ROS 2 engineering workbench

01 / communication contract

Will these endpoints match?

The publisher offers a policy; the subscriber requests a minimum. Test reliability and durability separately, then together.

Publisher reliabilityMaximum delivery guarantee offered Subscriber reliabilityMinimum delivery guarantee requested Publisher durabilityWhether this writer retains history Subscriber durabilityWhether retained history is requested

Inspect each policy

Policy	Offered → requested	Result

The rule

Match = reliability compatible AND durability compatible Best effort → reliable: no Reliable → best effort: yes Volatile → transient local: no Transient local → volatile: yes

What this result means

A match here means only these two policies are compatible. It does not prove that endpoints exist, types match, discovery succeeds, data arrives on time, or the payload is correct.

Transient-local history is owned by the publisher and limited by its history and resources. This tool does not guarantee a sample count. Retained data is available to a late joiner only when both endpoints use transient local and the writer still has it.

Try a prediction first

Set a reliable, transient-local publisher. Which subscriber choices fail under these two policies? None: all four combinations are compatible, though historical delivery differs.

Primary referenceROS 2 Jazzy · QoS compatibility tables

Change one assumption.
Watch the result.

Will these endpoints match?

Inspect each policy

Is the measurement still usable?

Age is cumulative

Where does the tool end up?

Bent configuration