast they can deduplicate backed up data and how much data they can store. This inability to scale forces companies to purchase another VTL or disk-based backup appliance - either a larger unit and migrate data from the old VTL or disk-based backup appliance to the new unit or purchase a second unit.
Neither option is particularly desirable since in both cases they still have the same inherent performance and capacity limitations as the initial disk library. However using a second deduplicating VTL or disk-based backup appliance requires it to start the deduplication process from scratch so much of the new performance and capacity in the second unit is wasted re-duplicating data that the first unit previously reduced. This pattern holds true for each succeeding deduplicating VTL or disk-based backup appliance that companies buy - none of the benefits are shared between units so companies end up with lots of silos of deduplicated data.
I bring out these points not to denigrate deduplication. Deduplication is better than just using dumb disk libraries with no deduplication at all. But there is more to consider when bringing deduplication in house than just buying and bringing any product that supports deduplication. Companies need to buy a product with scalable architectures that support the technology of deduplication.
Deduplication is a game-changing technology as it changes how enterprises manage their data. But it does more than just change the game for how data is stored, it changes the architecture of the products that companies can and should use to store it. The sooner companies understand this architectural change and the need for grid storage in enterprise deduplication, the sooner they can stop spending as much time managing their backup and storage infrastructure and start spending more time managing their business.
Leave a comment