Monitoring engines are deployed in various locations across the infrastructure. Each monitoring engine is be configured to monitor a subset of the infrastructure and publish the monitoring results to the reporting engine. The reporting is deployed in a central location and it aggregates the monitoring results produced by each monitoring engine.
This is the simplest model of all. Both the monitoring engine and the reporting engine are deployed at the same location. The monitoring engine, monitors the whole infrastructure and pushes the result to the reporting engine. This scenario is commonly found in deployments within one administrative domain, in which there is no need for distributing the deployments across the infrastructure. Of course as the infrastructure might scale, it is possible to transition to a distributed model with a minimal effort.
Monitoring engines are distributed across the infrastructure. Reporting engines are deployed locally for local reporting and centrally providing a holistic view.