News

Red Hat Updates OpenShift AI Platform

• By John K. Waters
• May 22, 2024

Open-source solutions provider Red Hat announced updates to Red Hat OpenShift AI, an open hybrid AI and machine learning (ML) platform built on Red Hat OpenShift cloud applications platform. The updates allow enterprises to create and deliver AI-enabled apps at scale across hybrid cloud environments, the company says.

These enhancements also underscore Red Hat’s vision for AI, integrating customer choice into intelligent workloads from the underlying hardware to services and tools such as Jupyter and PyTorch. This update promises faster innovation, increased productivity, and the ability to integrate AI into daily business operations via a flexible, scalable, and adaptable open-source platform supporting both predictive and generative models, regardless of cloud environment usage.

"Bringing AI into the enterprise is no longer an 'if,' it’s a matter of 'when,'" said Ashesh Badani, SVP and chief product officer at Red Hat, in a statement. "Enterprises need a more reliable, consistent, and flexible AI platform that can increase productivity, drive revenue, and fuel market differentiation. Red Hat OpenShift AI is our answer, making it possible for IT leaders to deploy intelligent applications across hybrid clouds while refining operations and models as needed to support production applications and services."

Many customers face challenges transitioning AI models from experimentation to production, such as increased hardware costs, data privacy concerns, and trust issues with sharing data with SaaS-based models. As generative AI evolves rapidly, organizations struggle to establish a reliable core AI platform that can operate on-premises or in the cloud.

According to IDC, to fully leverage AI, enterprises need to modernize applications and data environments, dismantle barriers between systems and storage platforms, enhance infrastructure sustainability, and strategically deploy workloads across cloud, datacenter, and edge locations. Red Hat’s AI strategy emphasizes flexibility, enabling enterprises to progress in their AI adoption journey while their needs and resources evolve.

The latest version of Red Hat OpenShift AI, version 2.9, includes several key features:

• Model Serving at the Edge: Extends AI model deployment to remote locations using single-node OpenShift, offering inferencing in resource-constrained environments with intermittent network access. This feature provides a consistent operational experience from core to cloud to edge, with out-of-the-box observability.
• Enhanced Model Serving: Supports multiple model servers for predictive and generative AI, including KServe, vLLM, TGIS, and Caikit-nlp-tgis runtime. This enhancement allows users to run various models on a single platform, reducing costs and simplifying operations.
• Distributed Workloads with Ray: Utilizes CodeFlare and KubeRay to accelerate data processing and model training across cluster nodes. These tools simplify task orchestration and monitoring, optimize node utilization, and allocate resources like GPUs effectively.
• Improved Model Development: Features project workspaces and additional workbench images for data scientists, offering flexibility with IDEs and toolkits such as VS Code and RStudio. Enhanced CUDA support is also available.
• Model Monitoring Visualizations: Improves observability into AI model performance and operational metrics.
• New Accelerator Profiles: Allows administrators to configure hardware accelerators for model development and serving workflows, providing self-service access to suitable accelerators for specific workloads.

Red Hat OpenShift AI underpins IBM’s watsonx.ai, and organizations like AGESIC and Ortec Finance are leveraging it to drive AI innovation and growth.