The introduction of the Arm Neoverse family of CPUs opened the door for our partners to deliver the diverse set of high-performance, secure, and scalable solutions required for modern cloud infrastructure. Since then, the Arm ecosystem has grown to include all major cloud providers across the globe.

Simultaneously, the Arm global software ecosystem has been expanding to include a wide range of applications and workloads. Hundreds of popular open source-based projects hosted by Cloud Native Computing Foundation and Open Infra Foundation, as well as majority of independent software vendors (ISVs) have embraced Arm Neoverse-based platforms with tier 1 support. Linux operating system vendors such as Red Hat and SUSE, cloud native ISVs, such as Cloud Bees and New Relic, and traditional enterprise software industry luminaries, such as Oracle and SAP, to name a few, have committed to support 64-bit Arm systems on their long-term product roadmaps.

These efforts by our partner ecosystem have made Arm platform benefits of better price-performance and energy efficiency truly accessible to the broad swathe of developers further ensuring software availability and portability across all Neoverse-powered products.

As a result, customers are making their choices by picking Arm-based cloud instances, partners continue to invest in the future generations of Arm silicon to sustain customer demand and ISVs and open-source projects are busily porting their stacks to 64-bit Arm architecture.

In no small part this adoption trend is fueled by how closely Arm interacts with our ecosystem partners.

Expanding cloud footprint with Arm

Just over a year ago our cloud provider partner, Google, expanded Arm footprint in the cloud by launching Tau T2A VMs on Google Cloud Platform (GCP). These VMs, powered by Ampere® Altra® Arm Neoverse-based processors feature up to 48 vCPUs per VM and up to 4GB of memory per vCPU.

This innovative offering came as culmination of multi-year collaboration among Ampere, Arm and Google and is great example of how Arm Neoverse-based CPUs can support complex demands of the modern cloud-native applications in the most efficient and sustainable way. Combined with support for popular Linux operating systems, as well as seamless integration with container-optimized OS for efficient Docker container deployments, these VMs offer a compelling solution for diverse scale-out workloads that encompass an array of applications including web servers, containerized microservices, data-logging processing, media transcoding, and large-scale Java applications.

Additionally, T2A VMs deliver exceptional single-threaded performance at a compelling price and are integrated with a number of GCP services, including Google Kubernetes Engine (GKE), Dataflow, Cloud Batch, and Dataproc, allowing developers to stand up their applications quickly, efficiently, and securely.

Collaborating in open-source and on standards

Over the years, Google and Arm have collaborated on a number of open-source projects and industry initiatives. We have been working together on such prominent projects as Tensorflow and Kubernetes, and, most recently, joined forces in forming OpenXLA project that helps accelerate AI innovation.

To further aid customers on their journey to Arm-based platforms, we’ve introduced the Arm SystemReady Virtual Environment (VE) certification that addresses a growing need towards standardization of cloud infrastructure to seamlessly run workloads in multi-cloud settings. This standard ensures that Arm-based virtual environments, such as cloud instances, work across common operating systems and software, further solidifying the Arm ecosystem's foothold in the cloud landscape. Notably, Google Cloud Tau T2A VMs was among the first to achieve this certification, exemplifying a commitment to providing a unified and standardized platform for seamless cloud operations.

Google Cloud is also a part of the Works on Arm initiative, that enables software developers to build, test and optimize projects on the Arm architecture at no cost during a trial period by using hardware in the cloud.

Another pivotal milestone in enabling developer’s journey to Arm platforms was the creation of Learning Paths by Arm, a new way of providing a long form, written how-to content for software developers guiding them through selecting Arm hardware for their computing needs. For example, one of the common scenarios for deploying cloud native applications is to stand up a Kubernetes cluster. Arm developed a Learning Path on how to build GKE cluster with T2A VMs. There are several additional Learning Paths available for developing software on Arm or for migrating existing applications to Arm using GCP.

Running scale-out workloads on Arm

Built upon Arm's unmatched performance efficiency, Tau T2A VMs are becoming a popular choice for various enterprises running compute-intensive applications, including Arm as a company.

We are working on migrating our compute infrastructure to Arm architecture and GCP offers a compelling way to provision Arm-based VMs to our internal users that run electronic design automation (EDA) workloads, where we simulate our chip designs ahead of shipping them to our partners.

We started this process with our on-premises datacenters several years ago working closely with EDA ISVs, namely Ansys, Cadence, Siemens and Synopsys, as well as other software ecosystem partners, to ensure that their products run smoothly on Arm hardware. We then began to gradually include cloud-based computational resources. This year we’ve added Tau T2A instances from GCP to our compute infrastructure, providing us with a true hybrid cloud environment to run a wide variety of simulation jobs using internally developed tools as well as software suites from our partners.

The two scenarios where we use Tau T2A instances from GCP for EDA include:

• Deploying compute clusters using GCP infrastructure-as-a-service with the same architecture as our on-premises datacenters for seamless capacity increase.
• Submitting simulation jobs using Rescale that provides agility and choice across multiple cloud providers in real-time, based on the use case and tooling required.

This example underscores the versatility of Arm architecture and highlights Tau T2A VMs capabilities to run complex scale-out workloads.

Arm at Google Cloud Next 2023

We are excited to showcase these and other solutions at the Google Cloud Next ‘23 conference. If you are attending, please visit our booth #1841 to check out the advantages of running workloads on Ampere Arm-based T2A instances, have a conversation with our experts on-site and sign up for our developer program that offers even greater access to Arm programs and resources.

In the booth we will have a demonstration of a cloud-native application deployment in a hybrid GKE cluster with x86 and Arm-based nodes. We will show how easy it is to take an existing x86-based web application running in the Kubernetes (GKE) cluster and turn it into a multi-arch application with minimal changes.

If you are a developer looking to get started with Arm-based Tau T2A VMs, Google is offering a free trial from now until March 31, 2024. We also encourage you to check out Works on Arm program that provides free access to Arm-based partner hardware as well as our recently expanded set of resources available through the Arm Developer program.

For more information on Arm software ecosystem, please visit Arm Developer Hub where you’ll find documentation, on-demand tutorials, Learning Paths for cloud computing, and a community of developers working on Arm.

Build a Multi-Arch GKE Cluster