The public cloud: Friend or foe for storage vendors?
When storage vendors sell to public cloud providers, do they wind-up competing with themselves?
Last year, storage vendors were all about cloud. They saw major-league opportunities in the private, public, hybrid, and federated versions. No cloud was too big or too small. In fact, because clouds were "infinitely scalable," there was no limit to the number of yotta bytes they could sell.
Storage users and data center storage administrators in particular were decidedly more sanguine. You say cloud is a new services delivery model? Hey storage vendors, where have you been lately? We've been all about services delivery for some time now. Tell us something about cloud we don't know. And by the way, how does cloud make our jobs easier?
Cloud storage software start-ups have not done well over the last few months. Parascale is on life support and MaxiScale is struggling. Even EMC stumbled with Atmos Online. On the other hand Cleversafe is growing I think because it emphasized cloud data security right from the start and is well-connected to the open-source community. More on that later.
The major storage vendors now appear to be breaking into two camps with regard to how they approach storage in the cloud--one focused on private, the other on public clouds. And while both camps profess to loving both clouds at the same time, EMC's current marketing message is all about the Journey to the Private Cloud. On the other hand, during recent sessions for analysts sponsored by 3PAR and NetApp, both were showcasing their successes in the public cloud space.
Make no mistake. The real competition in the cloud arena is between private and public clouds. Right now, I'm thinking that EMC's private cloud journey may be the right approach. Here's the question I'm wrestling with: When storage vendors sell to public cloud providers, do they wind up competing with themselves?
Consider the following scenario: In 2009 a well-known storage vendor sold a petabyte of new arrays to a Fortune 100 company. This year, the Fortune 100 company signed a major deal with an "IT as a service" provider. Much of the storage the Fortune 100 company bought in 2009 walked over to the service provider. The sales rep assigned to the Fortune 100 account was heart-broken until the vendor he worked for decided to credit him for additional sales to the ITaaS vendor intended to support his Fortune 100 customer. The sales rep is now happy. The service provider is happy. The Fortune 100 customer is happy. In this case, the storage vendor is happy too. It's a win-win-win-win situation. Or is it?
Data is moving from corporate data centers managed by internal IT staff to data centers, either owned or managed by service providers of "fill-in-the-blank as a service" and all in the name of "cloud." Is the long-term impact to storage vendors selling to these service providers net positive or negative?
Selling to corporate data center customers has become a well-honed practice of the successful storage vendors. Value-add features like thin provisioning, logical partitioning, data snapshotting, etc. expressed in array-resident software allow vendors to justifiably boost selling price and profit margin well over that which they could deposit in the bank if they just sold JBOD arrays (just a bunch of disks in a box). Array-based software harvests the big profit dollars.
Service providers are now buying those same arrays and the same value-add software. In the case above, they are sometimes buying them to support the same customer. When data moves from the vendor's in-house data center arrays to the same vendor's arrays run by a service provider, where are the better profit margins? I dare say that service providers are getting better deals on average from storage vendors for the same products than enterprise storage buyers. Ergo, gross profit opportunities are likely better for the storage vendor on the enterprise data center side of the deal. And, assuming the service provider keeps the enterprise customer happy and thereby takes on more of the customer's data, this situation becomes irreversible.
Here's another more ominous trend: Storage vendors typically outsource array manufacturing to companies like Foxconn and Sanmina. Cloud services providers are now sourcing hardware directly from these same manufacturers, layering software on top that they either get free via open source or write themselves. In this case, storage vendors become middlemen that can easily be cut out of a deal--zero profit dollars when this happens.
I've mentioned this particular trend to some storage executives. The common retort is: "Do you think that open-source software can really stand up to the demands of a 24-7 production data center?" Go ahead and smile all you open-source devotees.