Home > Storage > PowerFlex > White Papers > Dell APEX Block Storage for Azure: PostgreSQL Deployment and Performance on Kubernetes > Logical architecture
The logical architecture that is described in this section shows Dell APEX Block Storage deployment across multiple availability zones in a US East region using locally attached NVMe drives Azure VM SDS instances. The setup includes three fault sets, and one fault set is assigned per AZ. In total the solution includes six SDS instances, where two SDS instances are assigned per AZ, spreading data across multiple AZs for high availability and resiliency. The placement of MDMs (and the PFMP cluster) is also spread across AZs to minimize the chance of a complete management stack failure.
The SDS is installed on six Azure VM instances of Standard L32as v3 type with SUSE Linux OS. The Azure VMs that are deployed as SDS nodes in Azure are backed by locally attached NVMe SSD drives. The four locally attached NVMe drives per each Azure VM SDS instances are configured into a single protection domain. A medium granularity storage pool is created, using all the storage devices available within the protection domain as shown in the following figures:
Figure 6. Azure portal shows the Azure VM SDS instances
Figure 6. Azure portal shows the Azure VM SDS instances
The Azure SDC instances are created to consume the APEX Block Storage exist in the same US East region. A Kubernetes cluster with one Control plane and three worker nodes is created. Dell PowerFlex CSI is installed on top of the Kubernetes cluster to facilitate dynamic provisioning of persistent volumes for PostgreSQL database pods from Dell APEX Block Storage running on each worker node as shown in the following figure: