(The following is an extract from a text which was prepared in 1994)
Gregor Bochmann was one of the first to recognize the importance of precise and formal specifications for the analysis and logical verification, the semi-automatic development of implementations and the systematic testing of communication protocols. More recently, this general area has been termed "protocol engineering" [Boch 93c]. Since the beginning of the research in this area, about 18 years ago, Prof. Bochmann has initiated several new directions for the research in this area and developed several methods and tools which are not only used for protocol engineering, but also more generally for software engineering of distributed systems and discrete event control systems. He has become an (if not the) internationally recognized leader in the area of protocol engineering.
More specifically, Bochmann has proposed an "extended state machine model" for the specification of communication protocols [Boch 77c], [Boch 80e] which is based on the combination of finite state machines with variables and statements of a programming language, which has found much application. In particular, it has been used as the semantic basis for the standardized specification languages Estelle and SDL, so-called Formal Description Techniques (FDT's) that have been developed by ISO and CCITT, respectively, for the description of communications protocols and services. Bochmann has participated, as a delegate from Canada, in the development of these standards, and his contribution to the development of Estelle has been publicly acknowledged in 1989 when Estelle became an International Standard.
Bochmann was also the first to propose the systematic testing of protocol implementations based on the formal specification of the protocol, in terms of an FSM [Sari 82], [Sari 84] and an extended FSM [Sari 87a] model. These germinal papers lead to many other research results in the area of protocol conformance testing. More recently, Bochmann and his colleagues were among the first to extend these results to partially defined and nondeterministic specifications [Yao 94a], [Luo 93h] and to stress the importance of a precise fault model for the estimation of the coverage of a given test suite [Boch 91d].
Explanation
The systematic verification of the design of the protocols to be used in a distributed system is very important for obtaining reliable systems. Because of the concurrency within the distributed system, the verification is much more difficult than for sequential computing applications. The use of formal description techniques, such as the FDT's mentioned above, allows the use of automated tools for the validation of the system design. Several industrial tools have been developed for the validation of specifications and the semi-automatic development of implementations.
The formal specification of the protocol may also be used for the automatic or semi-automatic derivation of a test suite which may be used for checking that a given implementation conforms to the specification. Conformance testing is particularly important for protocols because independently developed implementations are expected to be able to interwork.
The contributions made by Bochmann are very substantial, because he contributed directly to the development of an important new research area, namely protocol engineering, and made several key contributions that oriented the developments in this area. His work not only contributed to the theoretical understanding of the area, but also to the development of standard specification languages, methods and tools that are now being used in industry for the development of designs and implementation of distributed computer and communications systems.
[Boch 80e] G. v. Bochmann, A General Transition Model for Protocols and Communication Services, IEEE Trans. Comm., COM-28, 4 (April 1980), pp. 643-650, reprinted in "Communication Protocol Modeling", edited by C. Sunshine, Artech House Publ., 1981.
[Boch 93c] G. v. Bochmann, Protocol Engineering, contribution to Concise Encyclopedia of Software Engineering, Derrick Morris and Boris Tamm eds., Pergamon Press, 1993, pp. 266-271.
[Boch 91d] G. v. Bochmann, A. Das, R. Dssouli, M. Dubuc, A. Ghedamsi and G. Luo, Fault models in testing, Proc. IFIP Intern. Workshop on Protocol Test Systems, Netherlands, Oct. 1991 (invited paper), pp. (II-17)-(II-32).
[Boch 77c] G. v. Bochmann and J. Gecsei, A
unified method for the specification and verification of
protocols, Proc. IFIP Congress 1977, pp. 229-234.
[Luo 93h] G. Luo, G. v. Bochmann and A. Petrenko, Test selection based on communicating nondeterministic finite state machines using a generalized Wp-method, to appear in IEEE Tr. on SE.
[Sari 82] B. Sarikaya and G. v. Bochmann, Some experience with test sequence generation for protocols, Proc. 2-nd Int. Workshop on Protocol Specification, Testing and Verification, North Holland, 1982, pp. 555-567.
[Sari 84] B. Sarikaya and G. v. Bochmann, Synchronization and specification issues in protocol testing, IEEE Trans. on Comm., COM-32, No.4 (April 1984), pp. 389-395; russian translation: Express Information (overview of western publications), Information Transfer, 1985, no. 28.
[Sari 87a] B. Sarikaya, G. v. Bochmann and E. Cerny, A Test Design Methodology for Protocol Testing, IEEE Trans. on Soft. Eng., Vol.13, no.5, May 1987, pp. 518-531.
[Yao 94a] M. Yao, A. Petrenko and G. v. Bochmann, Fault coverage analysis in respect to an FSM specification, IEEE INFOCOM'94, Toronto, Canada, June 1994.
(1) Together with P. Merlin, Bochmann has shown in 1980 that it is possible to determine the specification of an unknown system submodule that, together with other known submodules, will satisfy given system requirements. Initially proposed for the derivation of a protocol specification, this method has recently been found useful for the construction of protocol converters and control modules for process control.
(2) Significant contribution to the state-of-the-art in protocol conversion. The following paper is the first to clearly separate the conversion issues related to incompatible communication services and incompatible protocols.
(3) Pioneering work on the use of temporal logic for hardware design. The following paper has been extensively cited:
(4) Pioneering work on the evaluation of semantic attributes which is important for the construction of compilers. The following paper has been extensively cited:
(5) Bochmann's group was among the first to develop methods for test selection with guaranteed fault coverage in respect to FSM specifications that are partially specified or nondeterministic, as documented in the following publications:
Between 1976 and 1986, Bochmann did many studies under contract for the Department of Communications, Canada. The study entitled "OSI Support in Canada: Policies and Implementation Perspective", prepared in collaboration by Bochmann's consulting firm, CERBO Informatique Inc., and PSC Inc. of Ottawa in 1986, gave the direction for the development of an OSI protocol conformance test center in Canada. This idea was taken up around 1988 by the Canadian Interest Group for OSI (CIGOS), and finally lead to the establishment of the Protocol Test Center of Hewlett-Packard in Montreal in 1992.
Bochmann has taken much leadership in the area of university-industry collaboration by participating in and/or leading important research projects under contract for industry or government, and by participating in the process of creating such organizations as the Centre de Recherche Informatique de Montreal (CRIM) and the Canadian Institute for Telecommunications Research (CITR), one of the Canadian "networks of centers of excellence". He is presently one of the scientific directors of CRIM. The mandate of CRIM is to promote joint research by university and industry, to facilitate technology transfer, and to contribute to the training of highly qualified computer scientists and engineers.
Bochmann holds the IDACOM-NSERC-CWARC Industrial Research Chair on Communications Protocols at the University of Montreal, which is financed by the IDACOM/Hewlett-Packard Protocol Test Center and the governmental Canadian Workplace Automation Research Center. He is also a project leader within CITR and the scientific leader of a large research project on object-oriented modeling of distributed systems and network management, called IGLOO, which involves CRIM, three universities and six industrial partners.