Brainwave Podcast: Introducing the Cloud-Native Computing Poster

Intellyx co-hosts: Jason Bloomberg, Jason English.

Topics covered:

• Why make a Cloud-Native poster now?
• Why doesn’t this big picture list a lot of tools and/or vendors in the space — not even Kubernetes?
• At a high level it’s broken down into three sections: Cloud Native Architecture, Infrastructure, and Applications. What’s the rationale for that division?
• Hone in on some areas of Cloud-Native beyond the containers, integration and orchestration of deployment aspects – Software-Defined Networks, DevOps, Cloud-Native Zero-Trust Security.

• Get your own copy of the PDF of this poster now: https://intellyx.com/cncposter
• Listen/download the Podcast on your favorite channel or player here: Intellyx channel on Anchor.fm
• Watch the YouTube version here: https://youtu.be/TQz6NpU17r8

Transcript of the podcast:

Jason English: Welcome to the Intellyx podcast. We’re on our Cloud-Native poster series here. and I thought, what better way to kick this off then and invite the author and a conceptualizer of this fine piece of work. It’s Jason Bloomberg president of Intellyx.

Jason Bloomberg: Hi, it’s great to be here with you, JE.

Jason English: Yeah. you know, we’ve had a lot of interest in getting this poster out. It’s been a long time in the making as you know. But, I’m just kind of curious, I think the readers and watchers of the show would like to know, why did you come up with the concept for this poster in the first place?

Jason Bloomberg: Well, Cloud-Native computing as an overall trend in enterprise IT is really gaining steam. It’s becoming essentially a paradigm shift in how, large organizations in particular leverage a technology to meet, application-enhanced business needs. But one of the challenges is there are many different parts of the story and that leads to a lot of complexity and confusion. So it was sort of a great topic for this kind of, this kind of poster. So essentially Cloud-Native computing means that, we’re taking the best practices of the cloud, like scalability and resilience and other core practices and extending them to all of enterprise IT, whether it’s in the cloud or on premises, or on the edge. In such a way that we’re bringing a unified abstraction so that we’re treating all of these different parts of the Cloud-Native story in a consistent manner. And in order to get that to work, it involves a lot of hard work under the covers.

And that’s really what the poster reflects. [Small view of Intellyx Cloud-Native Computing poster, below.]

Jason English: Hmm. what I’ve noticed is kind of other views of cloud-native in general, tend to focus more on, tooling or, different sets of tools or lists of vendors. So, now, why did you approach it this way? When you categorize the solutions in this way?

Jason Bloomberg: Well, in terms of some of the tools and technologies that support this vision for, computing, a lot of the, discussion centers on Kubernetes, essentially an open source container orchestration framework.

And Kubernetes is a key part of the story, but it’s not specific to any particular product or open source effort. So open source in general is a major, supporting player in the cloud-native story. But it’s not about any particular, product or a project. So, so you’ll see that in the poster, right?

There’s no mention of any vendors other than sponsors, of course. but within the poster, there’s no mention of any vendors or any, open source projects because it’s really, at a higher level, right? There are many different ways to implement Cloud-Native computing.

Jason English: Yeah.  So at a high level, you have it broken down into three major sections here.

It says, cloud-native architecture, cloud-native infrastructure, and cloud-native applications. So what’s the rationale for splitting it into those categories, from the start?

Jason Bloomberg: Well, a lot of the discussion of Cloud-Native computing centers on the infrastructure, or how are we going to, implement containers and microservices in a scalable way?

How are we going to deal with clusters and other parts of the infrastructure story? And then once we have that, how are we dealing with service meshes and security and, and all of these, you know, integration and all of these infrastructure-centric issues. but. If you only focus on the infrastructure, you sort of have a hammer and you still have to wonder about your nails, right?

What problems you’re looking to solve? The problems you’re looking to solve are really application centric, right? You’re providing an infrastructure that supports workloads, that represent applications that are going to meet business needs. So tying the infrastructure to the applications is a key part of the story.

