Platform Engineers take the spotlight at KubeCon 2024 | Jay Jenkins, Akamai
At KubeCon + CloudNativeCon, Jay Jenkins, Akamai's CTO of Cloud Computing, talked about the Akamai App Platform—a flexible, pre-configured solution aimed at streamlining Kubernetes adoption.
The Akamai App Platform aims to support multi-clustered, edge-extended deployments, leveraging community contributions and partnerships.
Jenkins emphasized the importance of edge computing and distributed cloud architectures, highlighting Akamai’s long history with edge-based services. He envisioned a future where applications and data are processed closer to users, enhancing performance, reducing costs, and ensuring compliance. Akamai’s distributed cloud would allow for true redundancy and resiliency, similar to how Akamai runs its CDN and DDoS protection systems.
At KubeCon + CloudNativeCon, Jay Jenkins, CTO of Cloud Computing at Akamai, discussed the Akamai App Platform, a pre-configured and customizable solution designed to simplify Kubernetes adoption. http://www.youtube.com/watch?v=rQ5i… @cncf.io @mikemaney.bsky.social
Reflecting on his first KubeCon experience, Jenkins lauded the audience for their expertise and curiosity. “At the Akamai booth, I met developers that are working with the Kubernetes platform and are here to understand what’s happening in that area, exploring their favorite upstream projects,” said Jenkins.
“We’ve found a lot of systems integrators as well to understand the landscape and see how they can help their customers,” added Jenkins.
Jenkins also noted the increasing presence of platform engineers, describing their role as pivotal in abstracting complexity for developers by centralizing platform management. Platform engineers are “building platforms for developers to work on, which is a super exciting space for Kubernetes. It signifies a level of growth where the complexities are taken away from the developers and managed centrally by platform engineering teams,” he said.
Jenkins also talked about the growing concerns around data sovereignty and how the US and Europe are better equipped to handle the fast-changing geopolitical landscape. He discussed the challenges especially in the AI/GenAI space where balancing data input needs with privacy concerns is critical.
Jenkins talked about the role Akamai and open-source cloud-native technologies can play in enabling organizations and countries to maintain data sovereignty without having to reinvent the wheel by leveraging these open source technologies.
How does Akamai’s strategy align with the changing geopolitical situation and data sovereignty concerns?
[read more]
Unedited Transcript (Note: the text is AI generated, it has not been edited or reviewed. It may contain errors, including incorrect names. It’s provided here under Creative Commons license (CC by 4.0) to be used by bloggers, journalists and analysts for creating their own content.)
Swapnil Bhartiya: Hi. This is your host, Swapnil Bhartiya, and we are here at KubeCon and CloudNativeCon in Salt Lake City, Utah. And today we have with us, Jay Jenkins, CTO of Cloud Computing at Akamai. Jay, it’s great to have you on the show.
Jay Jenkins: It’s great to be here. Thank you so much for having me.
Swapnil Bhartiya: It’s my pleasure, and we are here at KubeCon. You folks make some very big announcements here. Before we talk about Akamai, I would like to talk about what kind of audience you saw at your booth, what kind of discussions you have had when you talk to analysts and, of course, you know folks who are at your booth and journalists. So just, just give us a pulse of the community. Yeah.
Jay Jenkins: So, yeah, we’re here on the second day, right? And it’s been an amazing first day. A lot of the people coming at our booth are, of course, developers that are working with the Kubernetes platform and are here to sort of really understand what’s happening in that area, as well as taking a look at their favorite upstream projects. Also, we’ve found a lot of systems integrators as well, sort of understanding what’s going on in the area and how they can help their customers overall. I would probably also break those developers down a little bit. You sort of have your your engineers that are building applications, but I’ve seen also a growing audience of platform engineers as well, who are actually developing platforms for their developers to build on top of which I think is a super exciting space for Kubernetes, because that really shows a growth to a level where these complexities are taken away from the developers and are sent more centralized for these platform engineering teams.
Swapnil Bhartiya: One of the big announcements was the Akamai App Platform. We have had a couple of discussions with Akamai folks, but I want to hear from you, as a CTO of Cloud Computing, what specific challenges that App Platform addresses for developers?
