1. Most Significant Research Contributions
1.1. Quality of service management by applications
Within the CITR project, my PhD student A. Hafid and I, in collaboration with Prof. Kerhervé and the other members of our team, developed a framework for the management of quality of service (QoS) at the application level. In contrast to most work on QoS which considers the network or transport level, we concentrated on the issues which must be dealt with at the level of the applications, in particular considering distributed multimedia applications (see publications J1, J2, J3, J4 and many conference publications). Although this was a new area of research for us, our work was well accepted and recognized by the research community, as indicated by an invited paper at the 1996 QoS workshop in Paris and my invitation to program committees of several international conferences in this area.
1.2. Test development for non-deterministic systems
Since our seminal work on protocol testing based on FSM specifications in the early eighties (see [Sari 82, Sari 84, Sari 87a]), much work has been done in this area assuming fully specified deterministic specifications. In 1990 (see [Fuji 91c]), I started to work with Fujiwara on the testing in respect to non-deterministic specifications. Further work was done by Dr. Luo under my supervision and in collaboration with A. Petrenko [J12]. The adaptation of these methods to the testing of labelled transition systems is the topic of the submitted PhD thesis of Tan. This work could have many applications in practice, since specifications often have some nondeterministic features. All our work on protocol testing, including the testing in respect to partial specifications which is mainly due to Petrenko, has been the leading edge in the community of protocol testing. It had a clear impact on the direction of work within this community.
An indication of our leadership is given by the following paper count: In the international IFIP Workshop on Protocol Test Systems held 1993 in Pau, France, one invited paper out of 4, and 4 contributed papers out of 17 (selected among 37 submitted) were authored by members of my group, 3 were co-authored by myself. The situation was similar in 1994 when the conference was organized in Japan.
1.3. Tutorial on protocol and software testing
Being a leading researcher in the area of protocol engineering, I have often been sollicited to present invited papers or tutorials at conferences, some of which also resulted in journal publications (e.g. J1, J5, Favr 91b, Boch 90g). Recently I was invited to present a longer tutorial on protocol and software testing at the Brazilian Computer Networks Symposium in May 1997. In this case, I prepared a detailed text and many slides for presentation (with some contribution from R. Dssouli). I think that this tutorial provides a good overview of the main issues, methods and techniques used for the testing of reactive systems (including protocols and real-time control systems), and puts to perspective the different research results that have been obtained over the last couple of years. The tutorial was well received by the audience, and a similar tutorial, based on the same documentation, was given by R. Dssouli at the SDL Forum conference in Paris and the CFIP symposium in Liège later this year.
1.4. Protocol derivation from a given service specification
We consider here the automatic derivation, from a given specification of a distributed service, of the specification of a protocol which provides the given service. This approach was first introduced by my work with Gotzhein in 1986 (see [Boch 86c, Gotz 90]). More recently, I worked with my PhD student Kant on this topic, extending the power of the service specification language to basic LOTOS [J7]. I also worked with the post-doctoral researcher Khoumsi on the inclusion of real-time constraints [J11]. Although our work in this area has been much quoted in the literature, it is not clear how useful the automatic derivation of protocols really is for practical applications.
1.5. Object-oriented analysis and design of distributed systems
This is not really a single contribution, but a theme underlying many of my research activities. It is also the theme of the collaborative IGLOO project (see Section 2.7 below) which I coordinated during four years. I would hesitate to mention a particular important contribution in this area, but I think that all my activities related to this theme, overall, represent a significant contribution. In this context, I would like to mention (a) the publication [J8] which is based on work done under contract for BNR, (b) some work on object-oriented network management done under the IGLOO project which led to subsequent projects of CRIM with Positron and Eicon, (c) my non-negligeable contribution to Dini’s PhD thesis on reconfiguration management in distributed systems, and (d) a comparison of object-oriented specification formalisms [Boch 94g]. I think that the impact of my work in this area has been mainly in the local research community and in technology transfer.