Toffer Winslow, StackState | AWS re:Invent 2022

This is texturong TV. Well welcome Hey weird special location today. We are at the tech strong studios here in the win in Las Vegas for reinvent 2022.

Lots of exciting things happening and we have some fantastic guests lined up speaking of which Topher Winslow who is a CEO with sexy. Welcome great and great to be here. Thank you.

Good to have you good to have you so for a few folks who may not know who stack state is or yourself tell us just a little bit about the space that you're in and we'll kind of get into what's happening. Sure. So stack States an observability company.

It's a crowded space tons of vendors lot of money has been poured into it over the last decade really in parallel with all this transition that larger Enterprises are making to go from managing their own infrastructure to becoming Cloud native driving a lot of retooling and we've taken an approach that Embraces. I think many of the things that companies care about harnessing all this data that their environments are generating. But trying to provide a different kind of explanatory context around it by infusing all that with topology the relationships between all the building blocks.

The companies have that make up these environments that they're migrating to and understanding how that topology morphs over time and then correlating that with all the other Telemetry information. So we're a company that's seven years old a lot of big Enterprise customers and financial services telecommunications managed service providers, but it's a horizontal platform to help people deal with the increasing rate of change in their environments. Yeah, you know in the startup world, we always say if you can solve your problems in the financial services or Telecom or the big, you know cable companies, whatever it is.

Yeah, that's when you know, you've really kind of reached that level where you can do some big stuff. Yeah, and that's really where we were born. We were born out of a Consulting project inside a large Dutch bank that was having persistent performance problems with one of their big customer-facing apps.

They couldn't solve it themselves. They called in a trusted Services partner who sent a couple young hot shot performance Consultants they said about Trying to gather all the data that they could and pretty quickly realized. It's not that there's not enough data.

It's just that there's no integrated data set that puts it all together and there's no sort of explanatory explanatory context wrapped around all that data and that's what they set out to build starting back in 2015. Yeah. It's interesting because you know in the apps World, we'll expect to it log aggregation getting all the data together in one place putting some nice query language and app building capabilities kind of where that all started but it seems like the both the size and scale of applications and now Cloud native Microsoft complexity of the apps that we're building the distribution across clouds.

We're doing a lot much bigger and important variable. So it's not just about getting the data together in one place, right? Yeah.

I mean the volume of data that's thrown off by all this Cloud infrastructure is massive and I think what many approaches have focused on to date is okay. How do we gather more and more of this data? How do we store it for longer and longer periods?

How do we make it more queryable? And I think those are all good things. That's that's a step in the right direction.

Argue though is that you need a secret decoder around like a Rosetta Stone to separate the signal that's in all the noise of all this Telemetry and that's where it's apology comes in. Mm-hmm. Yeah topology.

Is that context you're talking about right? I mean, would you like I don't need more data right got plenty of data. Maybe I don't know if I have the right data, but I need that context around it to tell me what's going on exactly show me what change show me how the dependencies between different components, you know, different kubernetes containers different micro Services, you know, you're increasingly virtualized infrastructure show me how those relationships are changing all the time because anytime there's a problem.

The first question that you usually gets asked in a war room is what changed. Yeah. Nothing changes you.

Yeah, that's that's almost never cheer. In fact, I mean recent surveys that are saying that the rate of change and most large-scale environments is like one change per second. Exactly.

I can't have this conversation with several people of we need to think of a world of constant change. It's a moving stream. It's not us the state that stays in one place.

One point in time because it's always anywhere in the stack across the stack environment locations all of that. So that's a different that's a different mindset much bigger problem to try to solve and tougher one it is. Yeah.

I mean when we set out to do it, we sort of thought having a graph of the topology and how that graph was changing second by second was the the way to unlock all the insights the problems that we had in our early developments was trying to do it with a commercial off the shelf graph database and what we found is that none of them had the scalability properties to just deal with the rate of change that's going on in these environments. If you've got something an environment that consists of you know thousands or hundreds of thousands even millions of components and those are changing on a second by second basis. It becomes super hard to find a graph database that will accommodate that rate of change.

So we ended up building our own just tracking down the change things. That's what the problem exactly. Yeah.

It's interesting too because I know stack state is you know, yeah Ops that's part of the equation too. And you know, there's the market and I've been working on AI since the 80s so with you know, it's gonna solve problems back and it gets overhyped but there's a lot of value that we see now through machine learning algorithms and things will get apply and targeted ways. Yep, I think that that's true.

