On one of the LinkedIn groups I belong to, someone just posted this discussion item:
“Is Open Compute for Everyone? I guess the cloud is no longer technology’s darling. All the IT buzz now surrounds the Open Compute Project. If you are not familiar with the Open Compute Project, take a quick look at http://opencompute.org. You will see that this is really the brainchild of some bright engineers at Facebook, and the results are impressive.”
Here’s the comment I posted in reply, and I think I missed several more points in the heat of the moment:
“‘The cloud is no longer technology’s darling?’ Nonsense. If you look at the blogs, the conferences, the papers by IT analyst firms, the questions on various forums, the activity in standards group (about cloud interoperability), etc., it is still very much current, and quite in flux as people balance the promises of agility and lower cost with the risks in areas such as security and availability.
Open Compute is not in conflict with the cloud, it can complement it. A cloud provider could use Open Compute to create their data center, or an end user (with enough internal resources to handle the task) could decide to forgo the cloud but to use Open Compute to create a lower-cost on-premise center. The two issues are essentially orthogonal. Open Compute is a data center technology platform; the cloud is a sourcing and provisioning model. Apples and oranges.
By the way, you made a self-contradictory statement: ‘All the IT buzz now surrounds the Open Compute Project. If you are not familiar with the Open Compute Project…’ If it really were all the IT buzz, then everyone here would be familiar with it! You may be quite right that it is something people should learn about, although I would dispute that (if you don’t want to, or can’t, build your own infrastructure, then you could very well be interested in the cloud but not in Open Compute). But you’re exaggerating the point by pitting the two things against each other. Open Compute competes with COTS servers, not with the cloud.”
One of the things I missed in that reply is that the Cloud was never technology’s darling. The cloud is not a technology proposition, it is a sourcing proposition. It has nothing to do with which operating system will be run, or whether the hardware will have a brand name or will be a white-box machine. It has to do with choosing to plug into a utility socket vs. owning your own power plant, and with the financial and risk management aspects of this choice. So the assumption behind the question was flawed in the first place.
Furthermore, to the extent that the Open Compute project has anything to do with the Cloud (i.e., as a cost reduction option for people who chose to keep their servers in-house), then it competes with the IaaS model (Infrastructure as a Service), which is only one aspect of Cloud Computing.
I think the actual situation should be described as follows: for each aspect of IT (infrastructure, applications, platforms, collaboration tools, voice, etc.) there are roughly three choices, one of which has two variants:
- traditional outsourcing
- cloud computing
- in-house resources, which can be:
- commercial, off-the-shelf products
- assembled in-house from open-source components
The Open Compute project is indeed interesting, I don’t doubt that the inventors are smart, and the results may be impressive. But this by itself does nothing to challenge the Cloud model in general — it only provides another option specifically for infrastructure sourcing.