Topics RTiS'03

Software testing is challenging. A recent study by NIST has indicated that up to 80% of the life cycle cost for software is spent on testing and debugging. The increasing complexity of software, along with a decreasing average product life expectancy, has increased the economic costs of errors. Elevating to the level of embedded real-time software testing, the challenge is even greater. Real-time software is usually built on an aggregate of multiple concurrently executing programs, i.e., it is multi-tasking. To begin with, this entails testing of
multiple programs. What is worse however, is the state space explosion that occurs due to the interactions between the tasks when they execute concurrently. These interactions are not limited to the functional domain but are also a function of the timing and the ordering of the tasks' execution in the system. The majority of current testing and debugging techniques have been developed for solitary (non real-time) programs. These techniques are not directly applicable to real-time systems, since they disregard issues of timing and concurrency. In
this session we will have two presentations from the industry, one by Sigrid Eldh from Ericsson, and one by Mikael Adenmark from Scania.

Topic Programme:

Ericsson Software Improvement through Software Quality Ranking by Sigrid Eldh

Automation of Integration Tests by Mikael Adenmark

The topic is focused on the intersection between real-time embedded computing and control engineering. One focus of the session is codesign of flexible real-time control systems. Integrating control theory and real-time scheduling theory it is possible to achieve higher resource utilization and better control performance. The integration requires new tools for analysis, design, and implementation, something which now is starting to emerge.

The second part of the topic is the use of feedback principles in the design of real-time computing systems. Applications of real-time computing have gradually evolved from closed embedded systems to complex, distributed open heterogeneous platforms operating in unpredictable poorly modeled environments such as, e.g., the Internet. Hard guarantees are impractical on such platforms since load and resource capacities are very difficult to predict. Using control-based approaches for modeling, analysis, and design of real-time computing systems is currently receiving increased attention as a promising foundation for controlling the uncertainty in large and complex real-time systems.

Topic Programme :

Co-design of Resource-Constrained Embedded Control Systems by Karl-Erik Årzén

Co-Design of Real-Time Control Systems: The Control Server Approach by Anton Cervin

On Real-Time Control with Limited Feedback Communication Network Embedded Control by Karl Henrik Johansson

An overview of Quality of Control by Martin Sanfridson

Embedded systems are imperative for applications such as cars, telecom, aircrafts, and robotics. These systems are becoming increasingly more complex due to the addition of new and advanced functionality. At the same time there are strong requirements to reduce cost, increase quality and shorten time-to-market. When developing and maintaining these systems we have a lot of hinders but also means to develop cost efficient systems which fulfill the customers’ requirements. In this session we will have two presentations from industry where the first one will be about requirements engineering and especially focus on the elicitation of
customer requirements and the second about model based development of complex telecom systems in large and distributed organizations. The presentation about requirements engineering will be held by Torkel Finnström from SAAB Bofors Dynamics AB. Torkel has many years experience from requirements engineering in complex applications. Peter Cigéhn from TietoEnator Telecom & Media will give the second presentation about model based development. Peter is appointed senior specialist in software architecture for UMTS radio basestations, and has 10 years experience from software engineering of complex telecom systems.

Topic Programme :

Model Based Development of Large Embedded Systems in Distributed Organisations by Peter Cigéhn

Requirements Engineering by Torkel Finnström

Component-based development (CBD) has achieved remarkable popularity in software engineering community and considerable results in certain domains. CBD, and software reuse in general, has been extensively used for several years in desktop environments, graphical mathematical applications, Internet- and web-based distributed applications. On the other hand the extra-functional characteristics and constraints are not of the highest priority. In many other domains, such as real-time and embedded systems CBD is considerably less utilized, mainly because inability of component-based technologies to manage extra-functional properties. CBD approach is as attractive in this domain as for other domains, but is less
feasible as it cannot provide accurate solutions for extra-functional requirements. In the first case it is the question of real-time properties, but also others, such as memory or power consumption, reliability, safety. The main challange of CBD approach for RT and embedded systems are developing component models that will enable predictable composition, i.e. that will make possible predict temporal and other system properties from the component specifications and specified constrains.

Topic Programme :

Component Based Software Engineering by Johan Fredriksson

Presentation of component-based software development based on the Rubus concept by John Lundbäck

Towrds Timed Probabilistic Reasoning for Fault Tolerant UML2 Architectures by Heinz Schmidt