Intel Corp. and Alphabet Inc.’s Google Cloud said Tuesday they have launched a jointly designed chip that could make data centers more secure and efficient. The E2000 chip, codenamed Mount Evans, takes over the job of packing data for network cards from expensive central processing units (CPUs), which were originally responsible for calculating those tasks.
The chip also provides better security between different customers who may share CPUs in the cloud, said Amin Vahdat, vice president of engineering at Google.
The E2000 chip is made up of basic processors called cores. There can be hundreds of cores on a chip, so sometimes information can be lost between them. For environments where network data transfer reliability is critical, the E2000 creates secure routing for each core to prevent this from happening.
Businesses and organizations around the world are running increasingly complex algorithms and progressively larger data sets at a time when performance improvements in chips such as CPUs are slowing down. As a result, cloud computing companies are looking for ways to make data centers themselves more productive.
While the new chips are being developed with Google, Nick McKeon, who leads Intel’s networking and edge group, said Intel also has the option of selling the E2000 to other customers.
Google Cloud will soon begin offering the E2000 in a new product called C3 VM, which will be powered by Intel’s fourth-generation Xeon processors. Intel says the Xeon chips are currently Intel’s most powerful CPUs, and Google Cloud is the first cloud service to deploy the latest generation of these chips.