Estimating Execution Time using Discrete-event Simulation for large complex Aerospace Systems

Estimating

Emerging, high-speed bus technologies and migration to multi-core platform requires detailed high-level system analysis before scheduling implementation. Systems engineers run multi-variant workloads to conduct a battery of trade-offs including task response time vs. power dissipated; distributed vs. centralized processing; and throughput vs. cost. Analytical means and over-provisioning cannot handle the system complexity required to identify system bottlenecks, tight timing budgets and power dissipated. The increased scrutiny of aerospace budgets and larger number of projects means that the project team must be correct on the first attempt.

Model-based system design approach enables the user to distribute software tasks across the processors/cores, determine how much load can be handled by the bus and get an early power estimate. Later the user can refine the model with detailed logic and control information for Mission and safety-critical applications such as braking systems and anti-skid control require real-time response within a tight timing budget. This timing deadline must be met irrespective of the workload and despite the occurrence of any faults in the system. The worldwide recession, changing consumer habits and shrinking government budgets has forced companies to add more functionality with greater performance at affordable, sometimes lower cost. The shorter design cycles and the growing number of projects handled by the same team require the team to be far more efficient and any failures would cause significant and costly setbacks.

To operate this environment and create successful products, designers must simulate, visualize and explore design alternatives from the very beginning of the development process. Systems engineering tools, such as VisualSim from Mirabilis Design, allows the systems engineer to visualize performance, power and program data flow during the product specification phase.   These tools enable the systems engineer to also evaluate the system quality by injecting system faults, causing software program sequence errors and unavailability of data events. An upcoming Webinar from Mirabilis Design will discuss these emerging design requirements of large complex systems and how simulation can provide a realistic view of the real-time system, very early in the design cycle.