Dell Technologies has worked closely with H2O.ai to deliver AutoML for AI and machine learning adopters through a jointly engineered and validated solution to help organizations capitalize on the benefits of AutoML for machine learning and AI workloads. We validated H2O Driverless AI running with the following Dell Validated Designs:
The following figure shows these two validated designs:
The Validated Design for AI with VMware and NVIDIA incorporates Dell Technologies infrastructure, VMware vSphere with Tanzu, NVIDIA GPUs, and NVIDIA AI Enterprise. This validated design is a virtualization platform that enables an enterprise to manage clusters of both on-demand Kubernetes containers alongside traditional virtual machines, providing complete life cycle management of those compute and storage resources.
The Validated Design for Analytics – Data Lakehouse incorporates Symworld Cloud Native Platform for the Kubernetes layer. Symworld Cloud Native Storage provides the Kubernetes Storage Class by discovering disks attached to the PowerEdge servers. Applications that require Persistent Volumes use this Kubernetes Storage Class.
PowerEdge servers provide compute resources for the AI workloads deployed on AutoML. The PowerEdge servers can be optionally configured with NVIDIA GPUs to support acceleration for neural network training and inference. PowerSwitch switches are used for network connectivity and out-of-band (OOB) connectivity. NVIDIA GPU operators provide automated deployment of NVIDIA software components required for the GPUs.
Note: Other Dell Validated Designs for container platforms, including Red Hat OpenShift, can be used with similar results. However, Dell Technologies has not validated these container platforms for AutoML as of the publication of this white paper.