KubeCon 2023 Through Platform9’s Lens: Key Takeaways and Innovative Demos

A group photo of the Platform9 team at the KubeCon event

KubeCon + CloudNativeCon North America in Chicago has come and gone, but the Platform9 team is still buzzing from the event, brimming with newfound knowledge and innovative ideas. The conference provided an excellent platform for education and networking, attracting a global gathering of cloud-native technology experts from well-known open-source projects.

Keynote speeches and panel discussions delved into topics like sustainability in cloud computing, the evolving landscape of cloud-native use cases in the Kubernetes world, and the future of open-source technologies.

Platform9 was heavily involved in the event, actively participating in the AI/ML co-located event, discussing cost challenges in the public cloud, and engaging in insightful discussions about how to ensure Kubernetes’s high availability and scaling across private and public clouds.

Zero-Touch Kubernetes Provisioning from the Bare Metal Up – a live demonstration at KubeCon 2023

Platform9 Field CTO Peter Fray demonstrated our zero-touch provisioning capability for Kubernetes clusters in real time.

The setup included three servers — two forming a Kubernetes cluster — along with a switch, router, and 5G hotspot. This setup is used to demonstrate remote deployment in diverse environments like edge locations and data centers.

Fray demonstrated PXE booting a server, automating a Kubernetes stack deployment on it, and adding a new worker node to an existing cluster, highlighting Platform9’s technology’s simplicity and versatility in various deployment scenarios, including integration with video surveillance systems.

Running 10,000 batch jobs per minute across K8s clusters

Chris Jones talks about Armada, an innovative open-source tool designed to schedule multiple workloads across various Kubernetes (k8s) clusters. Perfectly fitting into Platform9’s management capabilities for clusters both on-premises and in the cloud, Armada excels in high-performance computing and machine learning contexts. It can actively manage up to 10,000 batch jobs per minute across diverse Kubernetes clusters.

Chris also spotlighted Armada’s ability to manage and balance resources by showing multiple jobs within a queue, significantly improving resource utilization across clusters. It also provides a real-time view of these jobs running on Platform9’s infrastructure. By offering these capabilities, Armada simplifies infrastructure management and batch computing across clusters, freeing data scientists from the manual labor of running jobs on various infrastructures.

Overall, Chris emphasizes Armada as a dynamic tool for automated resource management across multiple GPU-enabled Kubernetes clusters, underscoring its utility in high-performance computing and machine-learning environments.

Fully managed Machine Learning and Kubernetes on-premises with Union.ai and Platform9

Chris Paap discusses the integration of Union AI and Platform9 Managed Kubernetes for machine learning. Union AI is built upon the Flyte open-source project and relies on Kubernetes for on-premises scaling. Platform9 simplifies the deployment and management of Kubernetes, enabling users to easily run Union AI for machine learning tasks. This allows data scientists to focus on their work without having to worry about Kubernetes management. The goal is to provide invisible infrastructure, making it seamless for data scientists to scale their work as needed.

Maximizing the efficiency of EKS clusters and saving 50% on your AWS costs

Platform9 launched early access of their latest ground-breaking innovation, called Elastic Machine Pool (EMP). We had many insightful discussions with attendees and showcased the EMP demo. Most people we spoke with said their EKS cluster utilization ranged between 20 and 30%. This has a massive impact on cloud costs, with up to 50% of spending being wasted. These discussions highlighted a common quest for Kubernetes efficiency and cost-effectiveness in the public cloud, a challenge that cuts across industries and use cases.

Madhura Maskasky, our co-founder and VP of Product, being interviewed on how EMP works on TFIRMadhura Maskasky, our co-founder and VP of Product, being interviewed on how EMP works

Madhura Maskasky, our co-founder and VP of Product, was featured on theCUBE and TFIR interviews discussing how EMP works, whose approach to solving the EKS utilization problem is very different and unique in that it targets utilization at the hypervisor layer. Keep an eye out for those videos in the coming weeks.

Check out a short demo of EMP here:


KubeCon 2023 went beyond being just an event; it was a vibrant gathering that celebrated community, shared learning, and collective growth in the cloud-native ecosystem. The insights, conversations, and demonstrations we witnessed are indicative of a thriving and evolving community, one that is not just keeping pace with technological advancements but is also shaping them.

We left KubeCon inspired and energized, ready to embrace new challenges and contribute to this ever-growing community. The future is bright for cloud-native technologies, and the community’s collective wisdom and enthusiasm are leading the way.

Interested in delving more into on-premises AI/ML Kubernetes, Chris Jones can help you with your critical infrastructure decisions to scale your workloads. Schedule a meeting with him at: https://meetings.hubspot.com/cjones2

Looking to double your EKS utilization and cut AWS costs by 50% or more? Reach out to our EMP expert team to see a live demo and learn more about Elastic Machines Pool: https://go.platform9.com/meetings/madhura-maskasky/emp


You may also enjoy

Top 6 FinOps KPIs for EKS  

By Chris Jones

Kubernetes FinOps: Comparing Platform9 Elastic Machine Pool to the Karpenter Autoscaler for EKS

By Joe Thompson

The browser you are using is outdated. For the best experience please download or update your browser to one of the following:

VMware admin? Join our hands-on OpenStack workshopRegister Now