AARCH64, sometimes also referred to as ARM64, is a CPU architecture developed by ARM Ltd., and a 64-bit extension of the pre-existing ARM architecture, starting from ARMv8-A. ARM architectures are primarily known for their energy efficiency and low power consumption. For that reason, virtually all mobile phones and tablets today use ARM architecture-based CPUs.
With the increasing need to decrease power consumption on servers, data centers and desktop computers worldwide, there is a slow but steady shift towards ARM-based systems altogether.
Although AARCH64 and x64 (Intel, AMD, …) are both 64-bit CPU architectures, their inner basics are vastly different. Programs compiled for one platform, won’t work on the other (except with some magic), and vice-versa. That means, software does not only need to be recompiled, but often requires extensive optimization for either platform.
Today, 64-bit ARM chips are reasonably widespread with mobile devices, such as tablets or mobile phones. Same is true for set-top boxes, single board computers (SBCs), like the Raspberry Pi, and many other devices of our daily life. On the other hand, desktop computers and servers running ARM CPUs are still rare, though, the breakthrough may be around the corner now. With companies like Facebook testing massive AARCH64 based server deployments and Apple moving to their own Apple Silicon branded M1 processors, the low power consumption and computational efficiency is the top argument. We are on the edge to see a large shift in deployments, away from power hungry x64 based servers towards low-energy, large core-count ARM64 systems.
That said, if you already use one of the new Apple machines, there is a good chance you already work on an AARCH64 architecture on a day-by-day basis. Conversely, that means you should be able to use software built directly for that platform. While Apple provides some transparency layer to enable a lot of the old software to work on newer Macs, the performance penalty is visible.
It also means, if you use Google or Facebook on a daily basis, you also already have a good chance working on an ARM system, even though it just happens to happen in the data centers. Which, however, brings us to the important point, data centers and servers.
If your system runs in a data center or you run your own operational platform, earlier or later you’ll be confronted with the question to use ARM instead of x64-based CPUs. Either for cost efficiency or power consumption reasons. If you host on AWS, you already have the chance to select ARM-based EC instances. Others either already provide similar options or will follow soon.
Migrating systems to ARM is twofold though, with your own services only being half of the work. While migrating your own applications and services is certainly the most important element, being able to run all the supporting, third-party, services on the new environment is important, too. These services include services such as
- Resource managers, such as Kubernetes, Cloud Foundry, Openshift
- Networking layers or tools, such as Traefik, nginx, Istio, Ambassador
- Potentially databases, such as Postgresql, MySQL, or others
- Programming languages, either based on runtime environments like Java, or compiling natively for AARCH64 like go(lang) or C/C++
- And last but not least, Observability tools and system or application monitoring
The last reason is especially important, if you run a highly dynamic, microservices-based setup of independent applications. In this case automation is an integral part of the always up-to-date health and performance state of your system.
At Instana we love to be at the edge of technology. Being implemented mainly in Java, and thanks to the Java teams at companies like Oracle and Microsoft, we are now fully enabled on any ARM64 platform. Furthermore, Instana provides a ready-to-run Docker image straight from Docker Hub and the Agent service, as well as sensors are fully ARM64 compatible. It just works as you’d expect.
That provides you with all the convenience, automatic discovery, and on-the-fly instrumentation of services you know and love.
If you have a Macbook, Mac Mini with M1 chip, or a server running on an AARCH64 chipset, we’d love to see you try it out and let us know about your experience. Not only if you love Instana’s Platform, but also if you find features that don’t work.
If you do not have an Instana account yet, but you’d love to try out Instana in your environment, just sign up for the free Instana Trial, download your Agent and get started right away to get your full-stack visibility in less than 5 minutes.