Model of real-time information systems for problems with a wide range of input data

Аuthors

Dudakov N. S.^*, Makarov K. V.^**, Putyato S. A.^***

Group of Companies «RTI», 10/1, 8 Marta str., Moscow, 127083, Russia

*e-mail: nikolay.dudakov@gmail.com
**e-mail: kmakarov@oaorti.ru
***e-mail: sputyato@oaorti.ru

Abstract

The purpose of this article is database modelling for real-time special-purpose systems.

It is shown that the use of the all-purpose client-server databases for solving the air force and air defense tasks leads to poor query performance. In some cases, the use of more simple distributed databases allows reducing the query time by 90%, but the amount of data does not allow using them as a single solution.

Given the limitations of the scientific and methodological methods for evaluating the effectiveness of the database, the article offers a method of modeling the database, which allows to parametrize data and evaluate database performance. Database modelling helps to carry out the management structure of the database.

Database modeling is made by mathematical methods of queuing theory, the DBMS performance criteria are based on the average execution time of queries to the database.

This article provides a method for constructing an optimal database structure when using a distributed inhomogeneous database for processing of data with a wide range of characteristics. Restrictions in optimization problem are based on database load. The optimization problem is reduced to the Boolean satisfiability problem — the problem of determining if there exists an interpretation that satisfies a given Boolean formula. The variation of the branch and bound algorithm is used to solve the SAT-problem.

It is shown that the design of real-time database management system based on database modeling techniques can significantly increase the performance of data processing.

This article provides a model of a database management system as a queuing system.

It is shown that the application of optimal control database structure allows you to double the capacity in dealing with real-time heterogeneous tasks.

Keywords:

DBMS design, queuing theory

References

1. Kuznetsov S.D. Osnovy sovremennykh baz dannykh (Basics of modern databases), Moscow, Isd-vo «Internet-universitet informacionnyh tehnologij», 2007, 488 p.

2. Deit K. Dzh. Vvedenie v sistemy baz dannykh (An Introduction to Database Systems), Moscow, Izdatelskij dom «Wilyams», 2001, 1328 p.

3. Kondrashin M.A., Arsenov O.Y., Kozlov I.V. Trudy MAI 2016, no. 89, available at: http://trudymai.ru/eng/published.php?ID=73411

4. Kleinrok L. Teoria massovogo obslugivaniya (Queuing theory), Moscow, Mashinostroenie, 1979, 432 p.

5. Boros, E., Peter L. Hammer. Pseudo-Boolean Optimization. Discrete Applied Mathematics, 2002, vol. 123, pp. 155-225.

6. Kalashnikov V.B. Trudy MAI 2016, no. 89, available at: http://trudymai.ru/eng/published.php?ID=73368

7. Muraviev S., Dvoryankin S., Nasenkov I.I. Otkrytye sistemy. SUBD, 2015, no.1, pp. 22-24.

Download