Data-Swirling in Kubernetes – Nuri Golan, Sosivio

This is Textron TV. Of the great pleasure being joined by nirigoan Nuri is CEO and co-bounder of socibo. Welcome Neri.

Hey, how are you today? Did very good even better? We're gonna talk about kubernetes great topics.

So excited to talk with you. Well first introduce yourself a bit about you. Tell us a little bit about sicibio.

Sure, so, I'm very Golan. I'm the CEO and co-founder of Cecilio. We're predictive troubleshooting platform for kubernetes.

So basically we're keeping kubernetes environments and applications that are running on them running smoothly. We automate a lot of the troubleshooting process using a new approach which we call data swirling. We can talk a lot about that today, but we're helping devops teams, you know, optimize their their workloads and functional a lot better.

Excellent. Well, let's start out by I think it's you know, everyone says kubernetes is a complex environment. I don't think there's too much debate about that or it can get complex fast.

What are some of the challenges people have we're trying to manage and operate kubernetes environments at scale. I think that you know kubernetes does a lot of a great things and and the reason people want to use it is because it automates a lot of the normal processes that they have to go through when they're running anything in the cloud or on-prem, but they're you know with any automation that you use there's gonna be layers of abstraction and ultimately, you know, the expertise that's required to really keep everything working smoothly tends to be harder to come by as you rely more and more on that automation. So so we're seeing the same thing in kubernetes as it's become more of the standard container orchestrator out there.

There's there's a difficulty keeping these environments running smoothly especially as they grow and complexity in size. There's a ton of information available a ton of signals data that you can you can use to to troubleshoot these problems, but you know too much data is certainly becoming a problem. Well, like any technology you learn over time you get comfortable whether you get experienced with these sort of build up on knowledge.

That isn't necessarily help you when you're scaling very rapidly deploying more applications under kubernetes. It's not like the technology waits for you to catch up. Right?

It's it's gonna in the customers and the people you do using your software are going to do what they're going to do. So the oftentimes that's the challenge is like, how do we get up to speed fast enough? Why do I deal with all the messages alerts logs information data that I'm getting From kubernetes and know what to do with it.

So aside from having a better way of pulling it all together, right? We certainly need to wait and look at that information talk about data swirling and why that's approach why that approach helps what we when we got started with the company. We thought that there was really a gap in in the tools that can be used to automate a lot of the troubleshooting process or just you know, speed up the time it takes to fix environments or fix applications or anything running in kubernetes when when they do go down so data swirling is specifically built because and maybe I can quickly talk about what it is, but it's essentially a tightly coupled system using machine learning and data collection which runs on kubernetes environments on the edge.

So it runs inside a customers environment the entire system is built Cloud native. It's proprietary to us. So it includes the data collection the analysis insights correlation.

All of that is done locally in the customers environment and the reason And we did that is because there's just as I was saying earlier. The massive amounts, you know the velocity volume and variety of the data that are available when you start to offload that information elsewhere, you know at some point it does become cost prohibitive or security prohibitive where there's many many customers in a variety of different secure industries that are not comfortable with their information being pulled elsewhere aside from all the gdpr, you know, all the regulations around personal information. So there's a whole sector of you know, many Industries which don't have tools that are usable today or at least that the the barrier and the hurdles to use.

These tools are extremely high. And so data swirling was a you know Edge approach to analyzing and deriving insights from the data that can be in completely disconnected and air gap environments. And aside from the technology advantage that it's much more resource-friendly again, we're not taking anything off the environment.

So we only keep the bits of information that we need and we do all the analysis on the edge but there's also security advantage in that again. We're not taking any information elsewhere and so it not only allows us to work in many different types of environments but also with many different types of customers that have been slightly underserved by the tools that are out there today. I'm curious too often CEOs entrepreneurs start companies because they lived with the problem that they're solving in the kinds of Technology.

How did you come across this approach to helping with kubernetes? So I've kind of interesting story so our CTO and co-founder LeRon Cohen, he was at red hat and he was a principal kubernetes architect in Europe and Israel, you know, he was also a professional service provider. So he was the guy that was called into many of the larger organizations using openshift and and when kubernetes when they were facing kubernetes problems, he was the one that would get pulled into the War Room to really, you know, help them figure out what was going on.

And so he saw firsthand how many of these large organizations specifically were struggling with kubernetes adoption. And so it became you know passionate project of his to try and automate some of his life and the work that he was doing as a professional service provider. I met him several years ago in my previous life.

I was meeting the corporate Venture on for Leo corporation, which was the company that acquired my first startup. And at Leah I had actually been working on some other Edge compute application specifically with e-commerce Platforms in the vehicle. There.

There was also a problem with data the amount of data coming from all the sensors in the car that became was becoming cross-prohibited and we needed Edge compute and Edge analysis to kind of figure out what data does need to go back what analysis can be done on the vehicle itself. So it was a it was a really nice merge where the same approach in the same kind of problem in a bit of a different industry and we really connected and decided to start this company and push it Forward along with two other co-founders who one was a friend of Leon's and one was a co-founder of mine from a previous company interesting interesting crossover to connection between sort of the edge analysis compute from the environment you came in and you know kubernetes experience that Lauren had Well, I think we're seeing it across, you know, many different Industries, especially as AI becomes, you know, the it's a very buzzword but as Technologies adopted by more and more, you know Legacy Industries. There is a lot of common overlap between the tools or at least the approaches that are being used by these companies.

So it wasn't a big leap to go from Automotive to Enterprise Cloud software because the fundamentals are pretty much the same and yet operated scale right both do yep. Very interesting. So talk a little bit about wherein people's adoption curve of kubernetes.

