CloudDevelopersDevOpsLet's Talk

CloudNatix Helps Solve Planet-Scale Cluster Management

0

Guest: Somik Behera
Company: CloudNatix
Show: Let’s Talk

CloudNatix is a service platform that is currently launching. Their goal is to bring a planet-scale cluster manager to market that is delivered as a service across Kubernetes clusters, virtual machines, cloud accounts, and on-prem data centers to unify them as an aggregate cluster across the globe. It’s all about driving up human and machine efficiency.

Helping Companies with Agility and Cost in the Cloud

According to Somik Behera, Founding Head of Products at CloudNatix, the company uses a technology called Autopilot (which Google pioneered years ago). Founder and CEO, Rohit Seth invented containers at Google in 2007 and predicted some of the issues enterprises would face. Fast forward to today, and CloudNatix believes every company has infrastructure that spans multiple clusters and multiple clouds. Those companies are grappling with agility and costs. Behera believes the CloudNatix platform can help them with those issues.

How does CloudNatix manage Kubernetes at scale? According to Behera, “We believe the next evolution of this architecture is to build multi-cloud-native applications, which means the study of applications that will span clusters at the minimum and clouds on the other periphery.” He continues, “As you go to this more heterogeneous, complex environment, there are some key elements of hyper-scale companies that are becoming the problems of every enterprise out there. Namely, the cost goes through the roof and the complexity of manageability of multiple clusters is a lot harder.”

The CloudNatix platform is capable of abstracting clusters of machines, even clouds, and aggregating and stitching them all together as a single whole to create a giant warehouse-scale cluster that works at planet-scale and uses an enterprise’s existing cloud accounts or data center machines. The company also leverages machine learning and their own Autopilot technology to, “continuously optimize capacity of containerized workloads, as well as virtual machine workloads while automating its placement, as well as connectivity and security, so that the amount of cloud waste a company has is reduced,” adds Behera.

But what happens when a customer is at a different stage of their journey? With those that are new to adopting cloud technology, CloudNatix helps with insights and the ability to leverage the power of machines to recommend what the customer can do to increase efficiency on utilization and some levels of automation to increase availability.

For more advanced customers who are adopting Kubernetes, Behera says, “they’re looking at our core Autopilot technology, not to understand using insights, but actually act on those insights and then automate that life cycle so that your Kubernetes deployments, virtual machine deployments, your pods, your stateful sets are continuously optimized, reducing your cost.” Should CloudNatix detect there’s going to be an outage, it can ensure that you avoid an outage. Behera adds that it “doesn’t matter how advanced a customer you are, we can insert in your environment and take you through the journey to become a cloud-native and a multi-cloud-native enterprise and do it faster than any other way.”

CloudNatix sets itself apart by not only doing observability but cluster management as well. Behera talks about CloudNatix’s unique offering, stating “We help you operate and understand that visibility with insights, and then act on that insight.” Finally, Behera adds, “We do something unique that nobody in the market does, which is Autopilot. It not only acts once but acts continuously, optimizes continuously, and operates your global infrastructure continuously.”

The summary of the show is written by Jack Wallen


Here is the rough, unedited transcript of the show…

Swapnil Bhartiya: Swapnil Bhartiya here and welcome to TFiR Let’s talk about cloud-native Computing. It’s a special edition of Let’s Talk for KubeCon. And today we have with us, Somik Behera, Founding Head of Products at CloudNatix. Somik, it’s great to have you on the show.

Somik Behera: Likewise, Swap.

Swapnil Bhartiya: This is the first time I have met someone from CloudNatix. So I need to learn a bit about the company. Tell me about the company, your partners, and what problems are you trying to solve for the cloud-native community?

