How Platform Engineering is Empowering DataOps and DevOps | DataOps Day

Hello. Good morning, good afternoon, good evening. Um, hope everybody is doing great.

I'm very happy to be here. Uh, presenting, uh, and talking a little bit more about how, uh, DevOps has evolved into more of a platform engineering and how platform engineering, um, is helping, um, uh, developers build and run applications, um, anywhere and at a much faster pace. Uh, I'm, my name is Vanka Trauma.

Krishnan, uh, I run products and engineering, uh, for, uh, port Works, which is a, uh, business unit, uh, of a much larger company that's pure storage. Welcome all, uh, you know, for a long time. Uh, if you, if you look beyond, uh, before an application developers, uh, we're actually building, uh, products and software and services, uh, very often they are, uh, constrained by their environment or their underlying infrastructure.

So there is a lot of friction, and from writing code to building code, to running it, testing it, and moving it to production, and then iterating on it a lot faster based on, uh, what the customers need, uh, for those applications as these developers are developing. Um, in this, the much of the friction is actually caused by, uh, waiting for infrastructure provisioning, having the infrastructure, uh, be tuned for the specific application, uh, also waiting for, uh, you know, uh, some new environment, environment to be available. And then when the application had to be deployed closer to, uh, their customers in different regions, having to port those applications and making them run.

A lot of time were spent by the application and DevOps teams, uh, uh, you know, in just getting the AppSec to be ready and updated. And then of course, when it comes to, uh, patching these applications, you know, security patches, uh, um, you know, bug fixes and all of that, that became, um, you know, a lot more cumbersome. And when, when, when developers have so much friction, obviously they cannot, uh, deliver AppSec at a pace where the customers want, uh, and, you know, on their, their experience to be improved.

Um, and why couldn't they have self-service? I mean, why couldn't developers, uh, um, you know, deploy the, uh, AppSec in infrastructure they want? Uh, how, why couldn't the infrastructure bend the way, uh, the applications want that, uh, infrastructure to be, uh, behaving in?

Why couldn't developers declaratively request for SLAs, uh, specific, uh, uh, you know, data placement, uh, network policies, uh, C P U and memory resources and IO resources, uh, based on what their application needs? Why couldn't developers, uh, uh, you know, look at how their application's performing and tweak the underlying infrastructure without having to go through an extensive reconfiguration process declaratively and ask for, uh, better, uh, ss l a from their underlying infrastructure? If developers at all of this, if the entire world around them programmatically responds to them, they'll absolutely be able to push, uh, uh, the AppSec to kind of go to new heights and better deliver customer experience.

But if you really look at it, the, there is not a lot of self-services available in the past. Now, incoming is the platform engineering team, or the DevOps team, right? Which supports, uh, uh, uh, you know, uh, uh, thousands of developers, uh, in, in bringing, uh, in, you know, making the infrastructure, self-service provision by the developers, respond to developers' needs.

But most of the times, these platform teams are overwhelmed, supporting hundreds of sometimes even thousands of developers, sometimes in our customer environments. I see the platform teams are about 10 or 15 people. They're supporting an application development team that is thousands of people right now.

Are they overwhelmed? Absolutely, because, you know, providing self-service making, enabling one team to be self-service means someone else has to do a lot of underlying work, uh, to enable self-service. And that can be overwhelming when you look at a lot of applications, a lot of, uh, engineers working on those applications and enabling, uh, them to be running at a faster pace.

But, you know, so that can you scale from dev to test to production, uh, easily, then, you know, what is the best way to enable application teams to build and run and, uh, uh, ship applications faster? And also enable platform teams to actually have a better quality of life supporting, uh, all these developers, uh, supporting all the applications. Uh, you know, if you, you know, if you look at the customer environments, we have many customers who are building, um, simple, uh, applications, mobile applications or complex, uh, enterprise web applications.

