Cloud-Native GitOps – Alexis Richardson, Weaveworks

This is texturing TV. Hey guys. Thanks for the throw.

We're here with Alexis Richardson. Who is the CEO for weworks? And we're talking about some Milestones as it relates to Cloud native and what we're seeing with gitops and all kinds of good things Alexis.

Welcome the show. Hey Mike. Thank you.

You guys seem to be growing the Market's taking off. We passed that one billion container image pole Benchmark or Hallmark that a lot of people look at something that says, hey, we're here to stay and my question to you is it seems like this is taking a long time not get UPS per se but the whole Cloud native container shifts and is it your perception that this is happening slower than thought maybe on Pace or we have some ridiculous and expectations in the first place. We definitely had ridiculous expectations in the first place.

We always have ridiculous expectations in the first place. I think what's really important to set this in perspective is at least from my seat. This is like the 90s.

There is a fundamental cyclical change in the nature of the internet itself, which all the vendors are involved with this isn't like openstack where there's one vendor Amazon and this plan to kind of try to undermine them using something else or previously. We saw Red Hat against use as a weapon against Solaris back in the day, you know, you see these occasional battles, but now I think this is a cyclical change they move to a different kind of Internet IPv6. Everything is connected and you know, I feel like I'm closer to the end of my career than the beginning but I feel like you know that the fabled internet of devices is kind of upon us soon and you know, the Blade Runner world of arvr on all that stuff isn't too far away and this means that all of the substrate is changing the same time.

The telcos are moving to 5G and the success of Technologies, but hyper fast Wireless and Telco bandwidth the clouds and out everywhere. And the technology to run applications which Google list together has been with us for a few years now in the form of containers and kubernetes, which is a sort of iteration of Linux and app servers really and so for me when people ask me, where are we I think it's like 1997. We've just seen a couple of businesses start and do well and the new era which give us hints of the future, but we have not yet seen Mass uptake of the Technologies around the edge.

Yeah, but things like cars and planes and trains will all be connected soon. And this will become very important politically as well because of things like climate change and Net Zero, which I'm sure will be a big driver of consumption as well as how we run it to And so I think it's a slower change Than People expected. I must admit that we too were part of those who thought it'd be quick.

When I started we've worked with my friend Matthias. We thought this would be like the cloud where Amazon appears they changed the nature of utility computing suddenly. It's got a name.

It's called the cloud. It sounds a lot cooler and you can pay for it with a credit card just like buying a book on the bookstore. This is kind of cool.

So we start doing compute and then suddenly things to take off really quickly. And then you have new businesses like New Relic provide value over the top and Heroku. We all thought that the cloud business would be a bit like the container business and Cloud native, but it was held back by one key thing, which is we're bringing the Enterprises with us this time and that we're changing the infrastructure behind the firewall and in the edge as well as in the cloud so that slows things down by a factor of I don't know three four or five and that's why I think we're still in the early Hills of a big change.

It seems like Enterprises always move slower than people anticipate and even as we move along here and we start building out more of these microservices container base Cloud native applications. I'm the bulk of what they're running is going to be Legacy for a while. So what's your sense of how long this transition is going to play out for?

I think that that we've got at least two more decades. It's the transition to see it through and it's fullest. I think the next 10 years is gonna be doing over the humps.

So to speak where we start to attach the Legacy apps. And I think we'll see that in two stages. So we're in stage one right now where all the cloud vendors and other folks have moved behind the firewall now with Enterprise and hybrid offerings Based On Tools, like Cloud native and kubernetes and you'll hear Cloud native more and more referring to these hybrid deployments and not only to containers and kubernetes, but other things like paths and servers in that context.

This will take a few years. But then what we'll see is we'll start to see a way for actual Legacy applications to much more efficiently move into an automated data center or into a public Cloud where they will run until they're thrown away and I think the technology shift that we're seeing now and we've worked as part of this actually is through the ability to attach non-containerized workloads two kubernetes. So what we found is that kubernetes is a great tool for doing orchestration and management as infrastructure.

