Google Cloud CEO Thomas Kurian speaks at the Google Cloud Next event in San Francisco on April 9, 2019.
Michael Short | Bloomberg | Getty Images
Google Are trying to make cloud computing more affordable through customization armBased on server chips. The new processors will be available later in 2024, the company said Monday at the Cloud Next conference in Las Vegas.
With its new Arm-based chips, Google is going after rivals like Amazon and Microsoft, which have employed similar strategies for years. Tech giants compete fiercely in the growing market for cloud infrastructure, where organizations rent resources in remote data centers and pay based on usage.
Google parent company Alphabet still gets three-quarters of its revenue from advertising, but cloud computing has grown faster and now accounts for almost 11% of the company’s revenue. The division, which includes enterprise productivity applications, also turned a profit.According to Gartner, Google will account for 7.5% of the cloud infrastructure market by 2022, while Amazon and Microsoft together control about 62%. estimate.
market leader Amazon Web Services launched Graviton Arm chips in 2018. “Almost all of their services have been ported and optimized on the Arm ecosystem,” Chirag Dekate, an analyst at technology industry research firm Gartner, said in an interview with CNBC. Graviton has secured business from companies such as Datadog, Elastic, Snowflake and Sprinklr.
Alibaba Release Arm processors in 2021, and Microsoft The same thing was done in November.
Arm isn’t new to Google, which started selling access to virtual machines (VMs) that use OracleEarlier this year, the company backed Arm-based chips from startup Ampere.
For organizations looking to reduce cloud computing spending due to economic concerns, porting applications to Arm machines makes sense.When Arm Holdings filed to go public last year, it pointed to Amazon’s claim that Graviton was 40% more cost-effective than comparable server instances, such as the common “x86” model used by ARM AMD and Intel processor.
Google has used Arm-based server computers for internal purposes to run YouTube ads, BigTable and Spanner databases, and BigQuery data analysis tools. A spokesperson said the company will gradually move Axion to cloud-based Arm instances as they become available.
Datadog and Elastic plan to adopt Axion, as well as OpenX and Snap, the spokesperson said.
Wider use of chips based on the Arm architecture could lower the carbon footprint of certain workloads. Thomas Kurian, head of Google Cloud, wrote in a blog post that virtual slicing of physical servers containing Axion chips is 60% more energy efficient than similar virtual machines based on the x86 model. Arm chips are popular in smartphones, offering shorter instruction sets than the x86 chips commonly found in PCs.
These chips can also speed up applications.
Google says Axion’s performance is 30% higher than the fastest Arm-based general-purpose virtual machines in the cloud, and 50% higher than similar x86-based virtual machines.
“I think it competes with their product portfolio,” Decatur said.
watch: Patient Capital’s Christina Malbon makes bull case for Google, Canada Goose