This is the partition model of the Rockwell Avionics system from the tutorial to perform
partition-based latency analysis. The model shows 4 systems and the display as device. It
defines a flow from the display through the Display Manager, Page Content Manager, Flight
Manager to the Flight Controller and back. The flow is defined by providing flow specs for
each system component, by defining the connections between the subsystems, and by
specifying an end-to-end flow as described. The flow latency analysis determines a lower
bound on latency based on the fact that cross-partition communication implies a minimum
latency of the partition execution rate for each cross partition step. We have specified a
Partition_Period property for the rate. The end-to-end flow has an expected latency
property value of 300 ms. This is exceeded by the flow. Since we know that this is a lower
bound, we can conclude two things. First, the only way to reduce latency is to reduce the
number of partition crossings, e.g., by connecting the flight director directly to the page content
manager instead of via the flight manager. Second, the lower bound of latency implies an upper
bound on the rate at which the display request events can occur. This allows scheduling
analysis to become more realistic.