System Architecture and Analysis

For system architects, the SystemLab PS is an invaluable tool for performance analysis. SystemLab models provide performance measuring capabilities that allow architects to readily understand the impact of design decisions on the performance of the overall system. New functionality can be added to the system, and the effect on system performance can be measured and evaluated. What-if scenarios can be played out to provide answers to system questions in a fraction of the time that it would take to build physical prototypes.

Software Development and Debug

During the design and integration phases of system development, the vast majority of time is spent trying to determine the cause of system failures. Once the behavior is understood, the correction and validation can be quick, but getting to the understanding without visibility is extremely difficult. Much time and effort is spent devising experiments to glean information through external interfaces in order to develop theories, and often times the experiments themselves modify the system behavior such that the problem is masked or changed.

The SystemLab PS real-time system models provide unprecedented visibility into the operation of the entire system. Contents of internal memories, hardware state, and software state are exposed to enable engineers to quickly understand the behaviors of individual subsystems and the system as a whole. Software developers can monitor the execution of code on any processor in the system and, in fact, can monitor the activity of multiple software programs running simultaneously on different processors. Because the model operates in real time, the timing interactions between subsystems are respected.

Diagnostics Development

The availability of a real-time system model for diagnostics teams enables tremendous improvements in the quality of diagnostic tests, fault coverage, and development productivity. Because diagnostics development is a specialized form of software
development, all of the benefits of system-wide visibility and debugging capability apply also to the diagnostics development process. Code execution can be analyzed to prove that it is performing as intended, and breakpoints and tracing functions can be used to debug diagnostic software.

Fault Insertion

A major benefit of the large-scale system model is that it allows faults to be inserted in almost any part of the system. Diagnostic software that is designed to detect these faults can be tested to see if it meets requirements. For example, a broken wire fault can be forced onto a wiring harness in a subsystem to determine if diagnostics code can detect it. If not, the code can be modified to check for such a fault and then it can be re-tested for successful execution. In the physical vehicle or system, it may be impractical to test for such a fault due to physical inaccessibility, destructiveness of the test, or because it takes too much time to make such a change and then restore the system after the test. More complex failure scenarios can be modeled as well, including multiple simultaneous failures and faults between two or more subsystems. In addition, fault coverage analysis can be performed on a system-wide basis. A sequence of faults can be injected to the system model, and the system response can be monitored to determine if the fault reporting capability can detect each fault. Overall coverage statistics can then be compiled, which will direct further diagnostics enhancements and improve the robustness of the system.