Cloud Native ComputingDevelopersDevOpsFeaturedLet's TalkOpen SourceVideo

How Equinix Metal Helps Partners Go Cloud Native With Bare Metal


Mark Coleman (LinkedIn) | Equinix Metal (Twitter)
Sirish Raghuram (LinkedIn, Twitter) | Platform9 (Twitter)
Mohan Atreya (LinkedIn, Twitter) | Rafay (Twitter)
Narayan Sainaney (LinkedIn, Twitter) | CodeZero (Twitter)

Equinix Metal is building an open ecosystem around Kubernetes, working with a large partner community, instead of creating its own managed Kubernetes services. The company recently announced new partners who help their customers run their workloads on Equinix Metal. We hosted a panel with some of Equinix Metal partners to understand how they are leveraging Equinix Metal to help their customers in their cloud-native journey.

Topics we covered in this show:

  • Why did Equinix Metal choose to build an open ecosystem instead of building managed Kubernetes services?
  • Introduction to CodeZero, what do they do?
  • Platform9’s perspective on Kubernetes and bare metal
  • Rafay shares their vision for Kubernetes Operations
  • Mark talks about the importance of such partners for Equinix Metal
  • These partners are solving different problems for different users. What does it mean for Equinix Metal?
  • How is CodeZero using Equinix Metal to solve the ‘four’ problems Atreya mentions, including life cycle management?
  • Raghuram talks about cloud-native experience on bare metal and how they leverage Equinix Metal to offer that experience to their customers.
  • Sainaney talks about what defines the next generation of developers and how they leverage Equinix Metal to help the next generation of developers.
  • How is Equinix Metal helping the growth and adoption of Kubernetes?
  • Raghuram mentions some of the most exciting use cases of bare metal such as edge and 5G. He discusses how Equinix Metal helps with such use-cases.
  • Coleman shares his thoughts and insights into these new use cases.
  • Atreya, Sainaney and Raghuram  also share how they use Equinix Metal to help their users.


Swapnil Bhartiya: Hi, this is your Swapnil Bhartiya, and welcome our TFiR less stock panel. Equinix metal is focused on building an open partner ecosystem around Kubernetes instead of kind of building their own managed Kubernetes services. And today we have a great panel with some esteemed guests from the Equinix partner ecosystem we have with us once again Sirish Raghuram CEO of platform nine Mo TRIA, SVP of product and solutions at Rafe NA CTO of code zero. And once again, Mark Coleman, senior director of developer relations at Equinix metal. Welcome to the show. Mark, We had a great discussion last week, where we talked about Equinix metals, focus on an open ecosystem around Kubernetes. I want to reiterate that vision once again. Why is Equinix working on building a product ecosystem instead of your own managed Kubernetes services?

Mark Coleman: You Know, Essen, what it comes down to is we think that other people are better at this than we are, and as we’re moving this hybrid and now things like interconnection, which we’re making this much more of a horizontalized market and because it’s a horizontalized market, it makes more sense for us to stay in our lane and then work with the incredible smart people at all of our partners. So of whom are here today, where they’re focusing on what they’re best at too. And the combination of those we think is likely going to be better, not only for supporting companies in our ecosystem and for future innovation, but also of course, for our customers.

Swapnil Bhartiya: We have hosted both Rafay and Platform9 before, but this is the first time we have good, fresh on our show. So I would love to learn a bit it about the company. So Narayan, and can you tell us, what do you folks do?

Narayan Sainaney: What we’re really focused on is making developers life a lot easier. Kubernetes has a pretty steep learning curve. And when we look at setting up Kubernetes, everything from set up to development to deploying your payloads and then going all the way through to even operating Kubernetes, there’s a fair bit, that organizations have to learn. So we see that there’s a timing for this next layer tier of tooling that is going to help cross chasm reach the next group of developers that are not infrastructure experts, who are not experts at things like load balancers and so on and so forth, getting right into being able to write software for Kubernetes. So, that’s where we are here to help.

