Participation & Registration

Participation to this one-day symposium is open to anyone interested and is free of charge.
Registration is now closed

For more information, call Roland Mathijssen:  +31 40 2474720

Location

The symposium will be held at the Embedded Systems Institute on the campus of the Eindhoven University of Technology.

Time

The symposium is on October 4th, 2006, from 9:15 AM to 5:00 PM, with lunch at noon and drinks afterwards.

Program

The program has, in addition to introductory and concluding presentations, an exciting key note speaker and simultaneous presentation tracks where project results will be presented.
 

9:00	:		Registration & coffee
9:30	:		Welcome		Frans Beenker / David Watts
9:45	:		Keynote presentation		John Lefor, European Microsoft Innovation Center (EMIC), Aachen
10:15	:		Project outline		Teun Hendriks / David Watts
10:30	:		Coffee / tea break
10:45	:		Presentation 1		ESI – High Level Model
11:30	:		Presentation 2		TUD – Spectrum based fault diagnosis
12:15	:		Lunch
13:15	:		Parallel session 1 : Presentation 3 & 4		3) DTI – User Model 4) TUD – ELAN
14:00	:		Parallel session 2 : Presentation 5 & 6		5) TU-Twente – SARA 6) TASS – Error Injection
14:45	:		Coffee / tea break
15:00	:		Parallel session 3 : Presentation 7 & 8		7) IMEC – Scheduling Approach 8) Philips Research – INXS
15:45	:		Wrap-up		Teun Hendriks / Jozef Hooman
16:00	:		Drinks, networking

 

Key-note Lecture

John Lefor, European Microsoft Innovation Center (EMIC), Aachen
John Lefor, director of the European Microsoft Innovation Center (EMIC) in Aachen, Germany has more than 30 years of experience in software development, product management, networking and research in both academic and commercial settings.

Since joining Microsoft in 1990, John has worked on a variety of projects and technologies ranging from DOS based toolkits to online conferencing technology used for distance learning in higher education. As a member of the Programmer Productivity Research Center in Microsoft Research he was deeply involved in the development and delivery of many of the tools used by Microsoft product groups for measuring and enhancing the performance and correctness of Microsoft products. Several of these tools have been shipped to customers, most recently in Visual Studio 2005. John was also instrumental in developing and supporting the use of the Phoenix compiler backend technology for higher education and software engineering research through the creation of the Phoenix Academic Program (http://research.microsoft.com/phoenix) and is now devoting his time to commercial and academic collaborative innovation through the European Microsoft Innovation Center.
 

Summary of the Presentations

Below is a summary of the presentations.

1) Embedded Systems Institute - High Level Model
Techniques to introduce awareness in products often assume the existence of a high-level model of the product, which can be used to detect errors and suggest corrections. We will present an approach to define such a high-level model in the TV domain.

2) TU Delft - Spectrum-based Fault Diagnosis of Embedded Code
Efficient recovery from software failures requires finding the components which contain the faults that cause them. A prerequisite for diagnosing faults in a program is the acquisition of data about the system's dynamic behavior. Examples of helpful data are traces of events such as basic block execution or function calls. However, due to the constraints of embedded systems, full tracing is unfeasible. We therefore use a compact representation of a trace, called program spectrum. In this presentation we show how we use program spectra to automatically detect faulty components.

3) DTI - User model and near-real life laboratory
Within DTI the Trader activities initially concentrated on prioritizing field reliability issues on field feedback information. Since the feedback currently available in industry is structurally unsuited for understanding root-causes of field-failures DTI has adopted, within the Trader context, a more experimental approach where products, users and failures are brought together in an near-real life laboratory context. The first, successful, results of this approach will be presented.

4) TU Delft - ELAN
Static checking tools are useful as an automated software inspection step that can easily be integrated in the development cycle and assist in creating high quality, reliable and secure code. However, an often quoted disadvantage of these tools is that they generate an overly large number of warnings, including many false positives due to their approximate analysis techniques, which effectively limits their usefulness for large scale software development. In this talk, we present ELAN, a technique that helps the user prioritize the information generated by a software inspection tool, based on computing the likelihood that execution reaches the locations for which warnings are reported, thereby allowing developers to focus on the most pressing issues first.

5) TU Twente - software architecture reliability analysis approach (SARA)
Several useful conventional reliability engineering techniques exist in literature but traditionally these have primarily addressed failures in hardware components and usually assume the availability of a running system. On the other hand current software architecture analysis methods show the importance of early prediction of quality before a system is implemented. We propose a software architecture reliability analysis approach (SARA) that benefits from both reliability engineering and scenario-based software architecture analysis to provide an early reliability analysis of the software architecture.

6) Philips TASS - Error Injection
Fault/error injection is an attractive approach to the experimental dependability validation of fault-tolerant systems, as it provides the means for a detailed study of the complex interaction between faults, errors, failures and fault-handling mechanisms.
The confidence in the fault tolerant mechanisms used increases when these mechanisms are triggered during different operational activities. Fault activation is (or should be) a rare event for software of a certain quality and not all faults are activated during the limited normal testing time. Instead of waiting for a fault to be activated, it is beneficial to be able to trigger the fault activation and the error occurrence at will. Error injection uncovers potential fault tolerance deficiencies and also gives insight in the system behavior and sensitivity.

7) IMEC - design-time/run-time combined scheduling approach
To bring an effective timing-error inspection and prevention to real-time systems, a fine-granularity application-specific task scheduler is required. Due to the ever increasing dynamic behaviors of applications, traditional static scheduling based on the WCET (worst case execution time) becomes too pessimistic; while the pure dynamic scheduling at runtime often gives sub-optimal results because of prohibitive computation cost to achieve optimal solutions. We will present a design-time/run-time combined scheduling approach based on Pareto-optimal solutions pruning and selection, and show its applicability in the real-life application.

8) Philips Research - INXS
INXS is a methodology and tool to add aspect oriented programming to the television software development tool Koala. Thus tracing, error detection, and error recovery can be added relatively easily with INXS to the existing SDE4 software stack. INXS is currently used in the tv520 project to determine how the applications and middleware access the nexperia-platform. In the presentation, we would like to describe our results of this pilot project.