Arm a trend in the cloud now with the launch of Google Cloud • The Register


comment It’s been a difficult year for Arm. First, the British chip designer lost a financial boost that was killed by regulators with its sale to Nvidia. Then Arm laid off employees when it made plans for an IPO, and now market conditions don’t look good for that IPO.

The good news for Arm is that the cloud world is increasingly adapting to the alternative instruction set architecture. The latest sign: Google Cloud’s launch of its first Arm-based cloud instance on Wednesday, which the cloud service provider says will “offer exceptional single-threaded performance at a compelling price point.”

Google Cloud’s Tau T2A virtual machines are intended for “scalable, cloud-native workloads” and are powered by Arm-based Altra CPUs from Ampere Computing. That means the ARM-compatible VMs, now in preview in the US and Europe, are expected to offer strong performance-to-cost for things like web servers, containerized microservices, media transcoding, and rich Java applications.

Google Cloud seems pretty excited about what Arm can bring to the cloud world, as it plans to let customers and partners try the T2A VMs for free during a trial period to “speed up development.” Even if T2A becomes generally available later this year, Google Cloud said it will “continue to offer a generous trial program, offering up to 8 vCPUs and 32GB of RAM for free.”

The T2A instance is part of Google Cloud’s Tau VM family, which debuted last year with instances running on AMD’s third-generation Epyc Milan CPUs. The ARM-based instance type supports up to 48 virtual CPUs per VM and 4 GB of memory per vCPU, and network bandwidth can be up to 32 Gbps. It also comes with a “wide range of network-attached storage options.”

However, there are some limitations to T2A that also exist for the AMD-based T2D instance: no support for extended storage, sole tenancy, nested virtualization, or custom VM shapes.

While Google Cloud didn’t provide performance comparisons with x86-based instances, Ampere jumped up and said a T2A instance with 32 of its vCPUs is up to 31 percent faster than Google’s N2 instance, which uses Intel’s Ice Lake silicon with the same number used by vCPUs. This is based on an estimated score for the standard benchmark SPEC CPU 2017 Integer Rate.

Using the cloud provider’s VM pricing guide, Ampere said a T2A instance offers up to 65 percent better price/performance than the Intel-based N2 instance for on-demand pricing.

What about software support?

With the cloud world largely rooted in x86 chips most of the time, it’s right to wonder how Ampere’s Arm-based Altra CPUs can handle a wide range of software.

To that end, Ampere is doing its best to give people confidence that its processors are up to various cloud tasks. In a Wednesday blog post, the company noted how “the Arm-based server ecosystem has matured rapidly over the past few years with open-source cloud-native software stacks that have been extensively tested and deployed on Ampere Altra-based servers “.


TrendForce: AWS aims to secure Arm up to 22% of data center servers by 2025


“For example, Ampere runs over 135 popular applications on 5 different cloud-native infrastructures to ensure our customers across the market have confidence in the Ampere software environment,” wrote Jeff Wittich, Ampere’s Chief Product Officer.

The startup’s server chips also support multiple versions of Linux, including Ubuntu, Red Hat Enterprise Linux, and CentOS Stream.

Wittich pointed out that Ampere has a section on its website with a large list of applications, programming languages, and other types of software that have been tested on its Arm CPU cores.

Google Cloud managed to get testimonials from a few independent software developers who said porting their code to T2A was easy.

“From day one, we were pleasantly surprised by the ease of portability to ARM instances. The maturity of the T2A platform gives us the confidence to deploy these VMs in production,” said Khawaja Shams, CEO of Momento, a startup that provides serverless caching services.

T2A also received endorsement from academia, with Harvard University Research Associate Christoph Gorgulla saying the instance’s “better value for money” helped his team “screen more compounds and therefore discover more promising drug candidates.”

Several major cloud providers are now on the arm train

With the recent introduction of Arm-based cloud instances, the British chip designer’s ISA is now supported by six of the world’s largest cloud service providers: Amazon Web Services, Microsoft Azure, Google Cloud, Alibaba Cloud, Tencent Cloud and Oracle Cloud. Other cloud providers are also behind Arm, such as JD Cloud, UCloud and Equinix Metal.

All of this means it’s safe to say that cloud provider adoption of Arm is definitely a trend now.

A development that would have been unthinkable for some people a decade ago, like GitHub engineer Jaana Dogan post it on twitter.

Integrating ARM chips into server environments running operating systems like Linux has required a great deal of collaboration between the software and hardware worlds, mainly to agree on and stick to a standard base of features and expectations in those computers. This has made building and running software on ARM systems, especially server boxes, relatively tedious: it should just work the way x86 just works, and it seems to do it.

AWS also helped pave the way for Arm’s rise in the cloud with its decision to develop an Arm-based server CPU in-house, leveraging talent gained through the acquisition of chip designer Annapurna Labs won by Amazon in 2015. The cloud giant is now using the third generation of its Graviton chip, which is now available in Elastic Compute Cloud instances and for which it continues to make great price-performance claims compared to x86 chips.

However, looking at all the other major cloud providers rolling out Arm-based instances, as well as some of the smaller ones, there is a common element that ties them together: Ampere computing.

Founded by former Intel executive Renee James, the Silicon Valley startup recently said growing support for its Altra processors from a variety of companies and cloud providers shows the chips are better suited for cloud applications than those from Intel or AMD.

Like Arm, Ampere is planning an IPO at some point, anticipating that market conditions will eventually improve. If you’re curious about how Ampere’s chip designs are well-suited for cloud applications, we encourage you to read our recent interview with Jeff Wittich, Ampere’s CEO.

While ARM’s growing acceptance in the cloud world is a welcome sign for anyone tired of Intel’s dominance in space, the question now is how long ARM and silicon partners like Ampere and AWS can sustain this momentum. After all, Intel and AMD both have plans to introduce specialized cloud chips in the near future, and who knows, maybe RISC-V can shake things up even more. ®


Comments are closed.