Open Source Summit NA 2025 | Mukul Karnik, OpenSearch Software Foundation

Clips
News
More from Open Source Summit NA 2025

Mukul Karnik

GM, Search Services

AWS

Code, community and the future: 13 takeaways from Open Source Summit NA

Beyond uniting developers under shared ideologies, goals and pain points, Open Source Summit North America ignited conversations at the intersection of code, community and change.TheCUBE goes live with insights at Open Source Summit NA

play_circle_outline OpenSearch 3.0: Major Performance Enhancements Fueling AI and Analytics Evolution in April Launch

play_circle_outline OpenSearch 3.1 introduces enhancements, including hybrid and vector search capabilities.

play_circle_outline OpenSearch's innovation rate makes it competitive against proprietary analytics platforms.

Info
Transcript

Mukul Karnik, OpenSearch Software Foundation

Mukul Karnik

GM, Search Services AWS

In this interview from the Open Source Summit North America 2025, Mukul Karnik, Director of OpenSearch at OpenSearch Software Foundation, joins theCUBE Research’s Paul Nashawaty to explore how OpenSearch is driving next-gen innovation across search, analytics and observability. Karnik unpacks the advancements in OpenSearch 3.0 and 3.1, including hybrid search, vector database enhancements and support for agentic AI use cases. He emphasizes how OpenSearch is becoming a foundational platform for real-time insights, log analytics and anomaly detection at scale. ... Read more

explore Keep Exploring

What are the recent developments and improvements in OpenSearch leading up to the conference? add

What were the key enhancements and improvements introduced in version 3.1? add

What is the status of OpenSearch as an open-source platform, and what factors contribute to its appeal as a search and analytics solution? add

bolt Powered by CUBE AI

Mukul Karnik, OpenSearch Software Foundation

search

Paul Nashawaty

>> Welcome back to Open Source Summit North America 2025. My name is Paul Nashawaty. I'm covering the app dev space in all things open source. I'm joined today by Mukluk. How you doing?

Mukul Karnik

>> Good. How are you, Paul?

Paul Nashawaty

>> Good. Good to have you here. I want to introduce yourself to the audience and tell what you do?

Mukul Karnik

>> Yeah, so I'm Mukul Karnik and I'm the director of OpenSearch and here on behalf of the Linux Foundation to talk about OpenSearch.

Paul Nashawaty

>> Great to have you here. It really is. I mean, OpenSearch is like... It's lots new happening. 3.1 release. There's a lot of impacts around AI, a lot of impacts around the innovation around it. Why don't we start with, well, what's new about it? What's happening?

Mukul Karnik

>> That's right. As you know, OpenSearch is a powerful search and analytics platform. A lot of capabilities and in the world of GenAI. We have vector search and other capabilities as well. So a lot of things happening in the last few months leading up to this conference, of course. And in April we launched OpenSearch 3.0, which had major innovations, performance innovations. We saw about 9 to 10X improvements in performance compared to our first launch three years back. So that's a big set of improvements. But we are also seeing a lot of improvements in the new features that we've built. So as you know, the world is moving towards agentic and agentic AI, and so we launched support from oral context protocol and we launched different tools that now leverage the power of the search and analytics capabilities of OpenSearch. And so now agents can leverage that capability as well. So we've launched those capabilities. And then of course in the observability space, we've done a lot of, again, improvements and capabilities for improving the observability experience.

Paul Nashawaty

>> Yeah. No, you've been busy, right?

Mukul Karnik

>> Yes, yes.

Paul Nashawaty

>> You've been busy. So 3.0 came out, what'd you say? April, right?

Mukul Karnik

>> April.

Paul Nashawaty

>> And then now 3.1 is coming out. I mean, that's fast. That's really fast.

Mukul Karnik

>> It is. It is. And we are seeing a lot of interest in the community on some of the launches that we did in 3.0. And so 3.1 carried a lot of the things that... enhancements to those things already. Again, improving performance of hybrid search, which is kind of combination of regular search and vector search, improving the vector database capabilities in OpenSearch, adding more analytical capabilities, improving the model context protocol work that we've been doing. So really a lot of focus on continuing the innovation and the feedback that we're seeing from the community.

Paul Nashawaty

>> Well, you touched on a lot of innovation here. You touched on the agentic part of AI, of course that's top of mind for a lot of organizations and how search and agentic are coming together. It's a really powerful combination. But I want to double-click a little bit on some of things like the real time search and observability use cases at scale. Can you talk a little bit about that?