Swapnil Bhartiya: Sirish i have been covering platform lines since early Open tag days. So bare metal has always been a topic of discussion between us, but let’s talk about bare metal and Kubernetes equation.
Sirish Raghuram: I think, look, here’s something really interesting that happens with bare metal and communities. For a long time, people have looked for more and alternatives to the cloud, right? More freedom and more flexibility with the cloud. And what happens with bare metal and communities is that if you think of is communities clusters, are they running in the cloud or is the cloud running in a community’s cluster? And I think if you run, think about running like communities in a cluster in on bare metal, one of the things we’ve been thinking about for a long time is how do we deliver an experience where people feel like they have a complete cloud experience, but in the cluster.

Sirish Raghuram: So not clusters in the cloud alone. Yes, you can run communities clusters in the cloud, but how do we all also run, make, deliver an experience where you can get a full cloud experience from a cluster, which happens to be running on bare metal, which has lots of use cases, right? Like you can get better price performance, you can locate that infrastructure specific regions. You can get better economics out of it. Right? So we think that this is incredibly useful and we have some joint customers with Equinix metal. We ourselves were a very early customer of Equinix metal when they before they were acquired by Equinix. So there’s a lot of use cases for this, and we were excited to be a part of this ecosystem.

Swapnil Bhartiya: You folks just raised around investment. Congratulations, let’s use this opportunity to share with our viewers, what do you folks do? And what is your vision for Kubernetes operations?
Mohan Atreya: So we looked at this slightly differently, leads, hope based on what we are hearing from customers. So we try and solve the problem of streamlining the life cycle management of both the clusters, as well as the applications for companies. Why is this a big deal?
Mohan Atreya:So it’s a big deal because in today’s world organizations, don’t just have a handful of clusters. They like to get closer to users. They like to deliver applications with great performance to do this. An average organization would have clusters on premises on things like new services, like Equinix metal, which help them bring their services closer to users, could be on the public cloud and hundreds of edge locations managing all, this can be a nightmare operationally for people. We try and help streamline the whole thing for them. And we focus on specifically four problems, automation, how do I get all of this up and running quickly and efficiently and keep the life cycle going visibility. When I have both clusters and applications in all these various places, how do I know what’s going on? Right visibility, then you can’t do all of these and fly by the seat of your pants. You need to have the security and governance, right? we also make sure that these four items are taken care of at scale, all through a single pane of glass. That’s kind of our current mission. Hopefully that clarifies here.

Swapnil Bhartiya:Mark, what an excellent panel behavior. If you look at these three players here, they are solving different problems. They have different stories, but there is a common theme as well. So from an Equinix metal perspective, how do you look at these partners? They’re trying to solve different problems for different users, but they are relying on Equinix metal. What does it mean to you?

Mark Coleman: Yeah, absolutely. So, I mean, you’re right. There’s many more partners, of course just came on in this recent trust release and they do all have different characteristic. You know, some of them are very specifically good at one thing and they need to be assembled with other partners to be able to be useful. Some of them, like Sirish was talking about a little bit earlier, they really kind of bring, it’s almost like they’re in our ecosystem. And then they bring their own ecosystem with them and it takes a different type of enablement. So one of the things we’ve done outside of interfacing with all these companies and bring them into the ecosystem, we’ve also built out an entire team focused only on this led by the very talented Josh trans there’s five people in there, MASU, Nori, James Levison share on black man, Eric Rosa and Vil me.

And they’re trying to figure that out right now because there’s kind of two sides to it. Like different partners have different levels of enablement, but what it needs to come down to at a certain point is saying voila here’s 25 partners doesn’t necessarily help the customer that much because it just means now they have to go and understand all these partners. So a lot of the work we are doing is trying to be like a concierge for the partner options that we have. Like, it sounds like based on your needs, you might combine the partners in this way. So there’s a bit of a sort of pre-sale thing in there that we’re going to be leaning into to make sure that basically everybody gets the maximum value. I should say, out of being part of the ecosystem.

