DonNTU   Masters' portal

Abstract

Content

Introduction

Computers were created for solving computational problems, but over time they have been increasingly used for the construction document production systems, and more specifically, the information contained therein. Such systems are commonly referred to as informational.

Information Systems require the creation of a computer memory dynamically updated model of the outside world using a single repository – a database. [1]

In addition to information networks are essential dynamical systems. The dynamical system is any object or process in which there are definitely some changes in the state focused processes, which are considered as a change in the total value at any given time, which is accompanied by changes in the parameters, conditions for a certain period of time for which the law is given, describing the change in the initial state with over time [2].

Using the subsystem databases in a distributed parallel simulation environment can positively affect the performance of the system as a whole, increasing the speed, reliability and security.

1. Theme urgency

One of the most important tools for prediction and analysis in various fields of human activity is a computer simulation of complex dynamic systems. Using it to assess the effectiveness of the changes applied to the system, as well as to anticipate the possible implications of these changes. The results obtained in the simulation, enable professionals to make decisions on the possible optimization of system parameters in order to improve its efficiency and reliability.

Formation of database management systems (DPSE) has coincided with significant advances in technology of distributed computing and parallel processing. As a result of any database management subsystem, consisting of parallel systems. It is these systems are becoming the dominant tool for creating data-intensive applications. [3].

Thus, the current is a comprehensive computer support this process – the development of sub-system databases in a distributed parallel simulation environment (DPSE), which will significantly increase the efficiency of the processing of the data array.

2. Goal and tasks of the research

The purpose of this master's thesis is the design of optimal subsystem databases in a distributed parallel simulation environment.

We must perform the following tasks to achieve this goal:

  1. Consideration of the characteristics of decomposition of DPSE
  2. Determination of structural and functional features of the database subsystem
  3. Analysis of existing database structures
  4. Just develop the database subsystem of DPSE
  5. Development of interfaces for integration with other subsystems in DPSE

3. Expected scientific novelty

Scientific novelty lies in the fact that as a result of this work is expected to receive about optimal database structure that will meet the previously described criteria.

4. Review of research and development on

These problems are not new and has been studied for several years in many developed countries. This concept has been described in detail in a report by ASIM-symposium in 1994 [4] and further developed in the works of such scientific figures as Anoprienko A., Feldman L., Svyatniy V., Braunl T., Reuter A, Zeitz M. and others.

In work "Universal modeling environment" Anoprienko A., Svyatniy V. [12] resulted in a complete description of the concept of universal simulation environment (UME), as well as ways to ensure its universality. Were characterized by the basic components of such media, consisting of hardware and software, as well as the important features required of the UME.

With these issues were also involved in Moldovanova A., Solonin A. Their work reveals the peculiarities of parallel simulation environments.

In the direction of the subsystem databases DPSE engaged in Donetsk National Technical University Masters: Shilo A., Navoev A., Merenkov A., Musenko E., Melnikov A. and others. In their works, they considered and defined the basic concept of the functionality of the database in the DPSE and its infological structure, proposed a new approach for structuring data in the database and the use of modern database systems.

Quite informative source is the master work Musenko E. "Development of the subsystem databases in a distributed parallel simulation environment" [7]. It is painted in detail the structure of DPSE in general and the role of the subsystem databases in it. Special attention is also paid to the main features that should have effective subsystem databases.

5. Development of the subsystem databases DPSE

5.1 Description of the distributed parallel simulation environment (DPSE)

The distributed parallel simulation environment (DPSE) – is a systematic organization of the operation of parallel hardware resources, systems and modeling software that supports all phases of design, implementation and application of parallel models of CDS in accordance with certain requirements [5,9].

A characteristic feature is the ability to perform DPSE calculating individual parts of the object simultaneously and independently from each other, i.e. parallel. Parallelization can significantly speed up the process of modeling. Hardware for this act distributed systems: computer clusters, Grid, parallel structures using GPU [6].

5.2 Decomposition of the DPSE

DPSE is divided into 10 main subsystems:

  1. Dialogue subsystem (DS) is used to display the presentation of features and capabilities DPSE. It provides a dialog system, as well as coordination of tasks simulating, planning and management.
  2. Topological analysis subsystem (TAS) provides the internal topology of a dynamical system, its representation in the internal format conversion topology information in the form convenient for the generation of a system of equations, the output of the analysis given topology. Performs verbal and graphic description of the encoding for the primary topologies.
  3. Subsystem of generation of equations communicates with the subsystem of topological analysis to obtain the coded object topology, the transformation of the results of its work in vector-matrix form.
  4. The subsystem of virtual parallel simulation models provides tools to interactively display the hierarchy of the virtual parallel simulation models, depending on the options for parallelization.
  5. Subsystem of a parallel solution of the equations – it is in this system produces a solution of systems of equations using parallel libraries, the definition of convergence, stability, accuracy, optimization of variable parameters, as well as the results of the conversion solution for visualization.
  6. Communication subsystem includes a list of components and resources DPSE. Performs data exchange at the request of the components, the mapping of data flows initiated by the user. Ability to optimize a parallel program with exchange operations.
  7. The load balancing system the definition and management of workloads between virtual processes and processors.
  8. Visualization subsystem prepares visualization of the simulation results, displays dynamic charts during and after the simulation.
  9. The database subsystem interacts with all the subsystems of the DPSE and stores the data for each phase of modeling, as well as information about users, their queries and problems.
  10. IT-support subsystem is remotes WEB-based application for modeling

