Blocks and Files Facebook's Open cypher Project (OCP) could doubtless destroy the business models of the mainstream server, storage and networking vendors. it is a shot to turn back time and come back to a pick-'n'-mix server, storage and networking world where suppliers blithely build to common interfaces, and element sets are not integrated for larger overall efficiency.
Facebook is its own massive systems measuring device, but the rest people don't seem to be and OCP standards adoption by enterprises could lead on to much more valuable IT element acquisition and implementation processes.
How do I arrive at now, a point where i'm trashing the spread of Facebook's terrific OCP ideas?
OCP is Facebook's gigantic conceive to get the IT business to deliver parts in ways that suit Facebook; it wants the speed of hardware innovation to induce nearer thereto of code by disintegrating rack and server parts so that they'll develop independently but still work together through common interfaces.
In this manner, server CPUS and memory and interconnects and power supplies shouldn't have to be compelled to develop at a four- or five-year refresh cycle rate with all the parts in balance. Instead, for instance, one could separate out the ability supply boxes and have them supply all server parts from their own shelf within the rack rather than being integrated with the server. One could have memory and processors loosely connected via Intel's new chemical element Photonics cable with 100Gbit/s information measure and very low latency - so you cannot replace either CPUs or memory without heavy the other.
Do this and Facebook can purchase in its server parts from low cost Taiwanese Original design makers (ODMs) rather than complete servers from Cisco, Dell, HP or IBM. Facebook showed a Dragonstone database server at the metropolis OCP conference yesterday. It's for handling cold knowledge and is made from - and uses - 3.2TB of Fusion-io PCIe ioScale flash card memory; there aren't any disk drives in it in the least.
Elsewhere within the OCP knowledge centre vision, Facebook is talking about victimization huge and fat and slow disk drives with spin-down capability for storing cold knowledge, just like the photos that very few individuals will ever access again. it'll wish these within the OCP rack design, not the industry-standard 19-inch rack design that enterprises use. And it will not get OCP-standard cold knowledge disk vaults from EMC or NetApp, preferring the ODMs again; they're cheaper and it will not have to be compelled to pay EMC or NetApp knowledge management code costs because it produces that code itself.
What Facebook wants is affordable for itself but unreasonable for everyone else. If you are one in every of OCP's chosen vendors - Ohio lucky, lucky Intel if chemical element Photonics takes off - then wealth and Facebook-blessed glory expect you... as Fusion-io is hoping. but if you are get into the cold, with merchandise like servers or storage arrays or network boxes, then OCP could be a dagger aimed straight at your business model's heart, and blood is dripping from it already.
The currently unspoken marketing aim for mainstream server, storage and networking vendors is to prevent their customers adopting OCP standards and concepts. Heaven forfend!
At an equivalent time they have to play nice with Facebook and be OCP supporters, experimenting with producing their own stripped-down OCP-style product but solely merchandising it to Facebook-style customers while everyone else prefers the full-fat merchandise with full-fat profit margins.
Note that storage suppliers EMC, Fusion-io, Hitachi GST and Sandisk have all joined the OCP organisation. Fusion sees big bucks coming back from its membership - ditto SanDisk I dare say, conjointly conjointly Hitachi GST with helium-filled huge fat disk drives being hopefully adopted by OCP. but EMC? Why the hell has it joined? apprehend your enemy, I think.
If Facebook was buying a lot of cars a year then affirmative, we have a tendency to could see it would wish to separate engine development from that of the clutch, the transmission, the suspension, the breaking so on to hurry its ability to use the latest and greatest parts. the rest people ar completely different. we have a tendency to get cars in ones and twos, tens and 20s if we have a fleet, and don't wish to integrate the bits ourselves. we want a get and drive a vehicle, not a get and build box of bits.
That's why the OCP IT model will fail within the universe outside the hyperscale knowledge centres of Amazon, Facebook, Google, Yahoo and the different enormous cloud IT service operators.
Don't believe me? Shout out why within the Reg OCP forum topic
0 comments:
Post a Comment