Somik Behera: CloudNatix is a software, is a service platform that is launching right now. And we are trying to effectively bring a planet scale cluster manager to market that is delivered as a service that works across Kubernetes clusters, virtual machines, cloud accounts, on-prem data centers, and unifies them to a whole unified aggregated cluster across the globe. Driving up human efficiency, as well as machine efficiency. Using a technology called Autopilot that Google pioneered years ago. We started CloudNatix because our founder and CEO Rohit Seth, who invented containers at Google in 2007. He saw some of the problems that enterprises are seeing today back at Google in 2007. Namely, the reason containers were funded at Google, the reason it was invented was because Google’s entire fleet, when he joined Google in 2006, was running under 20% utilization. And, their environment looks much like today’s enterprise environment, where they had multiple pools of compute.

They had Borg, which are [inaudible 00:01:48] backyard, and then they had the new businesses coming online. And so that creates heterogeneous pools of infrastructure. Come 2007, the recession hits. Cost becomes a number one priority. Google is a public company. A new CFO comes online and this is front and center from a cost reduction and a business agility perspective.

Containers were the solution back then, but there were a lot of pieces to containers that were added at Google. Namely, Borg had to schedule containers and address these new Google services like Gmail at planet scale. And they had to invent a key technology called Autopilot and that’s what helped Google achieve 80%, 90% utilization across millions of course.

Fast forward to 2020. We believe this is what every company needs is working with. And every company now has infrastructure that spans multiple clusters at the minimum and multiple clouds on the other side. And they’re grappling with agility as well as cost challenges. And we believe the same problem that Google had 2007 is what enterprises have today. And we can help them solve it with CloudNatix platform.

Swapnil Bhartiya: These days, we talked a lot about challenges of managing Kubernetes at a scale. I want to hear from you quickly, first of all, what are the challenges that users, customers face and how are you approaching it?

Somik Behera: So the top couple of, there are few challenges, as customers have adopted cloud and gone into a cloud first architecture and increasingly a cloud-native architecture, we believe, and we see the next evolution of this architecture is to build multi-cloud-native applications, which means study of applications, which will span clusters at the minimum and clouds on the other periphery. As you go to this more heterogeneous complex environment, there are some key elements that were in the realm of hyper scale companies are becoming the problems of every enterprise out there. Namely, cost goes through the roof. Complexity of manageability of multiple clusters is a lot, lot harder. And therefore the business suffers, it suffers via, explosion of class or reduced availability of services that your SLO suffered different applications.

Swapnil Bhartiya: You’ve mentioned Autopilot already. If I may also ask, you explain how you are looking at the problem, not just look at the tech stack that you are building or that’s the major announcement here at cube as well. So let’s talk about the announcement and then we’ll talk about the core companies that makes up this service.

Somik Behera: Of course. So the cloud Datix platform can abstract clusters of machines, even clouds, but aggregating and stitching together all of them together as a single whole and create a giant warehouse scale cluster if you will, that works at planet scale, using enterprises, existing cloud accounts or data center machines. We just stitch it together to be one uniform. Once we do that, we leverage machine learning and our Autopilot technology to continuously optimize capacity of containerized workloads, as well as virtual machine workloads while automating its placement, as well as connectivity and security, so that the amount of cloud waste a company has is reduced. Outages are reduced because we remediate any [inaudible 00:00:05:45] issues that comes up during the course of a business. And that’s what cloud networks brings to customers, reducing their costs and reducing human toll, increasing efficiency, increasing application delivery agility, and without having to rip and replace anything. We can work with that existing virtual machines, existing Kubernetes clusters, or we can even kind of manage and bring up a new Kubernetes clusters and, manage it from day zero on books. And we are designed to meet enterprises where they are at,

Swapnil Bhartiya: When you say that you meet enterprises, where they are. So their enterprises are in the different stages of their cloud-native or digital transformation journey. Some are foreign ones and already have Kubernetes like things in production and some still have legacy environment. So how do you meet them in where they are in the journey? How do you help them with irrespective of what kind of environment they have?

