Facebook's Open Compute Project (OCP) could potentially destroy the business models of the mainstream server, storage and networking vendors. It is an attempt to turn back time and return to a pick-'n'-mix server, storage and networking world where suppliers happily build to common interfaces, and component sets aren't …
"That's why the OCP IT model will fail in the real world outside the hyperscale data centres of Amazon, Facebook, Google, Yahoo and the other enormous cloud IT service operators."
It seems somewhat likely to me that they are envisioning a future where that's where almost all of the compute hardware actually lives, and thus where the economies of scale end up focusing the market. Personally, I don't think such an one-sided outcome is going to happen for a very long time, if ever, but a lot of hype seems to assume it's inevitable.
First thing I thought of when I saw "OCP" was "Omi Consumer Products"; the corporation in Robocop.
Well, I kept misreading it as OCD, which should ensure that I'll keep my hands clean of it.
it will help anybody who uses "the cloud"
You overlook the growing number of IT shops who move applications into servers hosted by Rackspace, Amazon, etc. I don't want to bother with buying computing hardware from suppliers and plugging them in either - I lease computing services as needed. Because OCP will help Amazon build better datacenters it helps me and more and more other people every day.
Of course we still need some hardware, for running back-office applications, but I'm confident we'll still be able to buy cheap hardware for mail servers, Quickbooks, etc.
I realize my view is skewed towards small companies or small engineering organizations in larger companies, and there's a whole world out there of purpose-built datacenters and there's a whole other world out there, but you can't just ignore AWS-style hosted compute services in your OCP analysis.
"transmission, the suspension, the breaking"
I think the author meant a device to slow a vehicle, rather than damage it.
Even if he worst happens and the Earth splits asunder and fire and locusts spew forth and everyone+dog starts making this 1970s stereo-style kit, I'm *fairly* sure those who sell the other sort of kit will not only sell the new stuff but offer to put it together for you.
I mean, on one level that's what Gateway and Dell started out doing, and they drove the "box in a box" merchants out of business in about four years. Used to be you could walk into any electronics store and see two dozen brands of off-the-peg computer waiting to be bought. No it's not quite the same thing, but then again, yes it is.
Calm down for Azathoth's sake. It's Friday. Go for a pint.
One word. VMWare. Makes the hardware irrelevant.
I am at a loss. This article appears to be attempting to argue against the economies of scale by presenting arguments in favor of the economies of scale. There is not much of a counter argument unless you happen to be dependant upon one of the companies that have margins based on artificial scarcity and accelerated obsolescence.
I am more inclined to follow the lead of facebook, a company that has risen spectacularly from nothing less than a decade ago to its current prominent position. They have been particularly canny with their decisions thus far. The only criticism that could reasonably be leveled at facebook is that they have played fast and loose with privacy. However, because of the phenomenon of 'group forming networks', there was a good argument that they were vulnerable to destruction prior to reaching their current 'critical mass'. MySpace was, I am convinced, entirely defeated because of the difference in the 'promiscuity' of data that facebook leveraged. Arguably, the privacy was doomed from the outset and at least facebook has thus far been pretty faithful in keeping that data within its boundaries.
One good thing that would come out of breaking out and commoditizing the core elements of the data center would be to highlight the current disasterous balkanization of critical infrastructure elements, particularly bandwidth.