I sort of think of AI Ops in some ways is a bit of a rorschacht test right people sort of project what they want to what want to see in it. Yeah. We we like one of the definitions that we've seen out there that says it's fundamentally about cross domain data ingestion and analytics.

It's about wrapping topology context around it. It's about doing event correlation and it's about suggesting possible remediations when when problems happen those things resonate with us and fit very well in our model interesting. So we're we're tougher we're at obviously and Las Vegas at AWS reinvent and men would I go into the first reinvent in this thing is so massive him it was big then I was surprised how big the first one was, but tell us about can your perspective about working with you know, the hyper scalar the large cloud service providers like the reinvents in the world and how you fit into that ecosystem because you know, you're getting more than a platform when you're working with the hyper scale.

There's a big ecosystem for sure. And I think the strategy for many of the hyperscalers is they want to go beyond just providing commoditized infrastructure to more and more value-added services layered on top of that and they have their own approaches for observability, you know, providing managed grafana managed Prometheus services, and there's a bunch of sessions here at reinvent about that and then they've got their whole marketplaces to let a thousand flowers bloom where you have folks like us and our many competitors who have their own offerings for sale in those. No, I think it's a smart strategy right you provide a base level of capabilities that customers can get going with that no to low cost and then have a bunch of other value added players available through their Marketplace you make sure you started in the financial services.

Have you found that you kind of go with your customers into the cloud as they move into it do they find you in the marketplace more often than not which is I would imagine it's probably the the latter story or the first first where you're going with your customers. Into the cloud. Yeah, it's mostly that what it's one of the things that we find is a precipitating event to get customers rethinking about how they manage their whole it operations is these digital Transformations and the migrations of the cloud and as they start implementing all these Technologies, you know, ci/cd pushing new code into the environment increasingly sort of virtualize and containerized environments.

It puts a lot of stress on the traditional tooling and so, you know starting with a customer while they're still in there early faces of their digital transformation and then going from on-prem to hybrid and then increasingly launching pure Cloud native Services. There's a great path for for us to get engaged with I imagine in addition, of course, you know looking and talking to new potential customers here. You have a lot of customers that are attending What would you say some of the challenges that they're fitting?

You know, they've been with you for a while. They're they're moving matured with you in terms of observability and capabilities that they're implementing. What are the next set of challenges that people are looking for I've got to figure out how to solve these problems now.

Yeah, as people are rolling out sort of kubernetes as the core of their infrastructure and what we're seeing that across all Industries. I mean, it's a big part of Telco in the whole transition to 5G financial services are doing it as they increasingly rely on kubernetes kubernetes troubleshooting is hard, you know, this all this ephemeral infrastructure and there's been a handful of companies that have been created the deal explicitly with kubernetes troubleshooting. Our view is that this topology LED approach makes a ton of sense in that environment where the containers are being spun up and spun down and the relationships and dependencies are changing at a high rate.

So that's you know, when you combine that topology map with the all the Telemetry data that both the kubernetes infr. Extra generates as well as all the associated surrounding infrastructure. That's the key to really finding out, you know, where the core problems are and differentiating whether is happening in the kubernetes layer or some of the dependent layers interesting because we introduced the complexity of kubernetes all the great problems that it solved but it brings Swift at that complexity the microservices architecture itself right now, you're talking about, you know, thousands or intensive thousands of services across applications.

Not a dozen or two in a SOA kind of service so that a complex scene just to understanding what's happening there. Yeah on top of kubernetes on top of all the cloud services. It's a complex fact.

Yeah. Well and then you layer in sort of continuous code pushes to refresh those microservices, you know, the code is changing the configuration of components are changing. The relationships between components are changing.

It's this level of complexity and high rate of change that is driving this whole rethinking of how people tool tool there and their management infrastructure the ephemeral and natur. The time nature of you know, what that condition doesn't exist anymore. That code's not that goes exactly now that we're troubleshooting her and that's why you kind of need this this sort of time machine to be able to rewind the DVR movie of your environment.

Say, okay these problems happened, you know an hour ago. What was the change that happened one hour and one second ago that started to precipitate them. Yeah that when you have that insight, you're not chasing a thousand separate alerts that are really symptoms of the underlying cause you can really drill into the underlying cause itself.

It's it's a big Market. I mean observation with hot for a while. So a lot of players in it.

They're new people coming in too, or you know kind of native observability of the folks that have got to come up the log aggregation log management. How do how do you stand out? How do you kind of show your your chops?