And many of these times, a lot of these applications come together in the, uh, and then they have to work together to deliver a superior customer experience for their internal and external customers. Now, with so many dependencies and then so many environments to support, uh, it can become extremely overwhelming for the application teams and the platform teams and the DevOps teams to, uh, kind of keep it all together and deliver the superior customer experience they strive to deliver. But can, you know, as a team, how do you scale from dev to test a production?

How do you do it? Like, how do you test, build your app, test it in an environment that is similar to your customers, and then at the same time deploy the same code, uh, promote it to your staging or production, and then get the best, uh, the, the, the same experience as well as do less, spend less time and resources, uh, porting it over or running it in different environments? And how do you do that?

Can you, can you really scale that? Uh, is is one of the questions we see, uh, folks asking constantly to continue to reduce, uh, the friction of building AppSec to running them in production. So this is where the platform engineering discipline comes, uh, you know, really handy, right?

What we have seen is that overall, what, what constituted as DevOps teams that sat in either application teams or in operation teams are now coming together as a platform engineering organization and a platform engineering discipline and practice where, uh, a lot of the, uh, needs for providing, uh, infrastructure that can essentially bend to the application needs, but at the same time deliver, uh, uh, like self-service, but deliver that with, uh, you know, uh, that meets the company, uh, requirements or regulatory requirements or compliance requirements, uh, and put the sufficient guardrails around what images get deployed, what kind of application code detect gets deployed, what kind of security scans are run on those, uh, code, and how can, uh, uh, uh, application code be promoted from devastating to production. And bringing all of that into a self-service automated system, uh, that can essentially drive, deliver the superior customer experience at a much faster base, but also develop support, uh, uh, thousands of developers, um, uh, at the same time. And, you know, you know, deliver the, unleash the productivity of the application teams like, uh, like never before.

So here is where I think the platform engineering team is becoming more and more critical for modern enterprises or even traditional enterprises that are modernizing their, uh, businesses. And we have seen, especially in the Kubernetes space, in the container space, uh, ton of adoption around, uh, uh, around containers and Kubernetes to enable these platform engineering teams to deliver self-service developer platforms. Let me pause here a little bit.

There's a little bit of a, you know, logistics. I need to check the other tab to see if there are any messages. Uh, are there any questions so far?

No. Alright, so let's continue, um, uh, with the, uh, with the, uh, with the presentation here. Um, and so when we talk about this platform engineering, what is, what enables the platform engineering?

What enables self-service? What enables, uh, uh, application teams to build anywhere and run anywhere? Right?

I mean, there needs to be some underlying, uh, uh, fundamental technology that powers that. And those, that technology is containers, right? I mean, containers have been around for a long time, but over the last, uh, few years, uh, with Docker and Kubernetes, containers have become mainstream, right?

You can essentially package, uh, an application and run it in your laptop, and you can run the same app, the same code on a server with very little porting effort or very little, uh, uh, application porting effort. So that is the cloud native approach, essentially, where, you know, you could, uh, run your application, uh, in any cloud or data center infrastructure. Now, what this enables is, uh, it enables continuous app development.

It enables developers to, uh, build and run and continue to iterate on their code, uh, and push it into different environments, uh, that they can, uh, actively test against. It. Also, as we discussed, uh, runs, uh, lets you run AppSec anywhere.

Uh, that means you can, you know, build and run the AppSec anywhere and with very little porting and managing of the dependencies, um, of course it makes it a hundred percent portable. We have customers who run applications in different infrastructure today. Uh, we see users running applications, uh, that are in one cloud.

They also can replicate the same application stack in another cloud or in the data center or in an edge, uh, environment, uh, you know, and continue to manage and run their applications with a unified interface like Kubernetes, it delivers kind of a lot of powerful, um, uh, you know, a lot of power at the hands of a developer and a platform engineer to orchestrate and run these applications. Anyway, uh, now Kubernetes, and we, of course, we touched containers. It's important to talk about Kubernetes as well.