Jay Jenkins: As we are talking to our customers who are adopting Kubernetes, it’s obvious that Kubernetes is, is, is really just a start. You know, Kubernetes itself is also a platform for other platforms. And so what you have is this ecosystem that you need to build first before your developers can even start putting their applications on top of the platform for a lot of organizations, that can take a while to sort of figure out which parts of these upstream projects they want to deploy onto their Kubernetes platform. So we went out to the market and we took a look at what was going on out there, and we acquired a company called Red Kubes, which had built a platform called Otomi. And what was appealing about Otomi is is it allows a platform engineer or a developer to quickly start up a platform with everything they need in order to host applications. Right? You want to deploy Istio, you want to deploy Argo, you want to deploy Harbor, you want to deploy other upstream projects, input of which we support dozens. It’s just a few clicks away in order to do that. So instead of taking months to figure out exactly what you need in order to support the developers, you can do that in minutes. And the great thing about this is, even though we provide this opinionated architecture, these golden paths, which provide everything that you need, it’s also customizable. So for me, I go back to the words of computer scientist Allan Kay, which I use all the time, and I think about all the time, which is, you really want to keep the simple things simple. Which is why we want we provide these golden paths, but you want to make sure those complex things are possible, because that’s what Kubernetes is about, right? It’s about making sure that you have this capability to do these complex things that you couldn’t do before,
Swapnil Bhartiya: right? And when it comes to Kubernetes, some of the pain points that folks often talk about is, one is, of course, complexity, but I also, after having a lot of discussion and looking at what ad pump, it kind of simplifies. It lowers the barrier of entry for folks, because the fact is that this complexity is not going to go away. And complex complexity is part of, I mean, Kubernetes is not, you know, a simple LAMP stack there. So can you also talk about what role do you see app platform is going to play in kind of lowering the barrier of entry, making things easier, without compromising on the complexity that Kubernetes offer? Yeah.
Jay Jenkins: And one of the great things about this is, for those organizations that are big enough to have a platform engineering team, it makes it so much easier for a developer. But for those organizations that do not have that capability, or they’re just beginning that journey, imagine me as a developer thinking about all the tools I need in order to keep my application secure, in order to make sure I have a a software development life cycle taken care of. Imagine secrets, management, all those things that I have to think about, and now I can actually just spin up that platform and just with a few clicks, be able to deploy that in a consistent way, right so it’s already configured for me. Security is already configured for me, but it does give me the capability to modify that and change how I might deliver that. It’s also extensible, so you have the capability to create your own templates and patterns for other projects which might not be a part of CNCF or which we haven’t yet included in the app platform.
Swapnil Bhartiya: What are the other trends that you are seeing in this space. We hear a lot about, you know, web assembly, of course, you mentioned, you know, platform engineering is there. Of course, AI/GenAI is a big topic here, edge computing. And so a lot of things are happening now. The interesting thing is that when we look at all of this thing, and when we look at this whole ecosystem, Akamai seems to be a new player, you know, but Akamai, you folks have been around for a long time. You have pioneered the whole CDN security, and that’s how people look at you folks. So how should people look at Akamai in this new world?
Jay Jenkins: Yeah, it’s, it’s really interesting, right? So 25 years ago, Akamai invented the CDN, and so we come at Cloud, and we come at Kubernetes from a very interesting perspective, which is, rather than looking from sort of centralized computing outward, we’re looking for edge computing inwards, because that’s where we’ve always developed our applications. So if you think about the products that we have around CDN, around WAF, around our security products, they have all been developed at the edge. They’ve all been developed with distributed compute. And look, 25 years ago, we did not have the, you know, the capabilities that Kubernetes has today. So for us, if we thought about, sort of, if we could reinvent the cloud, knowing what we know now, how would we reinvent that? Well, the cloud is very much today, still very centralized, and it is still not as distributed, I think, as the original vision was right. The whole idea is that it’s a cloud of compute, and yet we still centralize a lot of our implementations, even if they are on Kubernetes, inside of a data center. So already we’ve seen a large move towards, obviously, things like content being at the edge, but also now serverless frameworks being at the edge. And we thought, what if we could actually have other types of applications at the edge with thicker compute. And we actually deployed virtual machines at 10 Edge locations this year to see how customers might begin to use those. And one of the things that they said to us was, you know what? We need more Kubernetes at the edge. We need it because that is the real enabler for that. We’ve done so many years of work on creating these cloud native architectures, and although they’re centralized, they are ripe for taking some of those services and moving them towards the edge. So there’s less affinity for a centralized infrastructure and more affinity for the customer or the device or the autonomous vehicle or whatever it is you’re trying to get closer to. We can actually take those containers, we can take those services and move them closer to those locations. And
Swapnil Bhartiya: when we look at CDN, and when they look at the whole push towards edge computing, I feel like it’s a very natural progression there. What is your vision? When I say your aka my vision for cloud computing, because if folks really pay attention to what Akamai has been doing, and what you folks are planning is, like, paving the path to where the next thing is, is, like, you know, saying it’s not where the hockey puck is, but it’s going to be?