Swapnil Bhartiya:Mark when I look at the use cases and these players here, I cannot stop comparing this to Linux external where different users, different players, we’re bringing solutions to their own problems, but they made Linux betters for everyone else. So how are these use cases? These players, these partners are while trying to solve their own problems are also making Equinix better for everyone else. So how do you look at this?

Mark Coleman: I think the theme that’s tying all of this together, like there’s many different ways to look at it, but I think the one that’s tying all of this together is still the horizontalization if the cloud model that I’ve been talking about for a while, you know, for the longest time, if you are over at Amazon, they have these incredible products and everything is managed. It’s very much like having an IBM mainframe in, in 1988, right? Like everything is verticalized. And now what we’re starting to see as you pointed out, so you could be running Intel chips in there, you might be running AMD chips, and then you move up a little bit, you’ve got a whole host of different operating system options. You move up a bit more, we’ve just interview 15 different ways to do Kubernetes.

And of course, innumerable permutations of how you could combine those. And on top of that, you’ve got different support models. So I think really it’s actually already a problem we’re aware of. It’s a solved problem, which is how do you make customer successful on a horizontalized market and the way you do it, it’s by consulting them a lot. I think it’s just like back in the days as we moved into IBM PC, suddenly, I’m like, who’s selling me my chip. Who’s selling me my memory. Who’s selling my own mother bought who selling me my case, my power, my screen, like somehow helping our customers to negotiate. Our complexity is definitely a challenge here. But I still think, even though it’s challenging, it’s the right way to go.

Swapnil Bhartiya: Mohan, you’re talking about life sack management, and you also mentioned four problems. Can you share, how does Equinix metal help you solve these four problems? And also of course, let’s talk about life cycle management as well.

Mohan Atreya: One of the common problems we keep hearing from organizations, I kind of touch on what mark was alluding to. They’re very quickly divesting their on-premises data center resources and they need a viable alternative. And many of the applications are not yet well suited for an extremely virtualized environment, etc. So they need to run this on a bare metal, kind of a service. Now, the economics of them having to build out a large data center and not just one, in many cases, they need global companies, they need to service the users in 20 different locations. And not only that, they also need amazing interconnects across these data centers because data is being back hauled, back and forth, etc. So it’s a build versus buy decision for these customers. And the economics are obvious that they should be looking at something like Equinix metal, which is kind of why we excited to partner with Equinix level because we see organizations making that shift, asking for something like that.

We think this is a fantastic option in the market for at least our customers on top of it, kind of pick on an interesting extending what cities just mentioned, which actually is a fascinating use case we are seeing in the market, which is kind of turning the whole model upside down on bare metal with Kubernetes underneath the covers. You can actually applications teams can actually run both VMs on Kubernetes and containers on Kubernetes. That is a fascinating trend we are seeing in the market today. Of course, VMs on VMs are not going to be great as we all know, but running VMs on Kubernetes and containers on Kubernetes. So I now have a unified management plan for both my legacy applications and my modern applications. I mean, that’s the state, most enterprises are so, no one’s a hundred percent containers yet, unless you are a startup. So that is something we see again and again with big enterprises. So very excited to partner with Equinix.

Swapnil Bhartiya:Sirish you were talking about cloud native, like experience on bare metal. I may be totally wrong, but I think cloud native is more or less way of doing thing versus a thing in itself. So let’s talk about how does Equinix metal help you bring that cloud native experience to bare metal so that know your customers can leverage all these technologies while also leveraging bare metal.

