Strategic
Being an R&D project, ASSERT has naturally focused its attention to the achievements of technical objectives. This does not mean that the project has to be considered as a technology centered project where the results have only to be judged with respect to their inherent quality. Development and demonstration of a new technology is challenging and attractive for technical oriented people but the value of a project such as ASSERT has also to be thought in terms of business and strategic objectives.
The main reasons why we decided to put this project on motion are due to the context of the project.
The main expected impact of this project is the innovation to be introduced into the system and software engineering process but also, and principally, the introduction of this innovation in operational projects through the support of the space industry. When exploring the set of strategic objectives in the assert world, the following topics have to be considered :
- The main goal is to improve the competitiveness of the aerospace sector and this objective is fully in line with the ESA mission.
- Space systems that will have to be designed in the future will use more and more computers and the current approach in system and software engineering is not adapted to the new challenges (increase of on-board autonomy, in-flight cooperation between multiple vehicles).
- Better designs safely built on safe foundations will reduce the risk of mission failures, the cost of testing and the number of errors discovered late in the life-cycle.
- After a consolidation of the results in the aerospace domain, the project will use the dissemination policy to attract more industrial domains through the A.NET initiative (community of assert-project network).
- This type of project is not an isolated action only targeting the three years duration to introduce innovation. The project vision spans the next fifteen years, calls for major change in the industrial engineering practice.
Technical
“Building safely on safe foundations”.
In many industrial domains, you can trust your designs as they are built on safe foundations : an aircraft wing is built after aerodynamic laws, a bridge after a soil analysis and knowledge of strength of materials. An engineering approach is trustable if it supports :
- The capture of the required properties at system level (mass, maximum and minimum speed of an aircraft).
- An approach to build foundations that will guarantee the required properties (A design approach to build a wing after aerodynamic laws).
- A methodology to define the acceptable range of system components characteristics that ensure the compatibility with the foundations.
- A process to control the design and enforce properties checking at all stages.
In ASSERT, the approach supporting the project completely falls under this paradigm :
- System properties are captured using a system modelling language (AADL in our case).
- The ASSERT Virtual Machine constitutes the foundations that guarantee the properties (timing properties for example). Naturally the foundations (the VM) have to be tailored according to the system needs.
- The component characteristics are used to taylor the Virtual Machine and define the acceptable ranges (Worst-Case Execution times and deadlines of system tasks).
- A process based on the use of model transformation techniques is used to control the design and preserves the properties at all stages.
System in ASSERT references embedded computer system.
For the sake of understanding, and to avoid any misinterpretation of the ASSERT solution, it must be made very clear that the notion of systems in the project is by definition limited to the so-called avionic system. In addition, ASSERT concentrates on software dominant system and does not address the hardware part beyond the need of identifying the computer system parts (Processors, memory, buses).From one single project to a family of systems.
The ASSERT paradigm (Safe designs on safe foundations) is fully applicable on any project but cannot be by itself the sole basis of a new system design approach. In order to boost the productivity of design teams and produce more designs at lower costs, the approach at project level has to be extended to a set of similar systems with the notion of system families. Implementation of the ASSERT paradigm at the level of a family of system requires :- The identification of properties common to a set of systems.
- The definition of a reference architecture that defines the basic components and their interconnections, together with the range of acceptable properties.
- The family requirements on the foundations (in our case, the Virtual Machine).
By having such an approach implemented, a new system will first start by the selection of the system family, the use and tailoring of the reference architecture and the application of the ASSERT process to implement the system.
