AI Conformant Clusters in GKE

AI Conformant Clusters in GKE

by Janet Kuo & Federico Bongiovanni, GKE

We are excited to announce that Google Kubernetes Engine (GKE) is now a CNCF-certified Kubernetes AI conformant platform, designed to provide a stable and optimized environment for your AI/ML applications. This initiative, culminating in a major announcement of the Kubernetes AI Conformance program by CNCF's CTO Chris Aniszczyk, at KubeCon NA 2025, is set to simplify AI/ML on Kubernetes for everyone. You can check out the Opening Keynote here.

During the keynote, Janet Kuo, author of this blog and Staff Software Engineer at Google, performed a live Demo, demonstrating the practical power of an AI Conformant cluster. If you are interested in the technical specifics, you can learn more about the demo here.

Why AI Conformance Matters

The primary goal of the Kubernetes AI Conformance program is to simplify AI/ML on Kubernetes, guarantee interoperability and portability for AI workloads, and enable a growing ecosystem of AI tools on a standard foundation.

Setting up a Kubernetes cluster for AI/ML can be a complex undertaking. An AI-conformant platform like GKE handles these underlying complexities for you, ensuring that your environment is optimized for scalability, performance, portability, and interoperability.

For a detailed look at all the requirements and step-by-step instructions on how to create an AI-conformant GKE cluster, we encourage you to read the GKE AI Conformance user guide.

What Makes GKE an AI-Conformant Platform?

A Kubernetes AI-conformant platform like GKE handles the underlying complexities for you, providing a verified set of capabilities to run AI/ML workloads reliably and efficiently. Here are some of the key requirements that GKE manages for you:

• Dynamic Resource Allocation (DRA): GKE enables more flexible and fine-grained resource requests for accelerators, going beyond simple counts. This is crucial for workloads that need specific hardware configurations.
• Intelligent Autoscaling for Accelerators: GKE implements autoscaling at both the cluster and pod level to ensure your AI workloads are both cost-effective and performant.
  • Cluster Autoscaling works at the infrastructure level. It automatically resizes node pools with accelerators, adding nodes when it detects pending Pods that require them and removing nodes to save costs when they are underutilized.
  • Horizontal Pod Autoscaling (HPA) works at the workload level. HPA can automatically scale the number of your pods up or down based on real-time demand. For AI workloads, this is especially powerful, as you can configure it to make scaling decisions based on custom metrics like GPU/TPU utilization.
• Rich Accelerator Performance Metrics: GKE exposes detailed, fine-grained performance metrics for accelerators. This allows for deep insights into workload performance and is essential for effective monitoring and autoscaling.
• Robust AI Operator Support: GKE ensures that complex AI operators, such as Kubeflow or Ray, can be installed and function reliably, enabling you to build and manage sophisticated ML platforms with CRDs.
• All-or-Nothing Scheduling for Distributed Workloads: GKE supports gang scheduling solutions like Kueue, which ensure that distributed AI jobs only start when all of their required resources are available, preventing deadlocks and resource wastage.

A Unified and Evolving Standard

The Kubernetes AI Conformance program is designed as a single, unified standard for a platform to support all AI/ML workloads. This reflects the reality that modern AI processes, from training to inference, increasingly rely on the same underlying high-performance infrastructure.

What's Next?

We invite you to explore the benefits of running your AI/ML workloads on an AI-conformant GKE cluster.

• Learn more about the program: check out the CNCF Kubernetes AI Conformance repository.
• Dive deeper into AI/ML on GKE: read the Introduction to AI/ML workloads on GKE.

The launch of the AI Conformance program is a significant milestone, but it is only the first step. We are eager to continue this conversation and work alongside the community to evolve and improve this industry standard as we head into 2026.

Unlocking High-Performance AI/ML in Kubernetes with DRANet and RDMA

DraNet Enters Beta! High-Performance Networking in Kubernetes

by Antonio Ojea & Federico Bongiovanni, Kubernetes/GKE

We are excited to announce that DraNet has officially entered a beta state! This marks a major leap forward in our mission to streamline and enhance high-performance networking for AI and HPC workloads within Kubernetes. As we progress towards a stable General Availability (GA) release, we are eager to gather your feedback on the current state of the project.

Why DraNet?

DraNet was born from the lessons we learned at Google, observing the challenges end-users faced when running AI and HPC workloads on Kubernetes. The existing networking solutions, often repurposed from traditional networking or bespoke and complex, fell short of providing a good user experience and efficient operational models.
For instance, managing RDMA (Remote Direct Memory Access) interfaces often involved a complex combination of CNI chaining and device plugins. This not only created an unnecessary operational overhead for administrators but also led to coordination issues between different components that needed to work in harmony impacting resilience and scalability.
Another significant pain point we identified was the need for fine-grained interface tuning. AI workloads, for example, are extremely sensitive to latency. The presence of some eBPF programs on network interfaces, or the need to configure specific NIC parameters, could severely impact performance latency and/or throughput. Users were often forced to create custom init containers just to apply these settings, adding another layer of complexity.

Introducing DraNet: A Native and Declarative Solution

DraNet is a native integration with Kubernetes that uses the core Dynamic Resource Allocation (DRA) API to address these challenges by treating high-performance network interfaces as first-class citizens in Kubernetes. Here's how:

• Simplified RDMA Management: DraNet manages RDMA interfaces natively, handling the different requirements to offer a unified and seamless user experience. No more need for coordinating different components.
• Declarative Interface Tuning: With DraNet, you can declaratively set interface properties. Need to disable eBPF programs to reduce packet processing overhead or set specific NIC parameters? You can now do this directly in your Kubernetes manifests, eliminating the need for custom scripts or init containers.
• Standalone and Secure: DraNet is designed as a standalone binary, allowing it to run in a distroless container. This significantly reduces the attack surface and the frequency of security-related updates for the container image. By interacting directly with the kernel via stable APIs like netlink, it avoids dependencies on third-party projects, improving both resilience and performance.
• Lightweight and Fast: The DraNet container image, with a compressed size of less than 50MB, has a direct impact on node startup times, allowing for faster deployment and scaling of your workloads.

Beta Release and the Road to GA

DraNet is now in a beta state, signifying that it is ready for broader community testing and feedback. This move to beta is aligned with the maturation of the Kubernetes Dynamic Resource Allocation (DRA) KEP (KEP-4381), a foundational technology for DraNet. We are continuing our active development as we work towards a future General Availability release.

We Welcome Your Feedback and Contributions!

DraNet is an open-source project, and we believe that community involvement is key to its success. As we work towards our GA release, we welcome your feedback, whether it's on the design, user experience, or performance.
You can contribute in many ways:

• Code contributions: We have a fast-paced development cycle and welcome new contributors. Check out our contributing guidelines to get started.
• Documentation: Help us improve our documentation to make it easier for new users to get started with DraNet.
• Share your opinion: Your feedback is invaluable. Let us know how you are using DraNet and what we can do to make it better.

To learn more about DraNet and get started, please visit https://dranet.dev/. We look forward to building the future of high-performance networking in Kubernetes with you!