Having paid very close attention to the Open Compute Summit this week because a client was there (and a member). What I have come to see is that Open Compute IS a good idea, but not for the reasons people want to believe. Let me explain…
Open Compute is good for Facebook. How doe we know? They saved $1.2B implementing Open Compute Hardware. That’s a great stat to be able to share. For Facebook. But what about other companies with say 1,000 employees whose size or revenue is less than $1.2B – what’s in it for them? How does Open Compute solve for their challenges? I for one am still waiting for an answer to THAT question.
I get standards, they have proven useful since before the metric system or the 8″ vs 5 1/4 inch floppy wars, or VHS vs BetaMax format. What Open Compute is working against is an entire industry that has developed and built to standards for 30 years. Power voltages, rack sizes, programming languages. Throwing the baby out with the bathwater is not gonna happen quickly, if at all because there are not enough companies at the size and scale to shave $1.2B off of their costs – and will spend to do it.
That said, the notion of Open Compute that I do like is that it can foster similar approaches in figuring out what DOES work for your company, over whatever period of time makes sense to adapt to standards that work today and will likely work in the future. In other words, you don’t HAVE to be like Facebook but you CAN work with your vendors to figure out what does make sense, is interoperable, and will likely remain interoperable for your company. You have the standards to augment what isn’t standard and still save money based on what you want to do/spend to fix it. It still may be a solution looking for a problem.
The alarm bells that are going off as a result of seeing where this can end up is astounding. I haven’t quite figured out if they are a new business model, a foregone conclusion, or electronic socialism.
What appears to be happening is that there is a rush to commoditization up and down the stack of what delivers value and makes stuff work. Data centers, servers, storage, operating systems, networking, all going ‘open’. Open Stack, Open Source, Open Compute. I call it OpenX. All these ‘opens’ are telling me that the race to bottom of cost is happening in the name of innovation. Innovation itself is a cost, not revenue. Is R&D in the cost or revenue side of a business? If you’re innovating something that will wind up being free or difficult to charge money for how do you pay people to keep making free stuff. Innovation doesn’t pay the mortgage. Or employees.
When we get to the bottom, what’s left? Apps. Mobile apps, desktop apps, endpoint consumed apps. I believe it’s why IBM bought Softlayer, IO got into the cloud space, and Microsoft bought Nokia. You can argue about the financial reasons for these deals all day long, but the market force reasons are this – if you don’t have an integrated stack – data centers, servers, storage, operating systems, networking, and cloud – you miss out on being valuable to the only things that will make money – applications. And capturing, washing, and understanding data, but that’s another post.
Think about it – when we commoditize things – wheat, corn, computer chips, the things with the value are food, fuel, and applications. The successful commodity businesses are those that operate at scale, with high risk, and razor thin margins. Look at Lenovo. They just staked their claim in the hardware business. They will make hardware and a ton of it. At scale, low margins, high risk. They will also own enough of the market to reduce the risk of innovation that’s not theirs.
Time will tell how it all shakes out – it always does. Keep your eyes focused on how to make money delivering vale – it’s still the only yardstick that measures how your business is doing. And there is no app for that.