When running real-time applications and guaranteeing the required performance from the hypervisor, design decisions must be made to ensure consistent resource availability for the vPAC workloads.
With vPAC installations, it is possible for a mixture of real-time (vPR) and non-real-time (vAC, also commonly referred to as Common Substation Platform or CSP) workloads to be running adjacent to, yet logically isolated from, each other on the same host. This design methodology of sharing hosts can provide additional benefits with further consolidation of functionalities, simpler networking, easier methods of sharing data, and a reduction in management tasks.
However, there can be internal or external standards, or interpretations of regulatory requirements that drive a design philosophy of physically separating critically classified functions that demand real-time resources against those with less strict operational requirements.
As the concentration of this validated solution guide is on VVS Reference Architecture and Shared Process/Station Bus, the architecture depicted in Separate Process/Station Bus is achievable with the guidance provided here, coupled with general Software Defined Data Center (SDDC) and VMware Cloud Foundation (VCF) documentation, available at docs.vmware.com.
Both virtual machines and containers are supported with Edge Compute Stack. When using containers, it is necessary to deploy VMware Tanzu Kubernetes Grid (TKG), which allows for customization of the base worker node image. Using this customization enables deployment of a real-time Linux-based OS for the containerized workloads to run.
Successful, low-latency baselines are achieved using VMware Photon OS and Debian, Ubuntu, CentOS, and other Linux RT-based virtual machines. The TKG layer contains multiple objects that enable the capability to run Kubernetes workloads natively in the ESXi hosts that can initially make the deployment and maintenance of vPAC-based applications simpler than bare metal applications.
