Software Architecture of the U-70 Accelerator Complex New Control System

The software of the new control system of the 70GeV accelerator complex is built around the distributed real time DBMS SSUDA with emphasis on control of technological process, not equipment. SSUDA was designed to store current states of dynamic parameters and supports 3-D tables. The tables are distributed around all levels of the CS, including equipment controllers. So, DB access protocols are used to access ECs. There are only two data-oriented atomic objects that applications deal with: vector and structure. All tasks interact exclusively through DBs and each task belongs to one of three weakly related types: data processing, hardware I/O and manmachine interface. The organization of parameter value storage is standardized, so data processing and I/O applications are highly unified. Only the console program serves the MMI for all tasks that are solved by operators and accelerator physicists. 1 GENERAL CONSIDERATIONS The U-70 new controls reasons, aims and plans were presented in [1]. The design study was done in collaboration with CERN [2]. The current state of the whole project is described in [3]. Due to historical reasons and in accordance with project design, one main and three local control rooms were placed in various buildings. Protection facilities are necessary to avoid simultaneous setting of correlated parameter values from different control rooms. As for software design, the basic idea was: the object of controls is a technological process. Inside of the control system the technological cycle may be seen as a set of parameters taking some number of discrete states. The software presents the states of each parameter as a vector of corresponding values. So applications deal with the set of vectors. Each vector corresponds to technological or physical parameter. Technological parameter vectors correspond to states of input/output hardware; physical parameters describe the physics of the technological process or software tune. Therefore the main attention during software design was paid to management of real time dynamic data and to the organization and standardization of applications. We build applications around the real time distributed data bases and separate traditional parts of conventional tasks: dynamic data, device access, data processing and manmachine interface. Figure 1 shows the next peculiarities of the transformation: 1) the user interface is supported by separate console session tasks; 2) data processing is done inside Work Stations; 3) device access is supported only by Equipment Controllers; 4) all data are placed in the Data Bases; 5) intertask communication is done exclusively through the Data Bases. Data processing and device access tasks do not use any graphical facilities. All the software components are data driven.