Sirish Raghuram: Yeah. I think we talked about this in the past, right? Like that cloud itself is a philosophy cloud native is a way of architecting and operating applications and consuming infrastructure and with consumption, right? Like in a very dynamic way. What Equinix does is really interesting to us because it gives us a programmable fabric that very few other companies in the industry have actually done a really good job of right. They give you a completely, they allow us to consume compute, like physical CPU, memory storage network in a way with a programmable fabric, which we can then take in our layer and then expose up for consumption through all the richness of the Kubernetes API and the operators and the application services, everything from databases to message queues and storage volumes and others.

So we can focus on that core competency of ours. So the end result to the customers, right? Whether a retail customer, it’s a telco customer, it’s an AI startup that wants to focus on their core competency of focusing on AI analytics and delivering the next generation of retail experiences. We are able to consume this programmable fabric from Equinix and deliver compute while the richness of the Kubernetes cloud API and dry all these end use cases such as the retail use case, the 5G use case, the CDN edge use case or the AI use case. And that wouldn’t be possible if we didn’t have a programmable fabric like Equinix. And this is also why, it goes back to do users need to run Kubernetes clusters and the cloud, or with something like Equinix metal, can they essentially treat the cluster as the cloud? And it happens to be running on bare metal and gives them a lot of richness and flexibility that you may not get outside of a framework like Equinix metal.

That’s the reason I think we have joint customers who are doing this today with us and we, our teams love Equinix because it gives us a great fabric to build on top of and deliver the next layer of experience on top.

Swapnil Bhartiya:Narayan you were talking about the next generation of developers to simplify their workflow. Can you define who would you call the next generation of developers? And then also, if you can talk about how are you leveraging Equinix metal to kind of enable that.

Narayan Sainaney: What Sirish is saying. First of all, having that infrastructure is critical. And the second piece of that is what Equinix is providing is this great vanilla Kubernetes experience. It’s undoubted it doesn’t have the thing about Kubernetes is almost living on original products of Java, which was right. Ones run everywhere. And if you need to run something regardless of geolocation, regardless of jurisdiction, and you need that reliability, you’re getting that from Equinix. And now with the partner ecosystem, one of the challenges for app developers, or at least even vendors ISVs like me, if there’s a gap in a solution, you don’t have a solution. The solution comes from mixing. What Mohan has what Sirish has with what we have, and now you actually get something that you can actually run, build your applications on.

So there is this notion I believe in, if you look at the Kubernetes ecosystem there are about a 100, 1000 attendees at Cube corn maybe according to the CNCF, a little over a million developers it’s growing rapidly. I believe the last report was 47%, but how’s it going to reach that developer who’s writing no JS applications as a Python developer cannot learn the cloud stack, the Kubernetes stack load balancers Kubernetes and its ecosystems is like a collection of Lego blocks, but each Lego blocks has its own 50 page manual that you want to read. And that gets very daunting. And as an app developer at as a microservice level and so on, you’re already having to build expertise in your domain, which is the application stack or if you’re front end and so on and so forth.

So, Kubernetes is going to be used by pretty much everybody at some point because it is so ubiquitous and it’s capabilities, and we need to help bridge that gap to that. That’s where the growth is going to come. So that’s where, or we’re coming in and with Equnix and the partner ecosystem helping reach that broader group and as I think Mohan alluded, the cloud almost becomes like a utility in the sense that it just needs to work. It’s pervasive and you get to experience it as an app developer.

Swapnil Bhartiya: Mark after listening to all these stories, of course, Kubernetes, this not need anybody’s help. It is growing at its own pace, but, I do feel that metal is playing a very critical role in adoption of communities in many use cases. So how do you see, what is your role in the Kubernetes ecosystem?

Mark Coleman: Yeah, so there’s a couple of things we do there. I mean, obviously, we’ve got this shoe each focus on Kubernetes container management partners, and that’s for a pretty simple reason as was just alluded to like most new products are going to end up on Kubernetes. Most re-engineered products are going to be at least looking at Kubernetes. So we think it makes a lot of sense for us. You know, most people, I can’t be very clear on the statistics here without doing bit of research, but like, I think a lot of people come to Equinix metal and they want to run either VMs on top of the metal or they want to run containers on top of metal. That’s it. Now, occasionally they’re also just running things directly on bare metal. So enabling those two and then maybe starting to look forward as well into some of the exciting stuff that’s happening, high conversion infrastructure again means that we’ve got that strong foundation.

