Admin workstation
The admin workstation is the system where IBM® Control Desk and docker are installed.
Admin will create the docker images of IBM Control Desk components on this machine and push it to the docker private repository.
Workload components that are deployed to POD are IBM Control Desk workloads components.
ICD components provide six different workloads (UI, cron task, report, API, MEA, and JMS consumer) that can be used independently based on the processing and isolation needs. ICD UI and cron task workloads are required to run ICD components, and other workloads can be deployed based on the need. To run a workload, the ICD code needs to be packaged to contain the needed code into an application bundle that can be used to create a container. ICD components provide tools that can be used to build these packages for each workload. For each workload that needs to be run, a separate container image must be built and deployed into the containerized environment. Details on creating the docker images are detailed in Creating Docker Images for IBM Control Desk Components on Admin Workstation.
ICD components use a WebSphere Liberty runtime to run the code in a containerized environment in the Kubernetes cluster. Running using traditional WAS ND is not supported within the containerized environment. The following link provides some information on the support for running Maximo code in WebSphere Liberty. Note that some limitations currently exist, follow link for reference https://www.ibm.com/support/pages/maximo-asset-management-761-websphere-liberty-support.
The ICD runtime is broken up into various workloads. A workload is a specific type of work that can be used to isolate the processing needs of that work so that it can be independently managed. For example, a UI workload is a type of work that allows a user to access the user interface of ICD components using a web browser. A cron task workload is a type of work that allows background jobs to be run. When a UI workload and cron task workload are deployed on two separate machines, the CPU and memory consumed by these workloads do not affect each other, and these workloads can be independently scaled and managed based on the needs. For more reference, follow Deploying Maximo 7.6.1.X to Red Hat OpenShift (this link is for OpenShift, similar steps can be followed for Kubernetes).