Mukul Karnik

>> Yeah. Yeah, for sure. So again, OpenSearch, if you think at the core of it is a search engine and search platform. And it lands really well for, of course, the search use cases as well as log analytics and observability use cases. And in OpenSearch you can index data in real time and get insights from data very quickly, whether you're doing search and wanting to get retrieved results quickly or whether you're doing log analytics and you want that sub-second response times. And so a lot of our work in terms of both scale and performance has been to make sure that we can do all of this in real time. And then with the evolution of AI, we are able to employ these AI tools within the context of both search and observability and even get you to more precise answers faster. Because a lot of times searching can take some time or log deep diving can take time. So you want to get to what you're looking for faster and to get through those insights, we're leveraging AI.

Paul Nashawaty

>> Well, I mean it makes a lot of sense. I mean, I think as organizations are looking at their own journeys, they're in different stages. They could be really kind of early in their stages, they can be really mature in their journeys as well. I think that when we think about this though, data is the lifeline of the business.

Mukul Karnik

>> Right.

Paul Nashawaty

>> Right? And we look at security and data integrity as a big part of top concerns for a lot of organizations. So can you talk about the enhancements that 3.1 helped strengthen the OpenSearch capabilities? And what does that mean for access control and what does it mean for encryption, auditability, all the things that organizations are thinking about when it comes to their data and how that impacts AI? Because, I mean, AI is very rapidly using that data for different things. We want to use it for the right things.

Mukul Karnik

>> Yeah, no, for sure. Again, you're right. In addition to making sure that the search and log analytics use cases work well, you want to make sure that the software is ready for enterprise use cases. And so there's a lot of, again, capabilities in terms of how you manage OpenSearch, how the cluster is performing, how access control is managed. There's a lot of things going on in there in the context of let's pick the model context protocol, making sure that the authentication works correctly, making sure all of those pieces work. Because these are new technologies coming out and you want to make sure that all of the enterprise great search that we've built over the years applies to these new emerging technologies.

Paul Nashawaty

>> Yeah. Of course, I mean that makes a lot of sense. I mean, a couple of weeks ago I did an app dev summit and I had four new research pieces come out around Day-0, Day-1, Day-2 and DevSecOps. And one of the things around Day-2 is around the observability front, as you can imagine, which is the operational side. When I see OpenSearch 3.1 and the new features for machine learning and the anomaly detection, this is an important part of the role of observability in AIOps, right? We look at AIOps and what that means for observability, I know that AI is top of mind, right?

Mukul Karnik

>> Right.

Paul Nashawaty

>> But there's also the fundamentals and the mechanics that make it... enable it to be successful. Can you touch a little bit about why this is important for the anomaly detection in machine learning, especially around AIOps and observability?

Mukul Karnik

>> Exactly. I mean, talks about GenAI and LLMs, but there are some core analytical, and capabilities like anomaly detection is critical to be able to finding just that particular spike that's resulting in an issue. And so we want to invest in both. You want a tool observability... The next generation of observability tool needs to have the right primitives like anomaly detection, other analytical capabilities, so that when you have agent capabilities, even in the observability space, they can leverage these tools to actually find what's going on in the system. So think of these tools as kind of the primitives. Previously we would've anomaly detection as the really important thing. Now it's become a primitive, it's become an important piece in what-

Paul Nashawaty

>> Table stakes almost.

Mukul Karnik

>> Exactly. Exactly.

Paul Nashawaty

>> Yeah, yeah. No, I-

Mukul Karnik

>> So as the world evolves towards using more and more agentic technology, these primitives and tools are going to be really important because the systems that have these powerful analytical capabilities are the ones that are going to be used more often in the agentic use cases.

Paul Nashawaty

>> Right, and I would even go as far as to say the ones that don't have it are going to become obsolete.

Mukul Karnik

>> I agree. I agree.

Paul Nashawaty

>> Yeah, I think that's really a factor here. So here we are, Open Source Summit and there's a lot of community growth here. Love to see the buzz around here, a lot of people walking around talking about things.

Mukul Karnik

>> Yeah.

Paul Nashawaty

>> So let's talk about that. Let's talk about how the community really has been instrumental to OpenSearch and what does that mean for the growth of OpenSearch?