Now, one of the other things we do is we have an open source program run by ed bill Medi. And I think, I can’t remember the exact number, but it’s around about a million dollars a year of like free hardware goes to the CNCF. And of course this is a win for both of us because most of the CNCF projects build natively on our machines because that’s where they’re building them. But also it means that they don’t have to have that, that infrastructure burden. And often if you’re just doing builds all day, every time a poor request comes in having a really powerful bare metal machine is going to be a lot faster than using something virtual was. So there’s that. And then, I don’t know, it seems like so much we do is around cloud basis, but they’re the two real specific programs that I’m thinking of right now, apart from just our general involvement with the CNCF.

Swapnil Bhartiya: Sirish I’ll come back to you. You talked about 5G, you talked about edge, you talked about a lot other use cases. If I ask you, are there certain use cases is where you felt that, Hey, we could not have done this without Equinix metal.

Sirish Raghuram: I mean, there’s some customers that we have jointly the retail analytics, AI startup comes to mind where their business model essentially I think, is not viable without a solution like platform lines, community stack on top of Equinix metal. So let me explain why, right? So they were running on Amazon. They’re a small startup and you know, AI in Amazon, AI is very GPU, intensive, GPU resources are expensive in Amazon. They’re also hard to find the supply issues, capacity mismatches and so forth. So they had to go kind of find a more efficient way to kind of consume GPU. So they had to kind of go and find like a place where they could get access to GPS more efficiently, and a place like metal is a perfect candidate for a use case like that. But then they’re a small startup and they want to focus their team and their precious time on engineering, on their core competencies, which is AI analytics for the retail industry, applied AI and running Kubernetes.

I mean, their whole team could just be spending time trying to run Kubernetes, right? Like they’re not in the Kubernetes business, we’re in the Kubernetes business. We make that super simple and easy. So it’s very remarkable to me that their founder and CEO said to set to our team that look essentially, if it wasn’t this option to run like at this cloud Kubernetes cloud experience, but on bare metal, we wouldn’t have a business. Like we wouldn’t exist today because we would’ve basically shut shop because we are upside down in Amazon. We couldn’t make the economics work. I think that’s one example. I think, if you talk more broadly about bare metal, I would just go look at the whole 5G ecosystem and say, there’s a massive, there’s an essentially millions of nodes that are going to come online and all running communities because that entire industry and the supply chain is transforming to containers.

It’s actually a very large industry outside of the generic, broader horizontal enterprise use cases that a lot of which actually happens in the public cloud. I think that the 5G industry in particular is very, very significant because of the sheer scale and size that they have. And I think almost a 100% of that is going to be running on bare metal. Like it makes no sense to be paying the virtualization overhead, especially for demanding like network function, containerized network function, workloads. They have like latency requirements, they have performance requirements, they have throughput requirements. There’s CapEx. Like the more hardware you need to deploy on site, the more CapEx here to deal with the more OPEX here to deal with. So the economics of bare metal, I think absolutely shine in a use case like that. One obviously of our platform that is doing a lot of work and we are going to continue investing in a great product experience there.

Swapnil Bhartiya:Mark, how do you see these use cases, especially edge? We talked about it earlier so many times in so many discussions. And here, when I talk about edge, we are not talking about a small IOT devices. We are talking about edge data centers, which maybe just resource constrained devices or resource constrains data centers and badge. So from Equinix metals perspective, how do you look at these use cases?

