Upcoming System Modeling Challenges

In the past, during the concept phase of a design, system models have typically lacked information regarding reliability. If at all, reliability was addressed late in the design phase shortly before tape out. As the functional safety aspect, and with it an extended device lifetime, gains more and more attention for certain applications, things have to change in the design processes. To ensure an effective development of safe systems, reliability information has to be considered right from the beginning of the concept phase.

Typical and well known functional safety standards, such as ISO 26262 for automotive systems and IEC 61508 for electric and electronic systems in general, require these changes implicitly. The overlooked and neglected issue is that functional safety for a device is directly linked to the product’s lifetime. Especially in automotive applications, these lifetimes last longer than the two to three years for a typical consumer device. Thus, it is outright important to include available information such as expected life times of the various components for specific application scenarios as early as possible in the design process.

The first gains can be made when using the system models for firmware development before an actual FPGA is available, which resembles the final product’s behavior. Using this approach makes the hardware and software design work in parallel, saving precious time-to-market while having both teams work together on design shortcomings. In contrast to that, the conventional serial approach would produce re-design cycles or possibly awkward fixes. Incorporating reliability of hardware in the firmware design makes the results more robust, which is what is expected in a functional safety design. Additionally, firmware designs can focus on being robust for parts, from which it is known how they degrade over time.

Many sources are available for reliability information, starting from field tests down to simulation of specific material aging simulations. The difficulty is to find relevant reliability aspects for system simulation, especially in such a way that actual simulation time and significance of simulation results lead to a more effective and better design and thus avoid re-design cycles later on.

Summarizing the above, major reliability challenges for system modeling on concept level include:

• Incorporation of reliability information into models
• Extraction of meaningful results from simulation in this context
• Reliability abstraction for concept level

These challenges have to be solved to find answers to questions such as: How can degradation of, e.g. an ADC or DAC have an impact on the whole functional design? What precautions have to be taken to guarantee functional safety under such circumstances during the final product’s lifetime?

Products getting more and more complex do complicate this issue further, so that the impact of reliability effects in a system grows as well. This highlights that addressing reliability during the concept phase will be an increasingly important part of the design process. In summary, reliability can no longer be thought of late, but has to be taken into account as early as possible in the design phase.