permabits and petabytes blog oem data optimization for next generation storage OEM Data Optimization Solutions

Dedupe 1.0 Requiem

As a follow on to my initial post on the pending Data Domain acquisition (by someone!), here are some additional thoughts (take 2!)

  1. For NetApp, they were basically left at the alter acknowledging to the world their product technology/portfolio shortcomings. They opted for the “buy” vs. “build” strategy. EMC has now one-upped them with an offer they basically can’t counter with cash. We’ll see if the latest NetApp counter of cash + stock holds water. What if EMC does win? Where does this leave NetApp now? The market gorilla just stole their bride. This leaves a major void that they will need to address one way or another.
  2. For Data Domain, their management team is meeting with their tax accountants! Again, well done!
  3. For EMC, this get’s interesting. They basically have piles of cash and will do anything they want to maintain their market domination. It’s analogous to the person who drives a yellow Hummer and takes up 2 parking spots at the mall just because they can.

hummer-park

    Let’s look further at what’s really happening to the dedupe marketspace:
    Effectively Dedupe 1.0 is over – Data Domain won. What is dedupe 1.0? It was a technology created to solve backup pain. The problem of storing the same redundant copies over and over in a smaller footprint. Job done. But what problem did that really solve? Why is everyone still backing the same data up over and over? How does this address long term storage challenges? How does this help with primary storage optimization? It doesn’t. It can’t.

Dedupe for backup created a market that had many players scrambling to become the market leader in:

  1. Data Domain – Set the stage, defined the market, now cashing out at the end of the era. Their technology cannot scale beyond 30 TB. It’s a fundamental design issue. It’s easier to cash out than try to figure out how to build dedupe 2.0.
  2. FalconStor – Had some initial OEM deals, but plagued with quality issues. IBM opted to acquire Diligent rather than go with FalconStor. If EMC wins, FalconStor is out as a partner.
  3. Quantum – If EMC wins (and even if they don’t), this is a nail in the coffin. The solution just isn’t cutting it.
  4. IBM/Diligent – Diligent is yet another 1.0 type solution. What’s happening there lately?
  5. Sepaton -Watching the chess match with Data Domain to see who doesn’t get them and will race to try and fill the void. They also have an OEM deal with HP. This is yet another 1.0 solution however.
  6. Others – Have pretty much been left for dead on the side of the road.

What none of the vendors above address is anything other than the backup problem. How do you take a point product with limited technology and attempt to solve a broader storage challenge? Look at what EMC proposed in their analyst presentation the other day:

  1. Avamar – Source level deduplication. Acquisition made sense. Addresses that problem with one set of technology.
  2. Data Domain – Target level deduplication. Makes sense. They’re the market leader in this space. Addresses that problem with a second set of technology.

But, they talk about next generation archiving solutions as well. Which one of these products/technologies can handle that segment? Neither! How can Data Domain be taken seriously as an archiving destination? With their scalability limitations, there’s no way a large enterprise would ever deploy hundreds of DD boxes to store long term information. First one is great, but by the time you’ve hit your 20th, you’ve created a management disaster (sounds kind of like the NetApp theme we here from customers). So what else does it have? Quantum technology that obviously wasn’t good enough for them otherwise they would have bought it by now. Strike one. FalconStor? Has anyone bought that solution? Toss that one out going forward. Strike two. What about Centera? Anyone hear the crickets chirping when you ask for a Centera roadmap? Again, great first generation technology, but now what? Atmos??
I saw the market-techture slides released by EMC about a single control layer to “tie” together the point solutions they have assembled. Please, we’ve been hearing the middleware tune for 20 years and it has never really worked.
So, it looks like a great defensive move for EMC but doesn’t really help to deliver a next generation storage platform moving forward. It will just be yet another point solution they’ll need to try and figure out how to “integrate” with their product portfolio.
This market frenzy over Data Domain is basically marking the end of the Dedupe 1.0 era. It’s time to look forward to Dedupe 2.0. What is required to take dedupe “technology” and make it mainstream in all storage. Backup, Archive, Value (Tier 2/3), Primary? What are the requirements for technology that can solve all those challenges? Scalability? Reliability? Performance? What vendors play in this next generation game?

I would love to hear your thoughts! Dedupe 2.0 is coming to theaters near you this summer!

No Comments