Mark Coleman: Yeah, I mean it is very important to us, right? So we have the Metro concept and you find that a lot of our data sensors are basically within major metros. Another thing that’s been done by broader Equinix is I believe the news went out on Friday about a new specifically focused 5G testing facility in the states. So this is something that we’re investing in a lot. And I think, you can sort of touch it from a metal point of view, but you can also touch on a fabric point of view and how we bring those together. I think like series set is still emerging a little bit as we start to see the use cases come out. But I know that one of the things that we’ve been very busy with is the Lennox foundation’s edge foundation. I think it’s called edge foundation.

We created something called the state of the edge report in 2017 at Mount down, donated that to that foundation. And so that’s something that we are using to stay very close this industry. And as the use cases start to emerge, I think we’re going to be, we’re seeing some changes in how we architect our tool and to support that.

Swapnil Bhartiya: Mohan can you share some of the use cases where you see that it was not possible to do without Equinix metals partnership.

Mohan Atreya: Perhaps maybe a good way to describe this is maybe take a real life customer use case, which is a recent one. So the particular problem, the challenge they had was they had to start moving fast and gone on the days when you can wait for a year to build something out and plan for two years to launch an application. People need to get these done in weeks or days these days. And well, we have a customer who, who wanted to achieve that. Of course, that application is optimized for bare metal environments. And the key tenant for them was programmable, right? So they cannot wait for bare metal equipment to be bought. And you guys all know we are living in COVID times, supply chain is all messed up, right? By the time they can order hardware in the data centers and all of that, they’re going to be dead in the water.

So a programmable way to consume infrastructure and not just in some weird data center, in a place where their users or not there, they wanted the flexibility to move around. Right? So something like Equinix metal was extremely well suited for them because they could just literally consume this via APIs. Right? All they had to do was tweak their pipelines, deployment pipelines and infrastructure would come up in the right location. But it’s not just the infrastructure for these Kubernetes clusters on bare metal that come up. They also need the back hauls. The back hauls are really important because applications need data. It’s so hard if I have a data center, but I can’t get my data there, then there’s no use. So the, the combination of having a programmable consumable environment, closer to users with a fantastic back haul, because the networking at scale, the interconnects are really important. That combination is what is compelling. Many of these enterprises to kind of reconsider, like how do they augment their existing data centers, or maybe they even use public clouds today.

But this is surrounded by these services that they run all over. Why they want to deliver great experiences for the users, right? Whether it’s a consumer or an enterprise, you come to expect world class experiences, a consumerization, like the experience you get through a Gmail or like instant access to stuff. That’s kind what we see happening.

Swapnil Bhartiya: Excellent thanks Mohan. Narayan Same question for you from code zero’s perspective, can you share some use cases where Equinix metal played a Critical role there?
Narayan Sainaney: Definitely. And got the same thing. I believe the expectation now as an application developer is how dynamic you need to environment to be. So the fact that Equinix metal is fully programed both through their APIs. We were able to programmatically set up clusters in various jurisdictions on demand where needed and to be able to shift the loads around. And the level of dynamism is not just. I need a cluster here, I need a cluster there. And then with code zero, your developer workstation. But what we enable is you can literally attach your developer workstation to those clusters for inflight development. So some people liken what we do to be able to fix a jet engine mid-flight. And when you get to that level of expectation as an application developer, if you’re having an issue in production or having an issue that you are not able to see, because look, it’s working on my computer to not working in the production environment.

I need an exact replica, and I need it really quickly. Like Facebook just had a outage on Monday. It’s typically has to do with the fact that one environment doesn’t match another environment. So if you can actually through programmatic means be able to instantly replicate that and then try to get to what conditions are happening in production that you’re not having locally. You need your entire infrastructure to be able to be brought up without any manual intervention and so on and so forth. So we’re seeing a lot of those type of use cases and enabling that through our tooling. And that’s where it works really well for us.

Swapnil Bhartiya: Narayan, Mohan, Sirish and of course Mark, thank you so much for taking time out today and creating an excellent panel. I love the discussion and of course I would love to have you folks back on the show. Thank you.