Cloud Native Theater | Data on Kubernetes Day: Opening Remarks - Edith Puclla, Percona

This talk introduces the Data on Kubernetes Day, where Idet Bugja serves as the host and highlights significant advancements and findings from the Data on Kubernetes community. The speaker discusses the recently released 2025 report, revealing that nearly half of organizations run at least 50% of their data workflows on Kubernetes, emphasizing its role as mission-critical infrastructure. Key insights include the rising adoption of AI and machine learning workloads, with 70% of organizations recognizing vector databases as essential. Cost optimization is now a top priority, alongside strategies like GPU utilization and tiered storage to manage expenses. The session also emphasizes the importance of edge computing and real-time data processing for improving AI capabilities, while noting ongoing challenges such as skill gaps and storage performance bottlenecks.

Hello. Hello everyone. Good afternoon. Welcome to the Data on Kubernetes Day. Today, um I am Idet Bugja. I am technology evangelist at Percona, and today I'm going to be your host in this event. But first of all, let's say thanks to our community sponsors, EDB, and A9 Nice, uh who support us and let us uh bring this event today in KubeCon. Also, we are going to thanks

to our volunteers who prepared this schedule, and they were able to review all the CFPs for today. Thank you as well to all our ambassadors around the world who helped us to build this community. And I want to mention that uh one example of the work that being done in Data on Kubernetes community is a Data on Ku- the DOK Get Started Guide. This is a community-driven

resource to help guide Data on Kubernetes beginners with tools that they need to start running Data on Kubernetes uh workflows. You can scan that QR code to get access to the GitHub repository, and if you want to also contribute and add more things, please just feel free to open a pull request. I want to also mention uh the Data on Kubernetes release 2025 is out. You can

find it scanning that QR code, and I'm going to mention some key findings for that year. The Data Data on Kubernetes community research full production maturity. Nearly half of organizations run now 50% or more of their data workflows on Kubernetes in production environments. With leaders surpassing 75%, and over 60% attribute more than 10% of their companies revenues of their deployments. That is a clear evidence that data

in Kubernetes has become mission-critical infrastructure. Let's talk about AI machine learning workloads that are driving the next wave of this generation. While databases remain on the top of workloads for the fourth consecutive year, AI and machine learning workloads have surged to 44% adoption. This is a remarkable 70 70% organizations now view vector databases critical infrastructure, signaling to the rise or retrieve augmented generation architectures and real data

systems. Now, let's talk about cost optimization tops for 2025 priorities. For the first time, cost optimization surpasses all their other priorities with 50% of AI and machine learning focused organizations sitting as storage cost as their biggest priority and challenge. Teams are implementing strategies such a GPU utilization optimization, tiered storage, and cross-cloud data transfer reduction to control expenses expenses as a workflow scale. The edge plus real-time shift.

Organizations are re-architecturing for real-time and distributed systems. 71% said edge computing is essential for the data strategy. And 74 54, sorry, cited real-time data processing as a critical to AI success. This marks a decisive move away from centralized batch-oriented data data And last but not least, a and performance gaps remain. Despite maturity, 40% of organizations report a skill gap and storage input-output performance remains to the top

of bottleneck. Closing these gaps through better tooling, training, and ecosystem collaboration will define the future of the face of data on Kubernetes excellence. Feel free to download the report for data on Kubernetes 2025 using that QR code.