I spent the day courtesy of Gestalt IT at the Software Defined Data Center (SDDC) Symposium 2014 in San Jose, CA. The symposium featured industry thought leaders and sponsors who have some really interesting products around the software defined space and more “traditional” hardware solutions. This is year two of the symposium and well worth the 6 hours of flying it took for me to get here. SDDC is a pretty interesting early stage design concept and I walked away from the event a bit sobered by how far we have to go as an industry to reach a level of maturity that the non-web 2.0 enterprises can deploy.
Enterprise vs. Facebook and Google
Early on, presenters did a good job of differentiating SDDC in the enterprise versus SDDC in hyper-scale homogeneous application environments such as Facebook and Google. These companies control the entire stack to provide a single set or well understood applications as well as Computer Science PhDs to make sense of it all. Most enterprise use cases don’t give this type of stability or warrant the level of resources dedicated to Google and Facebook. Of course, much can be learned from these implementations for the tradition IT organization. The question left, what’s the level of maturity for non-web 2.0 implementations.
SDDC is whiteboard material
The session that got right to the meat was the panel discussion titled “SDDC in the Sunlight: Real World Use Cases Panel.” I couldn’t help but walk away thinking that SDDC was something a little less than vaporware for most enterprise use cases.
I need to follow the statement by saying I’m a big fan of the SDDC idea. I believe the goal of abstracting the data center and providing the organization with a construct that is independent of the physical underlay has tremendous value. But the overall of theme of most of the panel seemed to be a mix of repurposed cloud concepts sprinkled with devops and mixed with a bunch of SDN and traditional storage. I didn’t hear a cohesive example of a real world enterprise implementing SDDC as we see in FB and Google. If I were a CIO unfamiliar with SDDC, I’d have walked away dismissive of the approach due to the lack of clear products or even frameworks that help guide the migration to SDDC. The session helped highlight without a product or framework SDDC is little more than whiteboard material.
The panel was lead by Howard Marks, an Independent Storage Analyst, and included the following SDDC thought leaders.
Dave Write – SolidFire
David Meyer – Brocade
Jesse Laurent – Simplivity
Michael Letschin – Nexenta
Ramana Jonnala – Coho
Victoria Grey – Gridstore
No “SDDC products”
All of the vendors on the panel had shipping products and even seem to have pretty decent strategies around levering their products in abstracted models. Solidfire, for example, is a one of many “scale out” storage solutions that allow for the modular expansion of enterprise class storage arrays and strong software management layer. However, solutions bridging the gap between effective management software and the abstraction of the entire data center are not available from any single vendor.
This isn’t a knock on any vendor. Creating a bridge between the different infrastructure stacks and software is a very complex challenge that takes more than one vendor. In addition to the massive challenge are the competing agendas of each vendor. I’ve written that OpenStack may be a way to the SDDC because of the open source nature of the project. Even so, we are a long way away from the software defined data center in the enterprise.