What does it mean to use a "compute-block" model to design your infrastructure? Is it:
-Standardizing the size of VMs to use only certain standard specfications of number of cpus, amount of ram, and disk size?
-Insuring the placement of VMs with standardized sizes in clusters so that processing, memory and storage resources of clusters are used in proporition - so that cluster resources are neither underutilized nor overutilized?
-Having a process of provisioning that allows VM sizing to cooncide with cluster capacity - so that the provisioning software drives the "compute-block" implementation?
What is it that exactly defines a "compute-block" model - or if there isn't an agreed upon definition, your opinions as to what the definition should be are welcome - thanks!