ARIES WP2 Dependability
- how to model and manage dependability aspects in application development involving new and emerging technologies.
Short-term and long-term challenges
- Short term: develop a survey on error detection and recovery, preliminary analysis of decomposition of safety integrity levels, set a framework for dependable communication, determine the main topics for future long term research, base the research on application examples of the automotive industry
- Long term: apply software engineering methodologies to enhance dependability of embedded software-based systems, use research results to support methodology in international standards for functional safety, show techniques for verification and validation, develop techniques for dependable communication, demonstrate applications in several industrial branches
Dependability aspects can be found in all industrial branches. Process automation, machine control, automotive applications and medical systems are examples. Today safety-related real-time systems are ubiquitous in almost every field of the industrial society. Computer-controlled equipment, x-by-wire features, autonomous vehicles, dynamic sensor-actuator systems, and support systems for information exchange and decision-making all enlarge the scope for a variety of safety-critical situations, where severe losses cannot be tolerated.
Dependable software is a key component to provide a dependable software-based system. Faults have to be avoided during the design, coding, validation and maintenance of the software. This is managed all through the software safety life cycle. But faults will nevertheless occur, both in software and in hardware. The system (software and hardware) must also include techniques to detect and handle faults during operation.
As a discipline, software safety can be regarded as an integration of the system safety and reliability methodologies into software engineering. In fact, reliability has been an integral part of software engineering since its inception. This is illustrated by the focusing on fault, error and failures in the search for software defects and in the choice of analysis and verification techniques. Besides, the influence of the reliability approach can be found among the design strategies, e.g., in the incorporation of fault tolerant mechanisms etc. However, with the introduction of system safety to computer systems in the late 20th century, the necessity to interpret the principles of system safety for the software parts also arose. As a result, focus shifted from defects in general, to those representing unsafe constructs, i.e. those with a potential of being exposed to hazards, unsafe states and accidents. Also recognized was that software safety, as a property, has to be designed into the individual component from the outset and, even more importantly, into the system architecture. Accordingly, in order to promote the overall safety of the resulting system, the safety activities needed to address the processes used, as well as the people involved in building, operating and maintaining the safety-critical software system. Thus, the main effort in integrating the three disciplines remains in the incorporation of safety considerations into the established software engineering framework, i.e. the processes, the design principles, the development environment and the tools included. In this work, a seamless integration is important; not only of the safety procedures into the traditional software processes, but also of the safety analysis methods into the design artefacts. This allows the various safety analysis techniques to be applied directly to the different design models and views defined for the system
It is expected that dependable software-based systems will be used in new applications and will be developed based on new emerging technologies. Wireless communication will replace the present communication buses used for safety-related applications. Increasing trust will be placed on even more complex programmable circuits. The re-use of software components will be widespread, and industry will learn how to validate and certify dependable software components for safety-related applications.
The Work Package “Dependability” will address dependability aspects such as:
- wired and wireless communication in dependable distributed systems
- safety validation
- fault detection and handling techniques
The pilot phase of the project will be a stepping stone for the research cooperation. First results and a research plan for the continued work will be presented after the first phase.