News
Ubuntu 18.04 Adds Kubeflow Support, Performance Optimization for AI and ML Workloads
- By John K. Waters
- 05/02/2018
The latest release of the Linux Distro Ubuntu, 18.04, was designed to support artificial intelligence (AI) and machine learning (ML) workflows, according to Canonical CEO Mark Shuttleworth.
Included in the new features is support for: Kubeflow, the Google approach to TensorFlow on Kubernetes. Kubeflow is the open source project focused on making deployments of machine learning (ML) workflows on Kubernetes "simple, portable, and scalable," the project page states.
A range of CI/CD tools were integrated in Canonical's distribution of Kubernetes and aligned with the Google Kubernetes Engine (GKE) for on-premises and on-cloud AI development, according to the company.
In a conference call announcing the release, Shuttleworth also commented on the the performance improvements enterprises will find with the release: "Multicloud operations are the new normal," he said. "Boot time and performance-optimized images of Ubuntu 18.04 LTS on every major public cloud make it the fastest and most-efficient OS for cloud computing, especially for storage and compute-intensive tasks like machine learning."
The Canonical Distribution of Kubernetes (CDK) supports GPU acceleration of workloads using the NVIDIA device plugin for Kubernetes. Complex workloads like Kubeflow that leverage NVIDIA GPUs "just work" on CDK, the company said, "reflecting joint efforts with Google to accelerate machine learning in the enterprise and providing a portable way to develop and deploy ML applications at scale." Applications built and tested with Kubeflow and CDK are transportable to Google Cloud.
Developers working on Ubuntu can create applications on their workstations, test them on private bare-metal Kubernetes with CDK, and run them across data sets on Google's GKE. "The resulting models and inference engines can be delivered to Ubuntu devices at the edge of the network," the company said, "creating a perfect pipeline for machine learning from workstation, to rack, to cloud and device."
"Having an OS that is tuned for advanced workloads such as AI and ML is critical to a high velocity team" said David Aronchick, product manager in Google's Cloud AI group. "With the release of Ubuntu 18.04 LTS and Canonical's collaborations to the Kubeflow project, Canonical has provided both a familiar and highly performant operating system that works everywhere."
"Whether on-premise or in the cloud," he continued, "software engineers and data scientists can use tools they are already familiar with, such as Ubuntu, Kubernetes and Kubeflow, and greatly accelerate their ability to deliver value for their customers."
About the Author
John K. Waters is the editor in chief of a number of Converge360.com sites, with a focus on high-end development, AI and future tech. He's been writing about cutting-edge technologies and culture of Silicon Valley for more than two decades, and he's written more than a dozen books. He also co-scripted the documentary film Silicon Valley: A 100 Year Renaissance, which aired on PBS. He can be reached at jwaters@converge360.com.