Kubernetes is the underlying scheduling technology, which enables essentially instructs the infrastructure to place application code, uh, in the right size C P U or the memory, uh, you know, where it can run, uh, and get the most performance and continues to balance, uh, the resource across, uh, uh, a, you know, a fleet of, uh, um, computer and storage nodes to ensure that the applications, uh, are running au autonomously or automatically, plus they're getting the performance, uh, and the resources they need to run. Uh, and, you know, when this cloud native approach also, you know, kind of enables us to unify kinda APAs between, uh, two constructs that that existed. One is virtual machines and another is, uh, containers, which is, you know, what you were discussed.

And, you know, you kind of see all of this cloud native approach unifies, uh, management of VMs and containers today with the latest advancements in technology of like Bert. Uh, so you see that a lot, uh, as enabling technology for platform engineering. Uh, and of course, um, there is no serious business app without data, right?

So you need to have, you know, this cloud native approach, uh, the, the, uh, whole platform engineering and the self-service platform engineering delivers is enabled in, in, now in modern enterprises through global data fabrics, right? That, uh, that like, you know, enables, uh, the portability between data, uh, build, you know, data can be born anywhere, but can be si you know, can be moved and migrated and cycled through in other infrastructure. But this in a secure, managed, in compliance, in a regulatory compliance way, uh, you know, and all of that enables, uh, you know, you need a global data fabric to completely enable the cloud native approach.

Um, so essentially these kind of approaches, uh, is what constitutes and enables platform engineering to deliver the self-service to the application developers completely, or, you know, to a large extent, eliminating the friction in building applications and delivering applications and continuing to run them, uh, in production. Now, where does Port Works come in? This, uh, port Works is a complete data platform that delivers a whole bunch of AI and application serv, uh, and, uh, and storage services, uh, and data services, uh, to, uh, containers and Kubernetes, uh, in production, of course, in, uh, Devon staging as well.

So what do we mean by a data platform? Uh, you know, uh, when it, when it comes to, uh, modern AppSec and when it comes to how those AppSec are, uh, built and run, um, it is very important that the, uh, you know, when you build an application, unless it's a very purpose built, focused on, um, one particular, uh, you know, purpose, most of the modern AppSec, you know, um, deal with a multitude of users and use cases, and, uh, and most of the times the AppSec need to deliver, uh, the best customer experience in order for the usage to increase, in order for the customers to benefit from the existence of the app. How many times do we actually open and use an app?

If it is extremely slow to use, we never open it. How many times do we go to a website that is, uh, really slow, takes a long time to open pages and, and actually consume the content? We rarely revisit that website.

So that is the customer experience, that's the user experience applications needs to deliver. So when you develop these modern AppSec, in order to deliver that customer and the user experience, uh, that these, the, the consumers of the application desire and look for, uh, you know, what you need is not just a just a storage solution or a, or a data protection solution. You need a data platform, uh, because, um, the, it is important to take the data closer to the applications in order for application to perform very well, in order for it to deliver the best experience.

Because if you really look at it most of the time, uh, uh, when the application's run somewhere and the data is disaggregated and running somewhere else, because of course for the infrastructure engineers, it's easy to manage that, you know, by separation, you have your server teams, you have your storage teams, you have your network teams, and we have logically separated the responsibility, but the user experience suffers. So when we bring the services closer to the applications, the user experience excels and delivers the best user experience. But of course, you know, you now need, you know, a, a a better way to manage these, uh, these applications and better way to manage the infrastructure.

So, and point products don't offer, uh, a, uh, you know, complete, uh, solution for the problem. Only a platform can deliver that complete solution, can deliver the appropriate level of performance that is closer to the application, but can also secure the data, protect the data, uh, per app, and enable, uh, migration of the data if needed from one environment to the another to deliver the best customer's experience, to provide business continuity to, uh, and, and disaster recovery. And of course, to, you know, uh, integrate with, uh, different data services and, and overall wrap them all around, uh, uh, with AI and AI ready, uh, capabilities where there is less involvement, uh, you know, from and manual intervention needed, you know, but you know, more of the, uh, tasks are done autonomously at scale.