You don't necessarily want to see all of it if you're a developer but as infrastructure, it's fantastic for this if we could attach Legacy apps. Databases API gateways and all this stuff that people have been building for the last 15 or 20 years onto kubernetes that we could maybe use kubernetes to manage that stuff for us. Which would mean things like checking that the app is available and if it isn't waking up a spare copy active passive failover being able to root the app if it's moved these are really basic functions, but for some reason they've been quite hard to do in the past using proprietary virtualization Technologies.

So have some combination of kubernetes containers lightweight virtual machines may show us away to bring these Legacy applications into the the automated and managed future and I think that's the second phase after that. I think it will start and we'll see, you know, the the consequences in in a future decade or whatever that gets to. All right.

So some convergence is coming our way. I'm What exactly so you have accomplished some things in the last year and seeing some rapid growth? So when you think is you look at those numbers where we're gonna be six months in a year from now and what's your sense of?

What is the pace run? I think we very much at the Crossing the chasm to the mass Market stage for most of these Technologies. Now all of the big Acquisitions in kubernetes that happen a couple of years ago have had time to settle in all of the vendors are offering that as part of their services.

If every Enterprise now has pick and shoes of the commodity orchestrator offering. Yes, there are still complexities, but now we're actually focused in on shipping the applications and the transition I'm seeing is people going from a couple of apps done by hand to automated application Management on top using guitars and that's why we've seen two billion downloads of flux for example, and you know traditional analysts firms like Gartner are starting to talk more about good Ops than ever before because their Enterprise customers are asking them for it. We also seeing it embedded into Microsoft's platform with the we've works and flux and as you're launched GA with good Ops earlier in the Amazon is doing similar things and other big vendors to so I think that our role in this continues to be to help customers to play applications to that to their venue of choice multi-cloud or hybrid Cloud you choose your architecture will help you roll it out and automate the management using gitups.

Some new changes for us. We did this acquisition. I think we last time we spoke.

We spoke about acquiring maglix. That offering now is fully part of our Enterprise Products. So policy as code or automated compliance.

If you will can now be part of this. This is really important for these Enterprises very few cios want every single thing to be automated without also having checks and balances either manual or programmatic and that's now part of the standard stack then also as I mentioned we're seeing people asking well, hey, what about not just containers and no Enterprise has only containerized apps. So they're all saying I've got these virtual machines.

I've got bare metal. I've got Edge deployments. I've got data centers.

Can you bring all of this under one sort of management theme and the answer is yes with kubernetes and get Ops. We have an operating model that can bring all these things into one place and that's what we've worked product also give you today. So this is great news, I think.

Do you think at some point I mean, I'm not sure how many people actually know what getups me. And so maybe you might want to explain that a little bit but also as part of this do you think the ultimate success is maybe when we don't use the word get us because it's just way everybody kind of normally does things. Absolutely.

I think that get UPS like kubernetes will become part of the very standard. Jargon. So commonplace that people don't talk about it along with Cloud native as well.

So that's for sure and get up today basically means automated management based on artifacts that's stored in git and image repositories. Like, you know, Dr. Hub, for example, he's a place is where you store immutable information about variable systems, like data centers app stacks and clusters and the way that you manage them and get Ops is using agents which manage them for you remotely and securely without you having to touch them at any scale you like and that is a beautiful thing because it means that we can generate a lot more infrastructure and manage it without increasing the head count of our operation and while reducing errors because it's done much more programmatically.

And I think that to your point. Yes. I'm already seeing people starting to talk about what you can do with getups rather than get UPS itself.

So platform the next generation of application platform the next generation of serverless. We're seeing people bringing offerings, which you get Ops managed and when you look at the details of their marketing and their website, so good Ops is mentioned but in the small print because it's it's an information about how it is done. It's a checkbox rather than a super interesting new thing about how this about what you're getting.