But there’s still more to it than that. And that is how everything fits together and that whole, fitting together part of the story is the architecture. And we’re really looking at different ways of architecting an enterprise infrastructure to support applications, following cloud-native principles.

So it’s really three parts of this puzzle. There’s the infrastructure, there’s the applications and there’s also the architecture, and — It’s a really a good stool, right? You need all three of them, you can’t be missing any one of those.

Jason English: Right. do you think a lot of the initial cloud data projects were probably scuttled by, maybe a lack of architecture, architectural foresight or something like that?

It seems like we’ve seen, especially over the past two years or so at the beginning, there seemed like there were more companies kind of going down this path or they would just try like, “We’re going to implement Kubernetes and we’re going to start on this path,” but then they kind of fail and have to start over again.

So you think that’s a common challenge that we’re looking at?

Jason Bloomberg: Well, actually, you know, that that’s an important part of the history sort of how we got to where we are today. Cloud-Native architecture didn’t sort of spring out of the blue fully formed, it’s part of an ongoing process of maturation of architectural approaches.

You know, we talked about service-oriented architecture in the 2000s, that led to REST-based SOA, which is really a transitional approach to architecture. And that led to microservices architecture. And microservices architecture is essentially a precursor of cloud-native architecture, and people do confuse them because obviously we use microservices and cloud-native architecture as well.

But what microservices architecture was really missing is this overall architectural context that provided the scalability and resilience at scale that we really needed. Right. If all we had are microservices, then you end up with a lot of microservices that ended up talking to other microservices and you ended up with the whole thing bogged down as you try to scale it up, you have a point-to-point problem, and it’s very hard to build an infrastructure that resembles the cloud in the sense we’re looking for. So there was this whole phase where microservices came on the scene and we started using Docker.

Kubernetes was still really too immature. And there were actually a number of different container orchestration platforms. And it wasn’t clear that Kubernetes was going to be the leading one. And for a while there, it was about containers and microservices without the orchestration infrastructure.

And at that point we really couldn’t, didn’t have all the pieces in place to build something that was going to provide the capabilities that were required. So once Kubernetes sort of became the leading approach, then, Mesos and some of the others, Docker Swarm sort of fell by the wayside, then. The community at large, right? The open source community who drove this really put together this approach to cloud-native computing that would at least in principle provide the scalability, resilience, and other cloud principles that we require. but we had to go through that transitional phase that microservices architecture to get to that point.

Jason English: Hmm.

I think there’s a few areas of this poster that I thought were kind of interesting. Usually, I don’t know if people would normally associate them with cloud-native, but they do have a place in there. For instance, you talk a little bit about software defined networks in there.

So how does that fit into this context, do you believe?

Jason Bloomberg: Well, at the core of cloud-native computing is the cloud-native infrastructure and you can’t really build a enterprise IT infrastructure without a network, right? So the network is a part of the story, but we need to build, all of our infrastructure, including the network in a way that enables us to abstract the physical aspects of the technology. So we want to be able to abstract the physical network. We want to be able to abstract,  IP addresses and TCP ports and, VLANs and all of these other, physical parts of the network in a way that we can come up with, an approach to networking that supports the diversity of different environments that is necessary for, cloud-native computing, the HybridIT.

Which is another part of the Cloud-Native story, where we have a mix of different environments and mix of different technologies and deployment options, that are intentional because the mix itself meets the business need. And now we need to have a consistent coherent way of dealing with all parts of the infrastructure from the network on up.

So we don’t want to have to deal with a network and the cloud differently from how we deal with it on premises. We want to be able to abstract the differences so that we can have a single business-driven approach to the entire network.

Jason English: Yeah. And in some sense, I see an adjacent space would be this cloud-native edge computing, and Internet of Things, it seems like that’s kind of the differentiator between this Hybrid IT approach and, just deploying applications for use by, your normal browser or computer-based users. You have, all these mobile applications, things involving tons of devices that are starting to appear, and then the Telcos even getting involved. Do you feel like the edge is sort of driving some of this broader perspective?