So, uh, the platform teams can, um, deliver these kind of the, uh, SLAs to these applications at the same time, maintain their sanity, uh, ensuring that these applications are running, uh, at scale in any infrastructure that the application teams desire. So that is why, uh, a platform is critical, uh, because it enables and encapsulates all the capability per application on any infrastructure. Uh, so the platform teams don't have to be the integrators of different technology, uh, different solutions to offer these, uh, for their application teams, and that's what Port Works delivers today.

Port Works delivers storage capabilities, backup and disaster recovery capabilities, uh, data services, you know, we can orchestrate, uh, different kinds of databases, you know, MySQL or, uh, Postgres, Cassandra e sql, and no SQL messaging frameworks like Kafka and RabbitMQ, and, uh, you know, of course, uh, ML AI frameworks, uh, too, to kind of enable data scientists and everybody to, uh, you know, leverage, uh, uh, AI ML frameworks to kind of build their machine learning and data science capabilities. And, of course, you know, uh, provide, uh, DevOps services as well to enable, uh, DevOps teams and platform teams to work and deliver these services to their application teams, and so that they can seamlessly build and, uh, deliver their applications on the infrastructure. Let me pause here a little bit and check for any questions.

None so far. I'll go back to my slides. And so that is, you know, at a nutshell of what Port Works delivers, what do you get with this platform?

You know, of course, there's a whole suite of capabilities that you get with the platform and Port Works, um, to highlight a few. Um, you get, you know, automated capa, automated capacity management. Um, you know, when, when you're running applications and especially these applications are storing data in a large scale, um, you know, capacity management becomes a critical piece because, um, you know, your, your, your application can only, uh, function as fast as your underlying storage, uh, can receive and store the data.

Plus your storage also has to grow along with your application's data growth. And how do you do it in a policy driven, uh, way, uh, you know, as how, you know, like portex data platform delivers, uh, you know, provides enterprise class disaster recovery. You cannot mess with customer's data if you don't have, uh, a proper, uh, way to recover from infrastructure outages or networking outages, or even software outages.

So providing enterprise, delivering enterprise class data, disaster recovery is key capability, uh, that Port Works delivers to, uh, modern, modern enterprises building these modern applications on containers and communes. Uh, so broad, no, you know, port Works delivers a broad catalog of data services. We discussed, um, you know, before, like, you know, SQL and no SQL kind of databases and messaging frameworks and M L AI frameworks, um, and, you know, enabling application and data mobility across different clouds through can resource optimizations at scale, where, you know, you could run, uh, your application clusters, thousands of nodes, port Works can monitor the cluster, rebalance the application data load, uh, and enable, uh, these applications to run with, uh, you know, less IO contention and less resource contention.

So they can enable kind of multi-tenancy and host different kinds of class of applications in the same infrastructure and run them in an autonomous automated manner. Plus, you know, of course, you know, we, we have built all of this and we've built the APIs and the underlying infrastructure, and now we're powering them through AI for automation, you know, port access, portex autopilot capabilities, uh, you know, kind of like, that's kind of the, uh, uh, early stages of AI ops for, uh, Kubernetes and platform engineering as well. With that, we are at the conclusion of this presentation.

Um, I can, um, we can switch over to any questions from the audience. Thanks for dialing in, and, uh, if, uh, anybody wants to continue the discussion, uh, please, uh, feel free to reach out to me at Kat Christian on Twitter. com.

Um, there's always someone willing, you know, available to talk and answer questions and take, you know, we are happy to be part of your, uh, uh, application modernization journey. Uh, and let's go build AppSec together and make them, uh, deliver the best customer experience. Thank you.