www thalesgroup com PRESTO Improvements of Industrial RealTime

  • Slides: 13
Download presentation
www. thalesgroup. com PRESTO: Improvements of Industrial Real-Time Embedded Systems Development Process {Shuai. Li}{Michel.

www. thalesgroup. com PRESTO: Improvements of Industrial Real-Time Embedded Systems Development Process {Shuai. Li}{Michel. Bourdelles}{Arnaud. Trottet}@fr. thalesgroup. co m THALES Communications & Security 3 rd FITTEST Industrial Day May 31 st 2013

Project Context: Real-Time Embedded Systems 2 / 3 rd FITTEST Industrial Day May 31

Project Context: Real-Time Embedded Systems 2 / 3 rd FITTEST Industrial Day May 31 st 2013

Project Goal 3 / Current Design Difficulties u RTES constrained by resources available (e.

Project Goal 3 / Current Design Difficulties u RTES constrained by resources available (e. g. processor, power)… u …but it is difficult to proceed to performance analysis at an early stage in design when the execution platform is not available. PRESTO General Idea u Provide tools to evaluate Software/Hardware allocations u …by analyzing functional and non-functional properties… u …and integrating tools into an industrial development process. 3 rd FITTEST Industrial Day May 31 st 2013

Project Synopsis 4 / Modeling Software Requirements (Functional and non-functional) Software Test Design and

Project Synopsis 4 / Modeling Software Requirements (Functional and non-functional) Software Test Design and Implementation on Reference Platform Behavior Capture Software Model Platform Model Analysis Functional Verification Test Trace Generation Extrapolation Estimation Non-Functional Verification ? 3 rd FITTEST Industrial Day May 31 st 2013 Platform Requirements Software Test on Real Platform Prototype

Project Technical Relevance 5 / Behavior Capture u Instrumentation strategies u Instrumentation impact u

Project Technical Relevance 5 / Behavior Capture u Instrumentation strategies u Instrumentation impact u Trace generation u Trace interoperability Modeling u General RTES modeling with stereotypes (e. g. UML) u Domain specific modeling (e. g. application, analysis specific) u Formal modeling of functional and non-functional properties u Model interoperability Analysis u Execution and specification trace comparison 3 rd FITTEST Industrial Day May 31 st 2013

Consortium 6 / United Kingdom France 3 rd FITTEST Industrial Day May 31 st

Consortium 6 / United Kingdom France 3 rd FITTEST Industrial Day May 31 st 2013

Consortium 7 / Finland Italy Greece 3 rd FITTEST Industrial Day May 31 st

Consortium 7 / Finland Italy Greece 3 rd FITTEST Industrial Day May 31 st 2013

www. thalesgroup. com THALES Case-Study Demo {Shuai. Li}{Michel. Bourdelles}{Arnaud. Trottet}@fr. thalesgroup. co m THALES

www. thalesgroup. com THALES Case-Study Demo {Shuai. Li}{Michel. Bourdelles}{Arnaud. Trottet}@fr. thalesgroup. co m THALES Communications & Security 3 rd FITTEST Industrial Day May 31 st 2013

THALES Software Radio Case-Study 9 / Functional Overview u Software radio application u TDMA

THALES Software Radio Case-Study 9 / Functional Overview u Software radio application u TDMA radio protocol for mobile ad-hoc wireless networks Software Architecture Module Thread: Activated when message in buffer Buffer: FIFO message storage Module 3 rd FITTEST Industrial Day May 31 st 2013 Communication

THALES Software Radio Case-Study 10 / Functional Overview u Software radio application u TDMA

THALES Software Radio Case-Study 10 / Functional Overview u Software radio application u TDMA radio protocol for mobile ad-hoc wireless networks Software Architecture Business Wrapper Business code: message processing Wrapper code: message reception and transmission Wrapper Business Wrapper 3 rd FITTEST Industrial Day May 31 st 2013 Instrumentation

Experiment 11 / Objectives u Software radio application modeling u Instrumented code generation from

Experiment 11 / Objectives u Software radio application modeling u Instrumented code generation from model u Execution trace generation u Trace visualisation with PRESTO partner tool u Instrumentation overhead evaluation Experimented tools and methods u UML (MARTE) modeling language u Softeam’s Modelio modeler for PRESTO u Pragma. Dev’s MSC-Tracer analysis tool Demonstration Time 3 rd FITTEST Industrial Day May 31 st 2013

Results 12 / Instrumentation Overhead Number of message types 21 Instrumentation points hit during

Results 12 / Instrumentation Overhead Number of message types 21 Instrumentation points hit during execution 38 Single instrumentation point overhead (µs) Min Max Avg 31. 00 40. 00 33. 09 Ratio (single overhead) / (total time) 0. 03% Ratio (total overhead) / (total time) 1. 07% Discussion u The instrumented execution does not result in any missed deadlines when they are equal to 100 ms. u In other software radios, deadlines are on the scale of 1 ms. Theoretically a maximum of 30 points are then allowed. In the experimental execution, deadlines would have been missed. u Instrumentation overhead can be limited by choosing active instrumentation points at model level. 3 rd FITTEST Industrial Day May 31 st 2013

Conclusion 13 / Results u Trace generation flow u Functional information u Timing information

Conclusion 13 / Results u Trace generation flow u Functional information u Timing information u Instrumentation strategy and impact u Solution to ease modeling Future Works u Exploit traces with formal properties in the model u Energy consumption information 3 rd FITTEST Industrial Day May 31 st 2013