"In other to accomplish this feat, Facebook enlisted the help of Intel (NASDAQ: INTC ) for its silicon photonic technology, enabling transfer speeds of 100 gigabits per second, as well as a low enough latency for components to be spread out within the server rack. As a result, Facebook has been able to develop server designs that are 38% more efficient and 24% less expensive to build and operate than today's state-of-the-art servers. This approach not only allows for data center engineers to create custom modular server designs to exact specifications, but also maximizes the life cycle of components."
[I am telling you, in the end, it will be servers that define Intel. The deconstructed model of servers with centralized CPU's and the photonic "bus" is the future. The CPU's could sit in a cryostat of liquid nitrogen for ultra efficient cooling...no wonder INTC is getting back into memory business...it will all be wrapped around the (Mega/Giga)xTerabps servers of the future. ]
Also, if you were saying it was poor.....for a change, that means you were saying it's normally good. LOL......you were saying it was GOOD for a change, yet the article you linked is really bad for Intel.
Did you not read the rest of the article? The article actually predicts doom for Intel, sir.
The age of homegrown
We are no longer living in a one-size-fits-all world for data, and this reality directly threatens the one-size-fits-all approach of the server monolith. Data needs do in fact vary greatly between organizations, which emphasizes the growing need for custom-tailored data solutions. The organizations that adopt an Open Compute Project-type approach are poised to save gobs of cash in building and operating their proprietary data centers to exact specifications. However, this cost benefit may come at the expense of Intel and its data center storyline, which is already leaking air.
"However, this cost benefit may come at the expense of Intel and its data center storyline, which is already leaking air."
The punch line is the last one and the key word is "MAY". You seem to be the one predicting doom for Intel. On a quite regular basis too. The article is forecasting "change" in server designs and with change there will be some winners and some losers. Intel may lose but more likely it will gain larger revenues/profit in a larger market.
Facebook is trying to squeeze costs out of their servers to increase their bottom line. Intel and Facebook started the OpenSever project.
The expertise of FOOLISH writers is very spotty and the quality of the work lower than that of Seeking Alpha. The FOOLISH writers are just lead in draw for the real products: The FOOLISH services be sold at the bottom of the page. You know: "DEATH TO THE PC" "research".