Jason Bloomberg: Well, yeah, you could say the edge is driving, the broader perspective or cloud computing paradigm is driving the growth of the edge. It sort of works both ways, but the edge itself is a very dynamic fluid notion, right? Different people define it in different ways and there are many parts to it. Right?

You have the Cloud Edge, which is where the CDNs live, right? Bringing the edge of the cloud closer to the user for better performance. Then you have the Near Edge, which is essentially points of presence or cell towers the telcos are building out supporting a 5G rollout, but it also would include, IoT gateways and remote equipment phone closets. It’s in retail establishments and remote offices.

And then you have the Far Edge, which includes, all of the end user devices. So mobile devices, et cetera, as well as IoT sensors and actuators that are all part of the Far Edge.

So if you start thinking about this as a networking challenge, we have many different pieces of the story, many different protocols for connecting these devices. And 5G is rolling out with a whole new set.

So we have a lot of complexity again, with the infrastructure. What cloud-native computing brings to this story is a coherent abstraction. Like how can we deal with all these different parts of the edge in a business-driven fashion that it makes sense across the board?

SD-WAN technology is part of the story, but SD-WAN and of itself is evolving. Right? Talking about next generation SD-WAN technologies that bring software defined approaches to the broader context. And really the idea here is that the cloud-native computing paradigm is an organizational principle that gives us a way of thinking about the big picture in a way that we can support with the technology itself.

So it’s a way of dealing with the increasing levels of complexity that we have no option but to deal with.

Jason English: Yeah. another side of it really is, you know, that I think we’ve been encountering in our briefings over the last couple of years, especially as, you know, challenges around getting, teams up to the level of skill to deploy these things.

Or even the level of understanding to solve it. So, you know, what are some of the best practices on that side of the house in terms of, getting teams involved, making them more productive in this new paradigm?

Jason Bloomberg: Well, one of the interesting things that’s going on, and this is, some very good decisions by the cloud-native computing foundation, the CNCF who is, shepherding, Kubernetes and a number of the other open source projects related to cloud-native computing.

One of the things that they decided is to make the entire Kubernetes ecosystem, extensible, as opposed to customizable. And this was very important. If you think about it, this is not the way open source traditionally works. If you think about how Linux works, Linux is customizable, right? You could go in, and you could monkey with the kernel, and you could do anything you wanted and then redeploy Linux.

And so you had different vendors come up with their own flavors of Linux. And now that’s where we are today. Multiple, mostly compatible, but somewhat incompatible flavors of Linux. And the CNCF said, wait a minute, that is sort of the wrong direction. We want to build a core set of capabilities that are extensible in order to give the broader community, both of the open source community, as well as commercial vendor community, the ability to build additional solution.

Yes. So this is sort of a long roundabout way to answering your question. What’s happening now is we have this ecosystem of commercial vendors who are essentially filling in the gaps on what the open source offerings can’t offer, or really aren’t intended to offer. and as well as the support and maintenance and training that are needed to use those.

So we have this whole class of enterprise-class commercial offerings in the Cloud-Native world that are leveraging open source, but help provide enterprises with the skills and the training and the support maintenance that they need in order to really have an enterprise class, you know, deployment at scale.

So that’s one of the key things that’s happening. No, obviously we, you know, we have a lot of demand for technical people who understand this stuff. but in terms of how the community is handling this really it’s driving this extensibility, which is now supporting commercial opportunities, which of course drives investment.

And, that’s really the fundamental answer to your question.

Jason English: Yeah you can see that they’ve left some of the wires hanging, you’ll see things like, multitenancy would be something that’s kinda left out of the original equation and also, some things like defining which languages you’re using to write the actual code itself.

And especially I’d say security is another area where there’s a lot of white space for innovation going on right now. What are your thoughts on that side?

