Arm Logo
Industry: Technology
Location: United Kingdom

Arm deploys Rancher Prime stack to 2,500 engineers for companywide DevOps productivity

Highlights

  • Following successful proof of concept with 450 engineers, Arm deployed the Rancher stack across the company to 2,500 engineers.
  • Reduced complexity in DevOps processes through repeatable processes.
  • SUSE Support and Services teams act as trusted advisers and proactive problem solvers.

Products

At-a-Glance

In June 2023, SUSE’s general manager of enterprise container management, Peter Smails, spoke with Andrew Wafaa, senior director of software communities and fellow at Arm. They discussed how Arm, a leading technology provider of processor IP, is using K3s, Rancher Prime and other SUSE solutions to streamline its organizationwide DevOps processes. 

 

Watch the interview

“With SUSE’s professional services team and the support team, we have confidence that they understand what we're doing; why we're trying to do it; what we're trying to achieve. So, they can actually be proactive.” 

Interview transcript

Everyone's cloud native transformation journeys are different. 

Absolutely. 

I would love to hear your perspective from Arm, in terms of what your journey has looked like. What was the genesis? Where did it start? What were you trying to accomplish? Set the background and context. What was happening? 

The open source software team has around 450 engineers, and we have a disparate set of software stacks, hardware stacks. Our kernel engineers love to have a myriad of different platforms sitting on their desks, under their desks, whatever. And maintaining all of that is a pain, right? 

Complexity, possibly? 

Absolutely. Okay. That's an understatement, shall we say. And so, for our DevOps team, they wanted to try and wrangle things into some form of order. And so, we wanted to try and standardize: make things simple; make it repeatable; make it easier for new staff to join – they know exactly what they need to do, how to maintain things, et cetera. 

And so, we looked at trying to modernize our software stack, and we went to try it out on one of our smaller sites (just to make sure that everything worked because we still weren't 100% sure of this whole cloud native thing). Bit buzzword bingo-y kind of thing. So, we wanted to make sure it was the right thing to do. So, we deployed to a site. Originally, we started off with SUSE's CaaS product just as the pandemic hit, and we hit the supply chain issues of getting hardware.  

By the time we managed to get hardware, CaaS had been retired; Rancher had come in. We ended up pivoting, deploying K3s to the local site, and that was backed with SUSE Enterprise Storage. It seemed to work. Our engineers on site were quite happy. We thought, great, let's move forward. And so, now we're deploying the full Rancher stack. And it was going to be for the 450 engineers; it's now going to be accessible to 2,500 engineers. 

I love that. I think we refer to that as, ‘land and expand,’ a very adoption-led model. 

Absolutely. 

You mentioned that you do use quite a bit of the portfolio. So, from a technology standpoint, tell us a little bit more about what you're actually using, why, and a little bit maybe on why Rancher? 

I'll start in reverse. Why Rancher? We needed to deploy our structure on Arm architecture. So, we're coming up with all these great designs we need to ‘dog food’ effectively, right? One of the big requirements was we need a container-based solution that can run on Arm. K3s supported Arm. Great. We'd worked with Rancher pre-acquisition; we've worked with SUSE pre-acquisition, as well. We've had a longstanding relationship with SUSE team. And so, it was kind of number one on the hit list, effectively. 

We went to market; we looked at all the alternatives. At the time when we deployed K3s originally, there were only two choices. There was either SUSE/Rancher or another. When it came to deploying wider, there were a couple more alternatives; but from a cost perspective, it was quite inhibitive – that would require us to redo a lot of the architecture and such. So that's how we landed on the Rancher side of things. From a technology perspective, we've got K3s currently. We're deploying RKE2, we're deploying Harvester, NeuVector, Longhorn, SUSE Manager. I think there's some SLES in there. 

Good. There you go. Oh, by the way, quick shout out. So, we did announce NeuVector 5.2. That was another one of the things that Thomas pre-announced for us yesterday. You got to be excited about 5.2, right? 

Absolutely. 

So yeah, we're going all in. It's a collaborative engagement, so our engineering teams are working closely with the SUSE/Rancher teams. We are working on getting the NeuVector support in there. We're doing a lot of the onboarding. We've done a load of work with Harvester. And so, we've got a good relationship going with the SUSE team there, which is great. 

Perfect segue into my last question for you, which really is about the relationship. It's not just about the tech, it's about the relationship. Obviously, SUSE and Arm have a longstanding relationship. What does that customer intimacy, what does that trust, mean to you? 

Trust for us is having confidence in the colleagues that we're working with at SUSE. With the professional services team and the support team, we have confidence that they understand what we're doing, why we're trying to do it, what we're trying to achieve at the end of it. So, they can actually be proactive. They can be sympathetic to some of our problems. They can be honest with us. We're not going to cry. We can be honest with them – vice versa sort of thing. And so, it just helps everyone get to that end goal. 

Just focus on our common goals. And execute like crazy. 

Definitely. 

Brilliant.