Logical Breakdown of Compute Fabric

Solution Set - it can contain an arbitrary arrangement of Kubernetes resources that are deployed together and will be directed at a single problem

Maybe a parallel to Microservice boundary here

• Items in a Node can share a local data source 
• Synchronization Context - defines logical partition for calculation sets 
• Generation is a windowed view in compute environment - all resources are directed to a time segment
• SyncSet aggregation mechanism for SolutionSets
• SolutionSet - a grouping of kubernetes services, pods, etc. targeted at a particular problem.

Generations are time-sliced compute contexts and as generations complete their work they re-initialize to another later time-slice 

See https://techvidvan.com/tutorials/spark-streaming-window-operations/ for a description of how a similar idea is realized in spark.

In the above diagram, a generation is how k8's defines the resources for each windowed computation.