A cause-effect chain is a sequence of multi-rate real-time tasks with data dependency. Cause-effect chains are generally subject to end-to-end timing constraints, especially in safety-critical systems. Communication paradigms greatly affect the end-to-end latency of cause-effect chains. This paper compares different communication paradigms (implicit communication, LET, DBP) with regards to the end-to-end latency of cause-effect chains using them, and proposes priority assignment strategies to optimize the end-to-end latency with specific communication paradigm. Experiments with synthesized data based on an automotive benchmark and randomly generated parameters are conducted to evaluate our results.