V.N. Negoda - Dr. Sci. (Eng.), professor at computer science chair, Ulyanovsk State Technical University. E-mail: nvn@ulstu.ru
A.V. Kurepin - master of computer science, head of department of telecommunication, Ulyanovsk State Technical University. E-mail: a.kurepin@ulstu.ru

When designing a distributed technological control system, design solutions for the communicative subsystem and ones for control subsystem must be coordinated with each other. The same applies for simulation of design solutions. Technologies for these are known as co-design and co-simulation respectively. Binding these techniques into a single design process is very efficient. In co-simulation, a network modeling system (e.g. ns-3) integrates with a formal modeling system used for control processes simulation (e.g. Matlab). In this paper we consider a distributed system node as a set of controller, managed units and one shared or multiple separate networks. Control cycle time estimation is the sum of ten delays of different nature. During modeling, several design solutions are born, one of which is believed to be optimal, while others may be used in the future under different conditions of the control system application. To simulate individual components of a system one may use fragments of network driver source code, distributed system hardware, and design patterns, which are used for automated code generation. Co-design and co-simulation of communicative and control processes are based on state-machine models. During design process some design solutions are transformed. The major transformations are the wait-phase decomposition, the decomposition of a computationally intensive process, adding of composite state, combining of packet reception and parsing for different target devices. In the paper an example of the application of this approach is considered. As a result, network delays were significantly reduced and number of lost commands is reduced more than twice.