© 2015 J.B. Lewoc et al., licensee De Gruyter Open. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License. The article is published with open access at www.degruyter.com. Open Comput. Sci. 2015; 5:51–59 Research Article Open Access Józef B. Lewoc*, Antoni Izworski, Antonina Kieleczawa, Marion A. Hersh, Nicolae Bulz, and Erich Leitgeb Exemplary industrial system and network modeling in Poland DOI 10.1515/comp-2015-0005 Received November 14, 2011; accepted July 21, 2015 Abstract: The paper presents the exemplary industrial system models from the event-driven simulator up to A pproximate n etwork a nalytical p erformance e valuation s ystem t ool (Anapest). The basic approximations assumed in Anapest are roughly discussed together with the basic network configuration for which the tool was developed. The two basic ways of Anapest validation, i.e. the inter- nal network measuring tool Sitwa and event-driven simu- lation are depicted. In addition, the major network perfor- mance evaluation projects run till now described, together with the basic outcome from Anapest. Keywords: computer network; industrial; performance evaluation; performance prediction; tool 1 Introduction 1.1 The very beginning The applications of real-time computers in automation in Poland (the case-study country) began in late sixties, from the steel rod mill manufacturing control and coordination project [1]. The system (Ref. Figure 1) monitored and con- trolled the steel rod traffic characteristics, as well as co- ordinated the manufacturing process in accordance with short-term production plans. The solution was novel not only in the case-study country. The system designers had to cope with the two very severe local problems: – any detailed technical information on similar solu- tions known in e.g. well-developed countries was very hardly available (a result of so called cold war), *Corresponding Author: Józef B. Lewoc: Design, Research and Translation Agency Leader Wroclaw, Poland, E-mail: leader@provider.pl Antoni Izworski: Wroclaw University of Technology Antonina Kieleczawa: Instytut Automatyki Systemów Energety- cznych Sp. z o. o., Wrocław, Poland Marion A. Hersh: University of Glasgow, Scotland Nicolae Bulz: Romanian Academy of Science Erich Leitgeb: Graz University of Technology, Austria – the computer hardware and software available (in this specific case, the Polish Odra 1204 computer of a very nice logical design but slower roughly of one rank of order than computers available then to designers in Western countries) were obsolete of some 5-10 years (a result of unreasonable embargos: the Red Army had no problems with providing their missiles and fight- ers with the world most modern electronics while the designers working for civilian applications could not buy a good computer). This situation that was continued almost till the political change in Poland in 1989 exerted major impacts on per- formance evaluation in the case-study country: the per- formance problems were more difficult and important for actual industrial design since the available hardware and basic software were obsolete in comparison with those available in the West while similar applications implied similar performance requirements, and had to be solved autonomously since even if any actually useful informa- tion on performance were available, it would not be, ob- viously, disclosed to the industrial designers of the case- study country. And such was the case for the steel rod-mill controls system; a co-designer of the system did a thorough sur- vey of available queuing theory methods and found no one appropriate for their industrial problems even in the most renown references [2–5]. Therefore, he developed an Event-Driven Simulator (EDS) of the central computer based control system [6] and carried out the necessary sim- ulation runs. All questions interesting the two designers were answered positively (in particular, the impossibility to lose any photo-relay signals was confirmed). Moreover, EDS enabled for the designers to understand much better their own solutions from the viewpoint of performance. 1.2 Another Simulation Study of Performance The steel rod mill control system could be developed with use of two priority levels only: background and fore- ground, but this may be infeasible for other centralized control systems. Therefore, the Lead ing Designer (Leader) developed the other simulation algorithms needed for in- Unauthenticated Download Date | 7/28/18 10:33 PM