And in a market that's got a lot of a lot of good players in this in this area. Yeah good good companies. Well, well backed and and some you know decent open source options as well too.

Exactly. Yeah. Our view is that you've got to have be low friction in order to give people a taste of what you do, right?

So the the day of heavyweight implementations and you know having you know three months for for a rollout to really try whether something works in your environment or gone. We're really focused on fast time to Value that's focus on a set of use cases and a set of your technology stack and show you something that works within a couple of hours that works particularly well for us in AWS environments and you know kubernetes running in those environments, but then being able to show someone this map of Environment how that map is changing second by second rewinding the map so you could see what it looked like the second before a problem happened those. Yeah, those those Technical differentiators and making it easy for customers to experience those quickly.

That's I think is the key to having a high velocity business. I'm curious when you when you're in the vendor space is the technology provider you get to talk to a lot of people they're usually like the questions that you get that can tell. Okay?

No, these folks are on this edge of it, right? They're not I just beginning their Journey. They're getting to the difficult stuff.

What are some of those challenges that customers ask you about and you say, okay great. Now we can really dig in and help you with that. Yeah for a lot of us.

It's customers who have some level of Legacy infrastructure that they're fusing to the cloud. So it's a high it's a hybrid architecture and what they want to be able to do is how do I protect my investments that I've made historically, how can I gather the Telemetry data or other useful insights that are being generated by that? How do I combine that with the new data that's being generated?

ICloud environment and how do I wrap a context around all of it? So that I can really understand what's going on. I don't want to have four different tools for different slices of my infrastructure.

I want one thing that gives me holistic visibility and that can keep up with the rate of change that is putting pressure on the traditional approaches to Touring that that is that is a key thing that we hear from our prospects all the time interesting. I think something is fascinating about the observability market too as you mentioned open source, and there are tools available but it's also open Telemetry in the way that's been adopted by the technology companies not as a sort of playing the standards game, you know, like we've all watched and then we go implement the standards that's favorable to us. This is there's been some real kind of I think genuine adoption of telemetry and and open tracing and some of the Technologies coming out the London Linux Foundation.

That's I think it's pretty unique and interesting. I don't know if that'll happen again in other segments. But how does that shape your strategy?

We think it's super important. I can imagine that there's a lot of sort of incumbent vendors for whom find this threatening, you know, they've made their bones over the years by being having their agent or their way of collecting data be the standard and that having an open standard now creates higher risk of displacement for them. Yeah.

Our view is that this if the standard creates a ton of value it's going to get adopted and our job is to embrace it and help customers get more value out of that day. So we're big fans and we're sort of expanding our full support for open Telemetry and every subsequent release. Open source has always challenging us to kind of get the new model right and and customers know whether you really kind of get open source, open Telemetry and support for it is as opposed to kind of paying lip service for absolutely it's being demand.

It's being requested by customers. And at least we're seeing it in RPS and but I agree. I think I think it's highly supportive of this, you know shift left Trend right where you're getting your developers more engaged and actually doing the instrumentation of the code that they're writing and then having sort of a common language to be able to talk with the operation folks who are you know, managing this code and the overall platform that it runs on.

Yeah. If you have sort of a common Foundation of data, it's much easier for these groups of speaking to do their job together as the term devops implies. They should be well.

I think you're doing something right when new terms emerge like observability driven design and development right now. How do you build in? Just like we build in security.

You mentioned if left awful also for observability. Yeah, I think it's fantastic. It's right very interesting.

What's happening? Great. Yeah, I mean we actually think there's some interesting applications of this whole topology approach when applied to security right while our Focus has been how do you create value for it in the field of observability understanding the relationships between components and looking for relationships that maybe shouldn't exist create an interesting data stream that you can feed into various xdr platforms to help them understand what's going on in a way that they don't currently today great.

Great. Well, we're we have our studios here in Las Vegas. We're not live streaming.

So we'll be putting these videos up as soon as possible. Probably, you know, folks may watch this after the show's over. Where's the best place for people to get more information about stack State and learn about some of the things that you talked about today?

Yeah, people can find us tomorrow Wednesday and Thursday in the AWS Marketplace Pavilion. And then for those who actually aren't at AWS and can't see us in person. You can always find us on our website.

com there's free trials demos tons of videos customer reference stories. So good stuff. We should good luck at the show, and it's fantastic time.

Thank you. Thanks. Bye great.

Take care. We have some more great interviews coming back up just like toffer and stack States. So we hope you will stay tuned to with us check out the other videos.

We'll be back soon.