Somik Behera: Very good question Swap, so, we, our customer base, where today, it stands from customers who are in early cloud adoption, there are few folks who are [inaudible 00:06:47] automating. And then there are customers who are advanced global customers like viewers to tuning into CubeCon edition. And they’re using multi cloud-native or cloud-native architectures. For a cloud customer who is adopting cloud today our [inaudible 00:07:05] ability to have global visibility and intelligence cost and operational intelligence is key. For customers who have adopted and, cloud they’re looking for automation. So that’s where we bring in CloudNatix, insights, ability to leverage the power of machines, to recommend what the customer can do to increase efficiency on utilization, as well as kind of automating and increasing available.

And then finally, the more advanced customers who are adopting Kubernetes, they’re looking at our core Autopilot technology, not to understand using insights, but actually act on those insights and then automate that life cycle so that your Kubernetes deployment. So your virtual machine deployments, your pods, your deployment, your stateful sets are continuously optimized, reducing your cost. And then if, if it detects, there’s going to be an outage, it can size it up, back up so that you avoid an outage. And, and that’s how doesn’t matter how advanced a customer you are, we can insert in your environment and take you to the journey to become a cloud-native and a multi-cloud-native enterprise, and do it faster than any other way.

Swapnil Bhartiya: Some of the challenges that you mentioned have been addressed. There are a lot of players in this cloud. Native is a very, very busy space. How do you set yourself apart from them? You did mention how you are approaching it differently, but if you can talk about what kind of folks you’re competing against and what edge you have over them,

Somik Behera: Of course. From a market perspective, there are companies which focus on observability and that’s great, we work with them. But today there is no nobody who no observably company, that that is a cluster manager. Then there are companies which, focus on cloud-native architectures, Kubernetes deployment, and just bring up of an infrastructure. We believe the cloud providers themselves like EKS AKS are doing a great job at handling that complexity. Then there are cost optimizers for companies out there who, just kind of automate, give you insights on cost management, cloud health, et cetera. Out there may give you some recommendation. We are, if you will, an intersection of all of these players. We provide you unified visibility. We work with observability platforms. We help you operate, understand that visibility with insights, and then act on that insight, act on those insights because we are a cluster manager ourselves.

And finally, we do something unique that nobody in the market does is Autopilot. It not only act once, but act continuously and optimize continuously and operate your global infrastructure continuously. So to net it out, there are a lot of players and enterprise can go in and start to stitch it together. It is a very complex problem. We have seen it firsthand at a Hyperscalers and they had a tough time doing it. [inaudible 00:10:17] what’s called code zero. It’s like code red, while they had to adopt this technology of Autopilot, [inaudible 00:10:24] Linux containers. We believe [inaudible 00:10:27] enterprise can build it, this, this kind of architecture themselves by adopting multiple things. It’s going to be a lot easier having a unified platform, which can help you observe and understand act and automate your multi-cloud environment and your multi cluster environment.

Swapnil Bhartiya: Who is your target audience?. And in most cases, what happened is when you do announce these kind of projects, there is already a user based. People are already using it in production. So can you talk about first of all, target audience, and if you can share some use cases, of course you can or cannot name the company they want to, but I am really interested in the use cases and the target audience ,

Somik Behera: Our target audience are enterprise customers who are struggling with scaling a cloud, their cloud-native and cloud initiatives. They have more than one cloud environment, or at least more than one Kubernetes cluster, so that’s key. And the, the DevOps engineers, the DevOps managers, the VPs and CTOs are, are, are how we enter into customer account school. We are the first conversation. The first conversation starts off with the cost problem. The more we deep dig deep into it, it turns out to be complexity problems. And then finally, most customers don’t have time to adopt a solution, like we talked about observability vendors, we talked about different Kubernetes distribution, they just don’t have time. So, , so we have created a simplified SAS solution. You click a button, takes five minutes to adopt your cluster. We go in and we become your virtual operations team that stitches this thing together and drives up efficiency, priorities increases available and get to get to going, , on your business.

Swapnil Bhartiya: Somik, thank you so much for taking time out today and talk about CloudNatix and I look forward to our next conversation. Thank you.

Somik Behera: Thanks Swap. I appreciate it. Bye.

Read Transcript
Don't miss out great stories, subscribe to our newsletter.

Login/Sign up