Domain-specific Drivers and Limits for Avionics Architectures – A Critical Review in the Context of the Avionics Application Domains

Aerospace once initiated the development of digital computing. Avionics have long confirmed their key role in modern air vehicles. Standardized, resource sharing, real-time computing platforms such as Integrated Modular Avionics are state-of-the-art. Almost no aviation function can be implemented without its digital computing capacities and no development, safety assessment, security analysis, and cost calculations can be performed without considering the avionics architecture as a central element. Thus, these architectures are one of the key drivers for aircraft systems, software, and tool development.Today, avionics seems to be a more or less unfortunate and incapacitated consumer of regular IT. The trend for driving the development of avionics architectures could be summarized as: “Adopt the general IT trends with a delay of one or two decades.”, i.e. increase computing power, memory capacity, network bandwidth, and hardware and software decoupling by deploying new technologies. An inevitable dependency of avionics on the available technologies is obvious. Today’s avionics architectures still seem to scale (sometimes slowly) with the global IT trend, but in detail there are drivers and limits, which make a long-term continuation at least questionable. For example, processors improve the average performance with methods that commonly contradict determinism and security, which are both mandatory for avionics. Containerization enables modularity of functionality as microservices and reduced independence from hardware and software, but this independence collapses at certification, because only the integrated cyber-physical system can be certified. It is of major importance to rework what drives, but also limits avionics architectures. For the future, it might be reasonable to establish self-carrying aerospace sub-branches for some IT methods, tools, and products dedicated to the needs and domain-specific drivers of avionics architectures.The contribution of this paper is threefold: (1) We evaluate past and current architectures in avionics and other domains. (2) We identify the domain-specific external and internal drivers as well as limits and derived requirements for avionics architectures. (3) We classify their significance regarding the application domain within avionics.