intel cloud


As cloud computing gets bigger, some applications have to run on thousands of computer servers at once. Intel on Wednesday offered a way to tailor its server chip for this work.

Intel said on Wednesday that it would enable its Xeon processor, their server grade processors used in big computing environments, to handle a customer’s large-scale proprietary algorithms as if the chip were customized for that process.

The program, which calls for the integration of field-programmable gate array (FPGA) technology into the same package as the Xeon processors, is part of a larger effort underway at Intel to offer organizations a wider selection of silicon products that are more adaptable to individual workloads.

FPGAs integrated with standard Xeon E5 processors will enable companies to program the processor chips to handle specific workloads. 

Diane Bryant, general manager of Intel’s data center group said “applications should work twice as fast as before, As cloud computing grows, the infrastructure of it has to change.” 

Diane Bryant continued, “This is for ‘scale-out’ computing, when you want to run a single application across thousands of servers at once,” Ms. Bryant said. “Five years ago, we talked about how many applications we could make run on a single server.”

Intel has been aggressive in addressing the rapidly changing demands in the data center. Where once the chip maker would release a handful of models of each processor they released, now they can roll out two dozen or more to offer greater optimization for particular workloads. In addition, the company is leveraging its low-power Atom platform for dense microservers—and trying to fend off a challenge by ARM and its partners, which includes longtime rival Advanced Micro Devices—while introducing an even smaller and more energy-efficient Quark family for the Internet of things.

How many customers will need this level of customization is unclear of course but the change also shows different cloud computing is from its predecessors. Ms Bryant said “Intel was talking with both cloud service providers and telecommunications about using the chip, It is an easy way for them to download an application and make it run faster, then change their minds and reprogram the chip to download another application”.