Jason Bloomberg: Yeah. And that’s intentional, right? If you look at sort of the core Kubernetes codebase, there, wasn’t a lot of security built in and that’s intentional because it’s essentially, we left as a future exercise either for open source projects, but really primarily for the commercial vendor community,

Security in particular is very difficult to get right up front when you’re building, an entire open source initiative. It’s one of those things that you have to sort of figure out what you’re doing and then figure out the security implications. And they may have to go back and rethink some things, but it’s better to have, an extensible approach to security so that you can build robust security products, in the context of the existing infrastructure.

And that’s what’s going on now. So the new thinking about Cloud-Native security, and in particular, what I like to call cloud-native Zero Trust. Now, Zero-Trust is an old Forrester term. Forrester came up with it in 2009. And the idea is that you want to assume that all parts of your network, both internal and extra leader, corporate networker.

Untrusted until you explicitly assigned some sort of trust to them based upon the identity of the user of that particular resource. The problem there is that based upon the identity of users that is human users, as opposed to, Abstracted Identity, which is what we need in a cloud-native world, because we may have any number of different kinds of end points.

They may or may not interact with human users. We may have abstracted IP addresses. We may have abstracted devices or abstracted IoT. components, or we might have microservice endpoints or APIs, and all of these now have to have an abstracted identity that may not even correspond to a human person, a human identity, or may not correspond to an IP address and may be dynamic or, fluid and ephemeral in some way.

So we have this now and this new notion of cloud-native zero trust that takes the older principles of Zero-Trust, having a zero trust network fundamentally, but abstracting the identities so that we can assign trust as needed to any abstracted end point, regardless of what kind of end point it is.

And there’s a handful now of vendors who are, implementing those technologies as well. That’s going to be a critical part of the overall cyber security story, because it’s really addressing security of any kind of endpoint. So, if you look at the IoT, with security issues and, holes, because it’s such a large expansion of the enterprise threat surface and all of the existing technologies were built before the IoT came along.

So the notion of Cloud-Native zero trust security, basically says we don’t care if it’s IoT, we’re going to abstract all of the end points. And so it’s a way of essentially saying, I don’t care how big your threat surface is. We’re going to be able to deal with the whole thing. And, that’s the way we really have to think, in the cloud-native world.

Jason English: Yeah. And it changes the assumptions of how you handle logic and your expectations of all of the participants, in a chain of different services that might make up your application in the end.

I think we’ve covered architecture, infrastructure and then, rolling up to the upper-right side of our diagram here, we can talk a little bit about DevOps and how continuous delivery, agile into continuous delivery, into DevOps movement has evolved and kind of become part of the picture in the Cloud-Native development world.

Jason Bloomberg: Right. So that’s this, this is part of the applications part of the story.

And, you know, obviously if you’re building applications, you want to follow modern best practice. And that may very well align with DevOps and CICD but there’s more to the story, right? If you look at some of the fundamentals of DevOps and CICD, Infrastructure as Code is a part of the story.

We want to be able to create, essentially, declarative representations of our infrastructure, as well as our deployment context, so that if we have something deployed in production and we need to fix it, we essentially can change the, this abstracted declarative representation and redeploy, and that’s the notion of Infrastructure-as-Code, but we want to take that one step further, right?

Because we don’t want Infrastructure-as-Code to be code. We want it to be essentially an abstracted declarative representation of the infrastructure in question that we’re looking to deploy, whether it’s workloads or applications, microservices, containers, or, components of our infrastructure.

So, this is part of the Cloud-Native approach is saying, well, the whole thing is essentially declarative and configuration based, and we want to make any changes, we can change some sort of abstracted representation of our infrastructure and redeploy. We’re never in a situation where we have to go in, or we would want to go in and say, look in a Kubernetes pod or a Kubernetes cluster and change something there.

