Wikibon Blog

After sleeping on this I thought I’d share a few additional thoughts to my initial post.

What does this move by EMC tell us?

First it comfirms what we already know– when it comes to competing, EMC plays for keeps. There are three additional thoughts I have:

1. EMC realizes it can’t grow its core storage business organically and needs to acquire to grow.
2. This is a defensive move, albeit an aggressive one, to stop DDUP from getting in NTAP’s hands.
3. The market continues to consolidate and companies like Commvault, Falconstor, Sepaton…and even 3PAR and Compellent are worth more today than they were yesterday.

EMC’s ’vision’ for Data Domain is better than NetApp’s in my opinion but its ability to integrate this vision is suspect. Specifically, NetApp was essentially going to run DDUP as a standalone entity with no grand integration plans. EMC on the other hand is talking about building so-called ’Next Generation Data Protection’ (NGDP).

What does NGDP mean?

It means extracting intellectual property from various data reduction technologies and placing that IP at the right level of the storage hierarchy, creating sets of data reduction services that can be invoked as needed. For example, you might have compression for primary storage and data de-duplication at the source or at the target for disk-based backup and single instancing, etc, across the portfolio. Each of these data reduction technologies would ostensibly speak the same ‘language’ such that I could: 1) apply the best technology where it’s neeeded (as a service) and 2) move data throughout the hierarchy without having to ‘re-hydrate’ data and incur the normal penalties associated with such re-hydration (processing power overhead, performance loss, energy consumption, etc.). This would make my data reduction strategy ubiquitous, transparent, cumulative and incremental, generating savings of 35-50% or more across primary and secondary storage.

Wow! Where do I sign?

The ‘BUT’ is with Data Domain, Quantum, Avamar, compression, single instancing and who knows what other data reduction technology that comes along, I feel like pigs will fly before this all comes together.

Soooooooooo, we’re potentially talking about ‘silo-central’ here and I’m not sure that EMC is the silo-buster for many years to come. This means more customer integration pain and lots of services opportunities for EMC. Hmmm….maybe not such a bad business model for the vendor but buyer-beware.

I guess the bottom line question would be: Is this the best use of $1.8B for EMC’s shareholders? The answer is it depends on how big the market is. Tucci says it’s $10B but DDUP is mostly a tape replacement technology. The brilliance of DDUP is you don’t have to replace your tape backup processes to plug in its gear. The hardest thing to change in IT is process, not technology, hence Data Domain got it right. But that’s a $3B market, not a $10B one. The rest of the $10B is doing other data protection on primary storage (replication, copy services, compression, etc.). Not sure I need to own DDUP for that.

As I’ve written before, I’d rather see EMC invest in smaller companies to attack the $20B information management market (email archiving, eDiscovery, enterprise search, etc.), or shore up its cloud storage/ATMOS strataegy, than overpay for DDUP in a market (core storage for the data center) that isn’t growing.

It’s easy to understand the allure of Data Domain. Aggressive, hungry, smart, arrogant (in a justified way) and ready to compete. EMC will point to its successful track-record of acquisitions which, considering most mergers fail to meet business objectives is pretty good. But not all of EMC’s buys have been poster children for successful M&A. Legato, eh. Avamar took years to integrate. Documentum, do-over. Even VMware’s ascendency, while meteoric, has largely been a function of dis-integrating the company (at least to date) and not any EMC integration prowess.

Data Domain will be a story of integration with lots of cultural force-fitting, product and channel rationalization, organizational shuffling, etc, because after all is said and done…’data de-duplication is a feature, not a market…’