Bloomberg News reports that search giant Google Inc. is considering building its own processors based on technology from ARM Holdings PLC.
Building its own chips would allow Google to better manage the interaction between hardware and software, an unnamed source told a Bloomberg reporter.
This would be a blow to chipmaker Intel Corp., which is heavily dependent on server processor sales to make up for a flattening market for personal computer chips. Google is possibly the world’s largest single purchaser of server chips; estimates of the number of servers Google has range from 2.3 million to 2.6 million. Bloomberg says Google accounts for more than four cent of Intel’s server revenue.
A job posting for a position as a digital design engineer at Google in Mountain View, Calif., had attracted 98 applications by press time. The ad specifies experience designing application specific integrated circuits (ASICS).
“You develop from the lowest levels of circuit design to large system design and see those systems all the way through to high volume manufacturing. Your work has the potential to shape the machinery that goes into our cutting-edge data centers affecting millions of Google users,” the ad reads in part.
The network-intensive demands of Google’s enormous infrastructure – there’s more backend traffic than user-facing traffic, Google engineer Amin Vahdat told a conference this spring – make it a perfect use-case for software-defined networking or network virtualization, wherein most of the network functions above the physical link layer take place in software. Google refers to its software-defined network as B4.
Significantly, virtualization makes it possible to run networks on inexpensive commodity hardware rather than expensive switching equipment. Making its own chips for that infrastructure could be a considerable saving for Google.
Data Center Innovation by Cisco and IBM
Register for this webcast to learn how joint data center solutions from IBM and Cisco can increase resource utilization, support business continuity, provide higher availability and faster provisioning, and help you achieve productivity gains through virtualization, optimization, energy stewardship and flexible sourcing.