The *Pipeline* scenario is chosen to investigate the effects of
delay propagation. The main issue is whether the calculated delay
value from the processing of the first message is sufficient to
correctly adjust the variables when the second message is processed.
Figure 8 shows the two cases based on the
relationship of the adjusted delay value () sent by P2 to P3, and
the overhead and waiting time on P3. (Like our earlier models, we assume
here for simplicity that this is the first message P3 receives.
Clearly, in this case, .)

The interesting outcome of the models is that the analysis and update
of P3's variables during the processing of the second message is
effectively independent of the first message. Of course, the
delay value sent from P2 is derivative of the effects of the first
message and , but the expressions are invariant compared to those
when P2 sends a message to P3 without first receiving a message from
P1 (i.e., the *Two-Process, General* scenario). This conclusion
extends to cases where there is an arbitrary number of processes in
the pipeline.