Right. Because you know, Kubernetes. Containers pods and clusters are ephemeral, right? They come and go as part of the scalability and resilience functionality of the infrastructure. So the last thing we want to do is go in and monkey with a pod because it might just disappear the next day and then redeploy, and our changes would be lost.

So we have to work at that abstracted, declarative level for everything we do. So it’s taking infrastructure as code and really saying, well, this is a fundamental principle, not just a nice thing to have when we’re deploying software, right? So we’re taking this approach and extending it essentially architecturally across the infrastructure.

And that’s a fundamental part now of how CICD works: continuous integration, continuous delivery or continuous release is really where we want to get, with CICD. In the context of a cloud-native infrastructure where, everything is represented declaratively, and then we can deploy as needed, into an inherently ephemeral stateless environment.

Jason English: Yeah. It’s an interesting picture. I’d say. And, I definitely liked the steam punk theme of this one. Yeah, I think it’s a little bit fitting with all the other bells and whistles and gadgets that are involved in making this work in the end. But it’s, not as complicated as it seems when you think of it this way.

Jason Bloomberg: Well if you look at the poster, especially if you don’t have the big version. If you’re looking at it on a screen, which most people probably will do, it looks really complicated. There’s a lot of different things to it, but that’s not the problem with the poster. That’s the nature of cloud-native computing.

Right? So that’s part of the story we’re trying to tell, is that it does have many different moving parts. There are a lot of ways, different things are connected to different things, and the steampunk theme just, essentially emphasizes the fact that we have a coherent technology, technological perspective on the entire thing.

So we can think of it. You know, steampunk is essentially a new perspective on Victorian technology, a modern perspective on Victorian technology. We can say, well, we have a modern perspective on Cloud-Native computing, and that’s what we’re communicating with the poster.

Jason English: Yeah, it has a little bit of the mystical feel, but also the tinkering with engineering and stuff like that.

So it seems like a good fit. Well, great. well I hope lots of people download and find some use out of this poster. I think it could be something you might want to stick it on your team wall, if there is a scrum room, again in the future, but, I think it could have value in helping, describe these concepts for people who are new to it, as well as people who might have filled in some of the puzzle, but have other parts open.

Jason Bloomberg: Yeah. And if you’re familiar with our other posters, Intellyx had an Agile Digital Transformation Poster from three to four years ago, as well as other posters that I helped put together at ZapThink. Our SOA Roadmap poster was enormously popular. We went through three different versions.

and those, you know, we gave away. At conferences and we included it in magazines back when we had magazines, you remember those days, and, give them way and people would, buy them from us, on our website and everything.

Well, you know, obviously we don’t have conferences at the moment, so you’re not going to be able to get a print copy from us because there’s no reason to print them until the conferences resume.

But you can get the downloaded version, on our website. you can print that if you want and put it on your wall. And obviously when the conferences resume, and there’ll be many different ways to find them, you can find them, you know, from any of our sponsors or if you run into us, at other places.

And the CNCF is one of our sponsors, so I would hope that they would distribute them at all of their events as well.

Jason English: Great. Yes. And if anybody’s looking for the poster, just to mention here, you can get it at https://intellyx.com/poster — that will have this poster as well as some links to the previous one that we’ve done as a firm.

Thanks Jason. For joining me on this introduction to the Cloud-Native Computing Poster.

Jason Bloomberg: Yeah, my pleasure. And so the Intellyx Brainwave Podcast, right? It’s our new name for it?

Jason English: Yeah,

Jason Bloomberg: Well, no, we haven’t come up with anything better.

Jason English: We invite everybody to check it out, and visit our great sponsors as well. OpenLegacy, Broadcom, Cherwell, and Unravel, Volterra. MemSQL and the Cloud-Native Computing Foundation.

And I think that’s about it.

Jason Bloomberg: Okay. Very good.

Announcer: Thank you for listening to the Intellyx podcast. If you have any questions or ideas for future episodes, feel free to drop us an email at PR@Intellyx.com. Until next time, keep on transforming.