Jay Jenkins: absolutely, and that is definitely where we’re looking. And we’re excited about what’s going on here at KubeCon, because we can see a lot of these new projects really embracing those ideas as well. WebAssembly being one of those, it makes these components even smaller, so we can actually move them in and out towards customers or onto devices or back into the cloud, depending on where that service needs to be. But if I look sort of in the future in terms of where I think a distributed cloud might head, you really only have to take a look at what’s happened, you know, so far. So we’ve already had content at the edge for 25 years. Now we’re seeing a lot of serverless functions at the edge, mostly stateless at this point, but you can sort of see a growing trend to look at different capabilities of the edge. So okay, I can have stateless functions now I can look, sort of cache, bits of content at the edge. Oh, well, can I actually have a distributed database, globally distributed database? Well, those things are now possible as well. Ok. Well, if I can have a globally distributed database, I can have more stateful types of applications at the edge, and if I can have more stateful types of applications at the edge, that gives me new sorts of possibilities on what I might do and the types of things that I might build. To give you an example today. You know, we’re very obsessed with availability zones, right? But availability zones are really sort of just in one region, right? Is that really redundancy? Or. So if we take a look at even you know, what the hyperscalers do with their own products, they globally distribute those services, and that’s how Akamai runs at CDN. It’s how we do our DDoS protection, because that is true redundancy. That is true resiliency. It’s around a distributed compute platform and putting those services where you need them globally and giving you that redundancy. And then again, as we move forward, imagine sort of being able to have this cloud that kind of follows you or your customers around, one that can sort of wake up whenever they wake up, and scale back down whenever they go to sleep. So the ability to, and this is very similar to a CDN. So you know, the content isn’t at the cache until somebody needs to fill the cache. Same thing with a process. The process doesn’t exist until somebody needs to spin up that process on that edge. So being able to do that dynamically based on, certainly, performance. But what if your goals are costs? What if your goals are sustainability of that particular processing location? What if the process, a processor type, is only available in certain locations? Well, I want to make sure I only spin those up in those locations. What if there is a legal compliance question? So a regulatory compliance question, so I need to keep data in a certain location. All those things are possible if you truly have a distributed computing capability. Computing capability. So that’s where we see things are heading.