Do they first encounter this problem? Are they you sort of overwhelmed in the beginning because there's a lot of new things to learn in your best to kind of set this up from the start. You have to get to a point where you've got enough data here where you can really take advantage.

Of you know, what a technology like years can do for them. Yeah, there's yeah, we we look at it as like day zero day one day two problems. We started building it as more of a date to approach.

You know, somebody a company has already moved their workloads into kubernetes and they're looking for help and managing those environments better. But we've also realized over time that day Zero ahead of the adoption or as they're migrating to kubernetes. There's a lot of advantage to using a tool like ours because they get set up to work the right way from the beginning and there's also a big fear barrier for many of these companies who are who are scared to adopt kubernetes because of the challenges that are well known so having a tool that that alleviates some of that concern knowing they won't run into many of the troubleshooting issues after the fact has helped some of our customers in that curve in that adoption curve.

Well now do you do this through a cloud approach, you know a cloud offering you do it through libraries of software on-prem which are the models of how people work with you. So it's it's agnostic. We have you can get us off the Amazon Marketplace and we also have a an on-prem offering as well.

So that's one of the major advantages of data swirling that I mentioned earlier because it can work on-prem or in disconnected or air gap environments. Companies large financial institutions Banks insurance companies security companies. Some of these are our customers are able to use our product without the concern of where is the information?

Where's the data going? So can work on Prem it can work in the cloud. We focused more in the early stages of our company on the larger Industries the more disconnected environments because there's we have more Advantage there and there's there's less less or fewer Solutions in the market for for those types of environments.

So definitely like Financial Services Insurance. Those have been our Prime markets early on Yeah, definitely. Yeah being able to operate in print on premise super important for some of those larger and you know security sensitive Enterprises.

And surprisingly still a very large part maybe not so surprisingly but a very large part of the market even for companies, you know moving to full containerization and using kubernetes many of them still run a lot on Prem. There's a lot of stuff on Prem as much as the cloud is, you know, taking taking hold taking. Yeah a foothold and very popular what are some of the things I'm curious so I'm running a kubernetes environment.

Maybe I've been doing this for a little while a year plus two years something like that and I plug in your technology to what I'm doing. What are the typical things that people find pretty quickly because they're using your stuff. Yeah, usually they're you know, there's always we've had this with a number of customers where they install the product and there was some nagging issue that they had for for weeks and configure it out whether it was a repeated slowness on their website or a specific application that was was crashing constantly and they couldn't figure it out and there's so many different reasons why those things can happen and there's there's an endless list of the types of problems you can encounter but a lot of it is due to the fact that many of the tools they were using beforehand don't really see the the true root cause the initial, you know, Nexus of the problem what they'll normally see with their in-house observability stuff or some of the older tools is the first symptom of the problem.

So okay a pod is repeatedly crashing or you're seeing slowness on a website. But where is that coming from? Is it from specific problem with the applications resource allocation?

Does it have to do with how it's interacting with with other containers on a specific note or server? These are the things that we can see because we're on the edge and we're Rising everything in real time we can see everything that's happening and we have no constraint as to offload, you know, so many of the tools today will offload those resources and they can't excuse me offload that data and they can't take everything because it's it's resource prohibitive. We're able to see everything happening in real time.

Excellent. What are the kind of resources this takes? To operate are we talking about something significant in terms of load on systems, you know libraries that it has to go into code.

What are some of the requirements to be implement? Our mantras to be non-intrusive so very low it's built Cloud native. Everything is built as a set of microservices.

It's deployed through a demon set. So there's nothing permanently installed on any of the nodes. We essentially have a data collector on every node.

And then our our AI engines or the analyzers collectors. Those are microservices with scale up and down with the environment. And so it's it's negligible resource consumption.

You can use socio to see the resource consumption of all of our microservices and our our customers again are these are large financial institutions insurance companies who are very cognizant of what they're putting on their environments and we haven't had any issues to date with with installation. It sounds like you don't have to do anything to modify or include extra code and my containers that can do that in the Clusters notes that I'm setting up. Yeah, no broken artifacts.

No, no code injection. The only thing that you would have to do if you'd like is give elevated permissions to make changes. So we do offer recommendations.

They're through the cube CTL action. And if you'd like to do that, then we would need the permissions to do so, but that's the only only thing and if you don't want to use that you don't have to give it. Okay, I would think that also helps with stability code right?

You're not having to Inject your be part of people's code that they're putting in their in containers or in their own code builds. right Excellent. Well talk about so how do people get started with you you offer free accounts access to the technology to try for a while.

How do you work with folks? Yes a standard the product comes with a one month free trial for our premium version. So there's a community version and a premium version.

io very easy to install takes a couple minutes. If you'd like us to be there with you you can reach out and contact us. Well happy to be there but very simple to to do so and as I mentioned it comes with a one month free trial at the premium version so all of the insights recommendations that analytics level that is part of the premium version, but our community version looks very much like a typical observability tool with very granular metrics and and plenty of data and insights that you can use.

So we do encourage we have plenty of users using it for free we like that we get feedback from them. So even if you just want a good observability tool we encourage you to download it and use it and get the premium version if you'd like as well. But yeah, okay, perfect.

Perfect. io. Correct?

Correct. Good. Excellent.

Well, it's been fantastic talking with you Neri, and it's a great space to be in right folks are looking for help and you know improve their environments and manageability of it and insights what's going on what they need to do to make it better or solve those nagging issues that have been hanging around for a while. We can't figure out right. Yep, good.

Well, thanks for joining us and I hope folks will check out free account. io socio. All right.

Thank you, sir. We'll talk to you again soon cash. Yeah, thanks for having me Mitch, and I hope balance doing well.

Thank you. Appreciate it.