>> Presenter: From San Francisco, it's theCUBE. Covering GitLab Commit 2020. Brought to you by GitLab. >> Hi, I'm Stu Miniman, and this is theCUBE's coverage of GitLab Commit 2020 here in San Francisco. You might notice some of our guests have some jackets on. It is a little cooler than normal here in San Francisco, but the community and knowledge is keeping us all warm. Joining us for the first time on the program is Nicholas Klick, who is an engineering manager at GitLab. Thanks so much for joining us. >> Thanks for inviting me. >> Alright, so you had an interesting topic. The state of serverless in 2020 was the session that you gave. Definitely a topic we love covering on theCUBE, something I personally have been digging into, trying to understand. Definitely something that the developers, and especially the app devs that I speak with, are very bullish on, so what is the state of serverless in 2020? >> That's actually a good question. So, my talk was actually broken into two parts. One was, like initially I just wanted to help provide a clear definition of what serverless is. In my opinion, serverless is more than just functions. There are a lot of other a lot of other technologies, like backend is a service, API gateways, service integration proxies that you can stitch together to create dynamic applications. So, I created a more expanded definition of what serverless is from my perspective, and the other part was to really talk about three things that I'm finding exciting right now in the serverless space. The first was Knative, and the fact that Knative is likely going to go to GA pretty soon, so it'll be production ready, and we can finally build production workloads on it. The second is that running serverless at the edge I find to be an exciting topic. And then finally, talking more in depth on those, the service integrations. Of how you can actually create applications that don't include functions at all, so functionless serverless. >> Yeah, so a lot of things I definitely want to tease out of that, but Nicholas, I guess maybe we should step back a second-- >> Nicholas: Okay. >> And was there survey work, or was there something done, or is this kind of something related to your job that you put together as just an important topic? >> Yeah, I know this is just me speaking as someone that works in the space and sees the technology is evolving and just my opinions, I guess. >> Okay, when I talk to the practitioners, when you go and say, "Oh, they're interested in it." Chances are they're doing stuff on Amazon, is like what kind of the first piece of it tends to be. There are lots of open source projects out there, but it's still this kind of dominated by Amazon. Azure has some pieces, of course. Google has things they're doing. I liked how you teased out that serverless definitely isn't a thing, and the definition, and even the term itself, gets people all riled up and things like that, so I hate getting into the ontological arguments, but the promise of it is that I can build applications in a different way, and I shouldn't have to think about some of the underlying components, hence the name serverless, kind of-- >> Right. >> does that, but it definitely is a change in mindset as to how I build and consume environments. >> Right. Right, and like another point that I made in the talk, that I believe pretty strongly, is that serverless is not something that's going to replace monoliths and microservices. I believe it's another tool in the tool belt of the developer, of the operator, to solve problems, and that we should look at it like that. It shouldn't be, it's not the next progression in application architecture. >> Yeah, I've met some companies that are 100%, they've built everything on serverless, but that's like saying I've met plenty of companies that are all in the cloud. It depends on what you do and what your business is. >> Nicholas: Right. >> When we look at the enterprise, it is a broad spectrum, and making changes along that path is something that typically takes a decade or more, and they have hundreds, if not thousands of applications, and therefore, we understand. I've got my stuff running on my mainframe through my latest microservice architecture, and everything in between. >> Right, and I mean I'm speaking as an employee of GitLab, and we have a very well known monolith that we deploy, and so for my opinion, I don't believe that monoliths are going to die any time soon. >> Alright, I'd love you to tease out some of those pieces that you talked about, the three items you talked about: Knative. You know, Knative is interesting. The thing I poked at when I go to KubenCon and CloudNativeCon is today I mentioned when I think about customers, most of them are using Amazon. The second choice is they're probably doing Azure, and today Knative directly doesn't work with EKS, AKS, or the like. I know there's a solution like trigger match that actually will interact-- >> Right. >> Between the Amazon and there, but don't you need the buy-in of Amazon and Microsoft for Knative to be taken seriously. And the other thing is, Google still hasn't opened up the-- >> Right. >> the Google controls, the governance of both Istio and Knative, and there are some concerns in the ecosystem about that, so what makes you so bullish on Knative. >> Yeah, so I'm definitely aware of some of the discussions around Knative. From my perspective, I think that Knative is, if someone is already operating a lot of Kubernetes infrastructure, if they already have those, that infrastructure running, then deploying Knative to it is not that much more of a it doesn't require additional resources and expense, so it could be, again it depends on their use case, and I think that, when I think about serverless, I try to remain pragmatic, so if I'm already using Kubernetes, and I want a simple serverless runtime, Knative would be a great option in that situation. If I want to be able to work cross-cloud, like this is another opportunity that Knative provides, is the ability of deploying to any Kubernetes cluster anywhere, so it has that, you know, that, there's not a vendor lock-in issue with Knative. >> Yeah, and absolutely there was initially some concern that, could serverless actually be the ultimate lock-in? >> Right. >> I'm going to go deep on one provider and don't have a way. There, open source groups like the CNCF trying to help along those ways-- >> Sure. >> Knative absolutely along those ways looking at that environment. From a GitLab customer's standpoint, GitLab's not tied to whether you're doing containers or serverless or VMs or in the environment. What does it mean for GitLab customers? If I want to look at serverless, how does that fit into my overall work flow? >> Yeah, so initially at GitLab we focused on providing the ability to deploy to Knative. That was, we were very early in the Knative space, and I think that as it's matured, as those APIs have matured, then our product has kind of developed, and so right now we enable you to be able to create Kubernetes clusters through our interface and then deploy your function run times directly from your GitLab repo. We've also, are kind of growing in our our examples and documentation of how to integrate GitLab CI/CD with Lambda. That's another big area that we're moving into as well. >> Great. As you look forward to 2020, we've got a whole new decade in front of us, what should, what do you think people should be watching on in the maturity of this space. >> Yeah, so I think that the point that I touched on earlier of the service integrations, I think that that is something you're going to see more and more of. Of the providers themselves linking together their different services and enabling you to create these dynamic applications without a lot of glue that you have to manually create in between. I think that we're going to see, you know, more open source frameworks, like, for example, Service Framework or Terraform that people want the, I mean, I know that a lot of people use, for example, AWS SAM. People want easier ways, and faster ways, to be able to deploy their serverless, so you have the bootstrapping of serverless. I guess, another thing that I expect is that the serverless, the serverless development life cycle will mature, in that whether going from bootstrapping to testing, deployment, monitoring security, I believe you're going to see companies that will start to really fill in that entire space, the same way that they do for monoliths and microservices. >> Yeah, absolutely. Thank you so much, Nicholas. Definitely something we've been tracking over the last year or so. You start to see many in the tool chain of cloud native environments digging into serverless, helping to mature those solutions, and definitely an area to watch closely. >> Great. >> Alright. Lot's more coverage. Check out theCUBE.net for all the events that we will be at through 2020 as well. If you can go back and see we've actually done Serverlessconf a couple of years, many of the other cloud and cloud native shows. Search in our index. I'm Stu Miniman, and thank you for watching theCUBE. (energetic electronic music)