So I agree with you 100% What do you think so far has been the challenge about making the transition to Cloud native microservices get UPS, whatever you want to kind of combine together to kind of make all that happen, but it seems like organizations are still finding in a little bit difficulty and it goes beyond kubernetes itself. It seems like they're struggling with processes and how they bring all that together. So what are you hearing from folks?

So many things it's complicated. There's too many moving parts. We lack the skills.

We don't have a plan. What's the operating model how to make it secure? Those are the continual themes.

I think it's it's not a it's not a overnight change to fix those. What I've seen is continuous improvement over the last five years and I see that continuing just as in the 90s. I remember people building their own servers and then they built their own Linux and then they got Linux from a vendor and then suddenly there was only one vendor, you know, it's it's an evolution.

I think also that there's a fundamental problem here, which is If I talk about Cloud native, I mean in the cloud then there's an alternative which is just to use the existing cloud services. whereas if I talk about in the Enterprise then the maybe no alternative but I'm already in this kind of World of Pain where everything is unique. It's like digging down millions of layers of sediment through the years to get to the bottom of different pieces of it and very hard to integrate things and make new systems out of old ones.

So you've got two fundamental objections to going Cloud native in the public Cloud. It's like, why do I need to do this if I can do it already using virtual machines and and services from the vendor and the other is this is really hard. I don't want to make my life any harder by introducing more things.

And sorry. I think that what's happened is We've seen it push through the gaps. So new application classes of machine learning, especially I'm now seeing a lot of movement around sustainable Computing and also finops the automated management by Financial metrics of your actual stack so that you're always optimizing for cost rather than say CPU signs or something like that.

What up time and finance and if you like Green Ops are both things we can do with it get UPS because they're systems that are automating the management of a system towards a specified goal. And everybody's going to have to have sustainable Computing very soon and that's going to drive adoption too. So the point is you've got to have a business reason that's bigger than the pain point.

Now that's why for example with the public Cloud a lot of the adopters have been big digital shops that have highly skilled stuff. And in the Enterprise, it's been people like Banks who are constantly renewing their it anyway. But I think for the majority we've got to have a slightly simpler way of adopting.

Yeah, and we're seeing that incrementally. So as you look out a year from now, where are we going to be? Well, we have crossed that Chasm and Achieve that level of Simplicity and if so what ultimately will drive that.

I think in a year's time, we're all gonna be talking about how much things cost a lot. And we're going to be talking not only about money costs, but also heat cost and we're going to be worrying a lot more about where things are running and how they are how they're managed by different people based on, you know, geopolitical another non-political concerns about you know, where what the local cost of carbon is, for example, all of these things are going to be forced Upon Us by the macroeconomic climate volatile shares volatile bonds inflation Etc. I mean some of these factors May blow over a bit, but I think we've kind of in a new regime now, So I think that then you have to ask what will drive consumption in a world of low cost and I think that open source is going to be a huge Factor because people find that it's great to not have to pay for something until you've decided you actually going to use it.

That's really appealing. I think anyone who can offer very low cost services in the cloud that are useful equally do very well or people who have freemium business models which are cheap at one end and only grow expensive. If you have serious consumption, all of that is good.

So I think that will continue to see growth of the cloud. I think that we'll see a lot of developer offerings coming through which aimed at joining together the benefits of Open Source and SAS with some of these very powerful infrastructure engines, like kubernetes and gitops. We've Works will certainly be part of that second group as we bring out more, you know, simplified platforms for application developers and Lobby other things too, but I think those will be big factors and cost will be the fact that drives all of it.

Not well, there's just no way that it exists in a vacuum. Right? So all these things that are affecting our daily lives whether it's inflation or somebody having a fight with somebody else in a country far far away.

Eventually come home the roost. Hey Alexis. Thanks for being on the show.

Thank you very much. My pleasure. All right back to you guys in the studio.