"description": "This deployable architecture illustrates the deployment of a GPU-enabled Red Hat OpenShift cluster on IBM Cloud with the OpenShift AI add-on, showcasing the key components and their interactions for AI workloads support.<br><br>You can create a fully-configured <b>VPC</b> by selecting the [Cloud automation for VPC](https://cloud.ibm.com/catalog/7a4d68b4-cf8b-40cd-a3d1-f49aff526eb3/architecture/deploy-arch-ibm-vpc-2af61763-f8ef-4527-a815-b92166f29bc8-global) component or you can use an existing VPC. The VPC can be provisioned either in a single-zone or multi-zone configuration, depending on your availability requirements. The default configuration deploys a VPC with three zone subnets, configures the predefined ACLs, and attaches a public gateway with one of the subnets. <br><br>To utilize the capabilities of <b>OpenShift AI</b>, you must configure <b>at least one GPU-enabled node</b> within your cluster. The architecture supports a flexible compute environment, allowing you to deploy a mix of <b>GPU</b> and <b>non-GPU</b> nodes across different worker pools. This solution offers two worker pools that are spread across all three zones, with the OpenShift AI add-on set as the default configuration.<br><br>A <b>Cloud Object Storage (COS)</b> instance is required for the OpenShift cluster, which serves as the internal registry storage for the OpenShift cluster. The [Cloud automation for Object Storage](https://cloud.ibm.com/catalog/7a4d68b4-cf8b-40cd-a3d1-f49aff526eb3/architecture/deploy-arch-ibm-cos-68921490-2778-4930-ac6d-bae7be6cd958-global) deployable architecture can be used to create a COS instance. <br>Additionally, the storage bucket can be encrypted using <b>Key Management Services(KMS)</b> to encrypt the cluster and the boot volumes used by the cluster to enhance security.<br><br>It can be integrated with <b>Observability services</b> such as [Cloud automation for Cloud Monitoring](https://cloud.ibm.com/catalog/7a4d68b4-cf8b-40cd-a3d1-f49aff526eb3/architecture/deploy-arch-ibm-cloud-monitoring-73debdbf-894f-4c14-81c7-5ece3a70b67d-global) to provide robust monitoring capabilities and capture essential metrics such as CPU and memory utilization, to help you proactively monitor system performance and resource consumption, [Cloud automation for Activity Tracker Event Routing](https://cloud.ibm.com/catalog/7a4d68b4-cf8b-40cd-a3d1-f49aff526eb3/architecture/deploy-arch-ibm-activity-tracker-918453c3-4f97-4583-8c4a-83ef12fc7916-global) to monitor how users and applications interact with the cluster, to support compliance and auditability, and [Cloud automation for Cloud Logs](https://cloud.ibm.com/catalog/7a4d68b4-cf8b-40cd-a3d1-f49aff526eb3/architecture/deploy-arch-ibm-cloud-logs-63d8ae58-fbf3-41ce-b844-0fb5b85882ab-global) to collect, view, and analyze platform logs related to the cluster and logs generated by applications deployed on the cluster.<br><br><b>Secrets Manager</b> Integration can also be enabled by using [Cloud automation for Secrets Manager](https://cloud.ibm.com/catalog/7a4d68b4-cf8b-40cd-a3d1-f49aff526eb3/architecture/deploy-arch-ibm-secrets-manager-6d6ebc76-7bbd-42f5-8bc7-78f4fabd5944-global) for centrally managing Ingress subdomain certificates, ensuring the secure management of sensitive credentials.<br><br>This architecture provides a robust and secure foundation for deploying and managing your AI workloads on Red Hat OpenShift cluster, enabling seamless integration, enhanced security, and comprehensive monitoring for your applications."
0 commit comments