Please enable JavaScript.
Coggle requires JavaScript to display documents.
Real-Time Systems - Validation - Coggle Diagram
Real-Time Systems - Validation
Validation vs. Verification
Types of system representations
Model of user's intention
Model of system specification
SUT: system-under-test
The prime validation method is testing, while the prime verification method is formal analysis.
Overview
Validation: the consistency between an informal model of user's intention and the SUT
Verification: the consistency between the formal specification and the SUT
Specification errors: gaps between Validation and verification
Testing Challenges
Overview
Problems
Probe effect
Temporal distortion due control halting
Controllability
Probe effect
Design for testability
What?
Mechanisms that facilitate the testing of a system
Framework design
How?
Break into smaller subsystems with observable and well-defined interfaces
Static temporal control structure that is independent of the input data, this make possible test this structure in isolation
Reducing the temporal dimension of the input space
Ground state observability
Software determinism
Test data selection
Some methods
Requirements coverage
White-box testing
Model-based testing data selection
Operational profile
Peak load
Worst-case execution time (WCET)
Fault-Tolerance Mechanisms
Cyclic systems
Random test data selection
Common coverage criteria
Function coverage
Statement coverage
Branch coverage
Condition coverage
Fault coverage
Test oracle
System evolution
Validating new versions
New-function testing
Regression testing
Designing a testing campaign: a
5W2H
, without the
WHO
Testing of Component-Based Systems
Component provider: a component independent from the context of use (Technology Independent Interface)
Component user: performance concerned testing
Communicating components
Formal methods
Formal methods in the real world
Conceptual model building
Model formalization
Analysis of the formal model
Interpretation of the results
Classification of formal methods
Use of concepts and notations of discrete mathematics
Use of formalized specification languages with some mechanical support tools
Use of fully formalized specification languages with comprehensive support environments, including mechanism theorem proving or proof checking.
Benefits of formal methods
Model checking
Fault injection
Testing and debugging
Dependability forecasting
Software-implemented fault injection
Predictability
Reachability
Less effort than physical fault injection
Physical fault injection
Sensor and actuator failures