A Performance Analysis Tool for Scenario-Aware Streaming Applications

Abstract

Dataflow models are often used for analysing streaming applications. The recently introduced scenario-aware extension of the synchronous dataflow model can capture the dynamism in computation and communication resource requirements of streaming applications that originates from different modes of operation (scenarios). This scenario-aware dataflow model uses a probabilistic approach to express the order in which scenarios (and different execution times within a scenario) occur. This paper discusses a tool for exhaustive and simulation-based analysis of various important performance metrics for scenario-aware streaming applications.