Ergatic Distributed Data-Processing Systems Control

Abstract

This paper presents a review of ergatic distributed data-processing system (EDDPS) based on network with Centralized computing and informational resources. According to application conditions, EDDPS is able to operate in several modes (regular, testing, training etc.) either with existing database (DB) or without it. There should be manual input to get the final result in the second option. Data processing in EDDPS is done by solving informational-computing task (ICT) including both standalone tasks and tasks consequently solved in technological handling data chain (THDC). Every ICT is a self-realized software product. Automated Workstations (AWS) are provided for data processing in EDDPS, functionally grouped ICTs are being solved on these workstations. The following requirements are imposed to the system: non-interrupted transition process between different modes and minimized data-processing time. EDDPS's efficiency may be ensured by the way of system control, i.e. provide with a capability to respond to control actions in a certain way, specific direction and timeframe. EDDPS control must ensure changing of the state in the shortest period of time to get the circumstances best for ICT solution; minimization data-processing time; steadiness to internal and external perturbations during the all system operation steps. Complex of technological and software resources, informational security and staff is the management object of EDDPS. It's covered by control loop which includes management subject (decision maker - DM) and management tools (technical, software and informational). Such approach enables to solve the time-minimizing problem for operation system both with data processing and with its state changing. EDDPS mode control is connected with adjustments of system parameters characterizing its functionality (e. g. Operating mode, used DB, AWS composition, ICT structure etc.) and must ensure the transformation from one steady state to another. Behavioral model reflecting all possible mode transformations may be used as management foundation. Data processing control is done in a current state in aim of time minimization of the data reduction to increase the speed of system functionality in total. Minimum of time may be achieved for the relevant system class by using the program monitoring function which can be presented by schedule of ICT decisions on chosen AWS. Presented two control modes are independent as their usability may be mistiming - as they are used on the different steps of the system operation. The basis of EDDPS control is the paradigm of several control loops usage each of which realizes method requisite for the exact step. Controlled functionality EDDPS with minimized data processing time is realized by software utilities and informational management support.