This month, Ampere Computing and the Cloud Native Computing Foundation (CNCF) are launching a pilot project to provide managed CI from Actuated for GitHub Actions to a number of CNCF projects. We are doing this for a few reasons – first as a Cloud Native Processor company, we want to support Cloud Native projects in any way we can. Second, we want to ensure that Arm64 is supported as a Tier 1 platform by as many CNCF projects as possible. And finally, we are interested in reinforcing the status of Ampere, and Arm64 servers in general, as a serious option for all your Cloud Native computing needs.
This initiative, from our perspective, grew out of a conversation we had a couple of months ago with Ed Vielmetti of Equinix. He shared that the Equinix Open Source Partner Program makes bare metal Ampere instances available to some open source projects, but that it is hard for many of these projects to take advantage of the benefits that Ampere Altra brings to the table – primarily, lots of cores – when building their projects. That is because many CNCF projects are hosted on GitHub and use GitHub Actions for CI to build and test their code, which makes it hard for them to efficiently utilize a powerful bare metal instance. While GitHub offers hosted and managed runners for x86, it does not yet offer this for Arm64. GitHub does enable projects to self-host runners and manage them, we have written a tutorial on this in the past, however, this does pose some problems for open source projects. First among these is that GitHub themselves do not recommend using self-hosted runners for public projects, as other people can then fork the repository and run arbitrary unauthorized code in your runner.
This sounded like a problem that we could help solve. We had previously spoken to Ampere Developer community member and well-known Arm enthusiast Alex Ellis of OpenFaaS about a solution that his company had been working on called actuated, which aims to solve these issues and more. Actuated spins up short-lived, immutable virtual machines using Firecracker virtual machines (VMs). It can efficiently carve up large bare metal hosts, and then manage all the interactions required to turn them into GitHub Action runners. In addition to solving the security problem, this solution offers other benefits that Alex covers in his blog post on the project- not least of which is that it is lightning fast. And Alex shared that he had been talking to our Linux Foundation friends about running a pilot program for several projects.
Once we set this plan in motion, it all came together quite quickly. Chris Aniszczyck committed some resources from the CNCF for the pilot project, and Ampere Computing funded the balance, to enable an initial pilot group of CNCF projects to take advantage of actuated. We have already onboarded our first two projects, etcd and fluent bit, and are actively seeking others. Our friends at Calyptia have also written about the benefits they have seen from running on Actuated.
This initiative offers benefits all around. For Equinix, this has meant that the servers they are donating for use by the CNCF are more efficiently used and enables them to support more projects overall. For CNCF projects, they will see build and CI run times on Arm64 come way down, and they will get that benefit with almost no work on their part.
That is because Actuated is faster even than GitHub hosted x86 runners, and using the Actuated service is almost identical to using the GitHub runners – you simply add the Actuated target to your “runs-on” line. For the CNCF, this represents a great additional option to their member projects for Arm64 CI. And for Ampere Computing, we get to see an improvement in the standard of support of 64-bit Arm for some CNCF projects who have held off on running complete CI tests on Arm64 because of runtime constraints. For other projects, we get the benefit of showing just what Ampere Altra can do, and we have the pleasure of reducing build times and increasing the community efficiency of some Cloud Native computing projects that make up the heart of the Kubernetes ecosystem.
Next week at KubeCon in Chicago, the Ampere Developer team will be on-site, and eager to talk to any project interested in taking advantage of this pilot. Our expectation is that we will soon be able to report on happy project maintainers reporting satisfaction with the service, and our hope is that the initiative will expand beyond this initial pilot to be available to any CNCF projects that want to take advantage of it.
Built for sustainable cloud computing, Ampere’s Cloud Native Processors deliver predictable high performance, platform scalability, and power efficiency unprecedented in the industry.
Talk to our expert sales team about partnerships or get more information or trial access to Ampere Systems through our Developer Programs, sign-up for our Developer Newsletter, or join the conversation at Ampere’s Developer Community.