Transformation and Verification of Data Flows Based on AADL

Jian SUN,Min XU
DOI: https://doi.org/10.3969/j.issn.1673-629X.2016.04.009
2016-01-01
Abstract:AADL supports structural modeling for software and hardware,and imports non-functional attributes description such as real-time and reliability in embedded real-time system. During the process of MDD ( Model-Driven Development) ,it is of great significance for ensuring the system real-time performance and improving the efficiency of system development to find potential design problems on critical aspects in the model design stage. In order to analyze the flow latency of AADL models,a method is proposed taking the analysis of data flows of AADL to form a formal description of data flows. The mapping relationship from formal description to time automaton semantics is regarded as the definition of mapping rules. On building the timed automata of date flows,methods and samples are given to transform both simple and mixed flows into timed automata. In the transformation of mixed flows,a template of non-periodic thread is presented to support the comprehensive analysis of data flows. At last the reference query statements is given to verify the properties of the data flows,and the necessary experimental tests of time automaton model converted from data flows are carried out.
What problem does this paper attempt to address?