Swapnil Bhartiya: And that’s also kind of the world we live in. The whole geopolitical situation is changing, yeah, countries, organizations are looking at data sovereignty. So how does that also play with the strategy that Akamai already has in place. Yeah, and
Jay Jenkins: That’s a, you know, a common question that we have from a lot of customers. So I’m actually based in Singapore, and so I do a lot of work in Southeast Asia. And, you know, the regulatory frameworks aren’t as robust as they are in the EU or as they are in the US. You know, there’s lots of different governments with lots of different policies. You know, for us, we think about distributed compute as a capability to also protect data. So as an example, let’s say that I have this fantastic AI, a healthcare AI, maybe it’s detecting cancer, and that means, in order to do that, I need to submit patient data into the model. Now, in much of today’s world, a lot of these hosted models would be hosted in the US as an example. So do I as a country want to allow that patient data to be submitted into the US in order to do the inference, even if it provides a lot of benefit for you know that that particular patient, is there a way that we know that the patient is in a certain location, and therefore I can keep the inference inside of that country, so that I can protect that user’s privacy and country? Can I go further than that? Can I actually say, Okay, so can I keep this patient’s data, even within the metro area. Can I keep this patient’s data in the hospital? Can they keep this patient’s data on the device? Now, of course, depending on the complexity of the model, the answer can be any one of those, right bigger models. You need more compute, so you might need to push that out to like an access point of the edge, but it doesn’t necessarily anymore have to go to a large data center
Swapnil Bhartiya: when it comes to this kind of data sovereignty, and I’m not talking about the CD and the cloud part, but sometimes organizations in different countries end up reinventing the wheel that is based on resources, and that also defies the whole person of open source. So the beauty of open source is that, of course, there may be trade embargos or whatever it is. But while countries can compete, but community and the code based, it doesn’t matter where you are based us, of course, Europe, Asia, Africa, even Antarctica, Antarctic so can you also talk about how to kind of stop that waste, how to stop the reinventing the wheel, how this edge model and data sovereignty can enable where you don’t have to move your patient data all the way to the US, and being able to leverage the technology locally? Honestly,
Jay Jenkins: It’s a very tough question, because a lot of future work in terms of data sovereignty and privacy depend on growing capabilities in space. So there are, there’s a lot of work going on around obfuscation, but still allowing the AI model to do its work. And certainly that’s a valid path, but that’s a path that’s going to take time, you know, and the models that are fast to market don’t have those sorts of capabilities. So, you know, I do think it needs to be a balance. I do worry about sort of over investment and waste in the space, in particular, always chasing sort of the biggest, best GPU, because. A lot of times that can encourage waste. There’s a lot of AI models that can be run quite efficiently on a CPU right or a smaller GPU, and in fact, can be more price performant for those particular use cases. I think for you know, as you look at sort of data sovereignty, there are other things you can do. So AI models are kind of difficult, because you do need to have that data, you know, those specific inputs that are required for the model. Other types of data sovereignty, however, can be sorted by various things. Now, imagine you have an access edge point, and you can actually do the encryption key based encryption at that access edge, but still have that data stored somewhere else. So now all of a sudden, you have a capability to store data someplace else, but you keep the keys inside of the country, and it gives you an extra level of redundancy, because that is stored now somebody else, somewhere else, but the ability to encrypt and decrypt that data stays at that access edge point. So the same thing that you know you might do on a device today in terms of end to end encryption, you can also do that on access edge again, if you’re if you are a government and are worried about that, being able to do that at the access edge before it leaves the country and when you bring it back into the country, what
Swapnil Bhartiya: What role do you see my app platform can play in all the scenarios that we just talked about.
Jay Jenkins: Yeah, so the Akamai app platform today is really focused, first of all on the community and upstream projects coming out of CNCF. We really want to make sure that it is a full Kubernetes ecosystem that gives you everything you need to get started inside of the ecosystem. However, with its extensibility, it certainly has a capability of number one, being able to put your own customized applications that might not exist inside of CNCF or your own in house enterprise applications into those projects as well. But over time, we really see this as being a way to establish a platform that might be multi clustered that might extend out to the edge, or, in fact, beyond the edge. So we’re working with a lot of partners in that space to help us think about how we might solve those kinds of problems. But we’re also attending, you know, events like KubeCon, and we’re Gold members with CNCF, because we understand, you know that the community can goes back to the that adage, right, that all of us are smarter than any one of us, so being able to be a part of this community and talk to people really helps generate new ideas around the types of problems that we might have as we do this, but also the types of possibilities to different ways of solving those same problems as well. Jay, once again, thank
Swapnil Bhartiya: you so much for joining me today and not only talk about Akamai App Platform, but also the broader picture, the broader world that we live in, and what role Akamai is playing and going to play in this Kubernetes ecosystem. Thanks for sharing those great insights. And as usual, I would love to have you back on the show, but I really appreciate your time with it. Thank you.
Jay Jenkins: Thank you so much for having me. It’s great to be here at KubeCon.
[/read]
What kind of industries rely on Akamai’s API security solutions | Rupesh Chokshi
By 