Mukul Karnik

>> Yeah, that's a great question. In the early days when the project started in 2021, it was largely driven by AWS. But over the years we've had more and more partners join the project. And last year we transitioned the project to Linux Foundation. And since then we've seen even additional increase or acceleration in the community adoption and growth. I mean, this morning we are looking at a stat together, and since the last six months, we've seen about 46% increase in our active contributions compared to the previous six months. And so it just the growth-

Paul Nashawaty

>> That's amazing. Yeah, congrats on the growth.

Mukul Karnik

>> Yeah, it's really amazing to see how being part of Linux Foundation has helped fuel the momentum that we are already seeing before. And so it's exciting to see that.

Paul Nashawaty

>> But that also leads to my next question. I think that the momentum is growing. There's a lot of adoption. People are exploring new ways of doing things, but they're also bridging the gap between old and new. If I'm an OpenSearch, an open source, and I want to understand, well, there's proprietary search opportunities out there and different analytics platforms that are proprietary. What's the kind of best case scenario here to go OpenSearch 3.1, it really has a strength for those use cases for being open and community driven, as you're talking about. Why would I choose one or the other?

Mukul Karnik

>> I would say, yes, OpenSearch is open source and is a great platform. But the reason why anyone would want to pick the next generation of search analytics platform is because of the rate of innovation.

Paul Nashawaty

>> Yeah, I like that.

Mukul Karnik

>> And what we are seeing is the rate of innovation coming from so many different partners, whether it's adding agentic capabilities or whether it's, what do you say, decoupling search and indexing. There's a lot of different innovations happening in the project at a rapid pace. And all of us in the community want OpenSearch to be the industry-leading search and analytics platform. And we see that vision, and especially in the world of agentic and GenAI, we want it to be the leading agentic search and log analytics platform.

Paul Nashawaty

>> Yeah, I could totally get that. And I also think that there's kind of that bell curve of where organizations are in their own journey, right?

Mukul Karnik

>> Right. Right.

Paul Nashawaty

>> And when I look at that, if you're kind of a leading edge versus a laggard, you may not be looking at... Moving forward, I don't think with AI and with the technology stacks, organizations, especially leading edge organizations, don't have the ability to be laggards anymore. They have to be right on the edge of what's happening. And we see big changes happening. Lots of announcements here at the show, including your own, right?

Mukul Karnik

>> Right.

Paul Nashawaty

>> All these announcements, this is happening real time.

Mukul Karnik

>> It is happening.

Paul Nashawaty

>> I want to kind of segue into one of our last topics here. Think about security, right? We've talked about security, we talked about the CRA, the Cyber Resiliency Act, right? This is coming fast. Right?

Mukul Karnik

>> Right.

Paul Nashawaty

>> We see that reporting has to be done by September 2026. All applications have to be adhered to these policies by December 2027. How is OpenSearch kind of aligning to that, that CRA?

Mukul Karnik

>> I mean, look, I think as you know OpenSearch has a wide set of community members, so we are working together as a group to make sure that we are going to be compliant and we'll be... So just given the pace of innovation that we are seeing, I'm not concerned about that because we have a lot of members in the community that are involved in that. But going to your point, again, just given the pace of innovation, we do see a lot of companies adopting OpenSearch and a lot of users adopting OpenSearch.

Paul Nashawaty

>> Yeah, I agree. I mean, I think that compliance regulation and governance is always going to be a thing. It has to be a thing, right?

Mukul Karnik

>> It's going to be.

Paul Nashawaty

>> All right, so final word. What would we leave the audience with thinking about OpenSearch and what do you want them to know?

Mukul Karnik

>> Yeah, look, I mean OpenSearch is an amazing search and analytics platform, but in the world of Agentic and GenAI, we are seeing it rapidly being used for broader use cases. Of course vector database use cases, but also agentic use cases. And there's some really cool innovation coming out in the next few weeks and months and we are excited about how the project's accelerating.

Paul Nashawaty

>> Well, I can't wait to hear about those innovations. I can't wait to see what's next. This is fast-moving, as we talked about. Thank you for your time and your insights, it's been really valuable. And I want to thank the audience for watching today. I know that it's really a lot to think about and consider here. So take into consideration all the information that we have to share. I'm Paul Nashawaty. I'm coming to you live from the show floor at Open Source Summit North America 2025. Thank you for watching theCUBE, the leading source of tech news.