Search

Collaborative optimization problems can often be modeled as a linear program whose objective function and constraints combine data from several parties. However, important applications of this model (e.g., supply chain planning) involve private data that the parties cannot reveal to each other. Traditional linear programming methods cannot be used in this case. The problem can be solved using cryptographic protocols that compute with private data and preserve data privacy. We present a practical solution using multiparty computation based on secret sharing. The linear programming protocols use a variant of the simplex algorithm and secure computation with fixed-point rational numbers, optimized for this type of application. We present the main protocols as well as performance measurements for an implementation of our solution.

This work introduces a formal analysis of the non-repudiation property for security protocols. Protocols are modelled in the process calculus LYSA, using an extended syntax with annotations. Non-repudiation is verified using a Control Flow Analysis, following the same approach introduced by M. Buchholtz and H. Gao for authentication and freshness analyses. The result is an analysis that can statically check the protocols to predict if they are secure during their execution and which can be fully automated.

In the EYES project, we are investigating self-organizing, collaborative, energy-efficient sensor networks. This study is devoted to the security aspects of the project. Our contribution is three-fold: firstly, we present a survey, where we discuss the dominant issues of energy-security trade-off in the network protocol and key management design space. From there we set out future research directions for our security framework. Secondly, we propose an assessment framework based on system profile, with which we have managed to carve out manageable design spaces from the seemingly infinite possibilities of ad hoc mobile wireless networks. Finally, we have benchmarked some well-known cryptographic algorithms in search for the best compromise in security and energy-efficiency, on a typical sensor node. Our preliminary investigations also cover an important parameter in the design space: the resource requirements of the symmetric key algorithms RC5 and TEA.

Secure perimeter schemes (e.g. DRM) and tracing traitor schemes (e.g. watermarking, audit logging) strive to mitigate the problems of content escaping the control of the rights holder. Secure audit logging records the user's actions on content and enables detection of some forms of tampering with the logs. We implement Schneier and Kelsey's secure audit logging protocol, strengthening the protocol by using tamper-resistant hardware (an iButton) in two ways: Firstly our implementation of the protocol works offline as well as online. Secondly, we use unforgeable time stamps to increase the possibilities of fraud detection. We provide a performance assessment of our implementation to show under which circumstances the protocol is practical to use.

The features of electrotechnical products are increasingly provided by software instead of hardware. Software is the keyfactor in creating products that are ‘better’ than those from competitors, with respect to quality, price, functionality, reliability and performance. During embedded software development, testing products is a major task, like it is for non-embedded software development. Real-life is the best testing environment. Therefore a testing approach widely applied for embedded products is a method termed ‘Field Testing’. This is the release of a small number of products to customers. The objective of a field test is to make sure that faults and failures are reported and taken out of the product, before normal distribution can start. Schlumberger RPS applies field tests to its products for that purpose. The actual number of failures reported from field tests is lower than expected. However, failures are found later that should have been reported from field test. Therefore, the effectiveness of the field tests must be improved. Our research concluded the need to emphasise two aspects. Firstly, the need to motivate people to detect and report failures by assigning responsibility for field test results. Secondly, the need to include functionality in the product to enable automatic detection, recording, and communication of failures.

Quality of software products has been a rather intangible concept for both users and software testers over the years. In spite of interesting results in research and practice, most users can not describe the quality characteristics of the software product that they need. As a consequence, software testers do not know what software quality requirements are desirable or realistic, which makes it almost inpossible to test a software product from a user’s point of view.Determination of software quality characteristics starts by listening to “the user” and his or her quality needs. This can be done by characterization of (a part of) the business situation the user wants to support, using a software product. The attributes by which a business situation can be characterized are clustered into three categories: business process characteristics, user characteristics and software product characteristics.Starting from the various business characteristics the software quality characteristics can be derived in a deterministic way. Software quality characteristics can be expressed in sub-characteristics, which are related to indicators and metrics. These characteristics, initially derived from the users’ quality needs, can then form a solid basis for a user-oriented testing process.Several (European) research projects are gathering empirical data to validate and calibrate the relationships between quality needs, business characteristics, software quality characteristics and metrics. The ultimate objective is to develop a practical basis for third-party testing of software quality. The various concepts will also be implemented as instruments for both developers and software testers to realize, verify and validate software quality.