5.3 Subsystem DB as DPSE

In this paper the database subsystem, one of the most important in the DPSE. It should have the ability to store and quickly access the information that is related to DPSE. This information includes:

An important function of the subsystem is the ability to archive and compress the data on the results of [8].

When choosing a software platform for the implementation of the subsystem database it sets the following requirements:

6. Scalable Data Storages

Recently, interest in cloud architectures is growing every day.

In-memory-data-grid – this is a cluster key-value store, which is designed for heavy-duty projects with large amounts of data and increased requirements for scalability, speed and reliability. The main parts are the IMDG caches.

Cache in the IMDG – a distributed associative array, providing fast concurrent access to data from any node in the cluster.

Distributed cache in the cluster

Picture 1 – Distributed cache in the cluster

Near-cache – is a local object cache for quick access, all the objects in it are kept ready for use. If the near-cache for this cache is configured, the objects get there automatically when you first get-request these objects.

Distributed and local cache memory in the cluster

Picture 2 – Distributed and local cache memory in the cluster

Thus, the technology of In-memory-data-grid allows you to raise system performance to a new level [11].

Findings

As a result of this work, which has, in general, exploratory in nature, we can confidently say that the subsystem database plays an important role in the integrated functioning of the DPSE, because it handles all of the data that must be processed and analyzed. Large volumes of data require effective organization of the exchange, storage and processing of data to ensure fast and safe operation of the system. And it is the use of new technologies, which will take into account the properties of both the software and the hardware is the actual current direction of development DPSE.

Thus, the research and development of this subsystem – this is another step in the development and DPSE and systems modeling in general.

This master's work is not completed yet. Final completion: December 2013. The full text of the work and materials on the topic can be obtained from the author or his head after this date.

List of sources

  1. Зеленков Ю.А. Введение в базы данных [Электронный ресурс]. — Режим доступа: http://www.mstu.edu.ru/study/materials/zelenkov/toc.html, свободный. — Загл. с экрана.
  2. Feldmann L.P., Svjatnyj V.A., Resch M., Zeitz M.: Forschungsgebiet: parallele Simulationstechnik [Электронный ресурс]. — Режим доступа: http://www.nbuv.gov.ua/portal/natural/Npdntu/Pm/2008/08flpfps.pdf, свободный. — Загл. с экрана.
  3. М. Тамер Оззу, Патрик Валдуриз – журнал Системы Управления Базами Данных # 4/1996, издательский дом «Открытые системы» Новая редакция: Сергей Кузнецов, 2009г.
  4. Святний В.А. Паралельне моделювання складних динамічних систем // Моделирование — 2006: Международная конференция. Киев, 2006 г. — Киев, 2006. — С. 83–90.
  5. Абрамов Ф.А., Фельдман Л.П., Святный В.А. Моделирование динамических процессов рудничное Аэрология.– К.: Наук. Мысль, 1981.-284с.
  6. Мирошниченко К.В. Разработка и исследование подсистемы топологического анализа сетевых динамических систем как объектов моделирования [Электронный ресурс]. — Режим доступа: http://masters.donntu.ru/2011/fknt/myroshnychenko/diss/index.htm, свободный. — Загл. с экрана.
  7. Мусенко Е.А. Разработка подсистемы баз данных в распределенной параллельной моделирующей среде [Электронный ресурс]. — Режим доступа: http://masters.donntu.ru/2012/fknt/musenko/diss/index.htm, свободный. — Загл. с экрана.
  8. Меренков А.В.Разработка и организация подсистемы баз данных распределенной параллельной моделирующей среды (РПМС) [Электронный ресурс]. — Режим доступа: http://masters.donntu.ru/2010/fknt/merenkov/diss/index.htm, свободный. — Загл. с экрана.
  9. Святный В.А. Моделирование аэрогазодинамических процессов и разработка систем управления проветривание систем шахт. Докт. дисс, Донецк, ДПИ, 1986
  10. Мельников А.И. Подсистема обмена данными в распределенной параллельной моделирующей среде [Электронный ресурс]. — Режим доступа: http://masters.donntu.ru/2012/fknt/melnikov/diss/index.htm, свободный. — Загл. с экрана.
  11. In-memory-data-grid. Масштабируемые хранилища данных [Электронный ресурс] / Хабрахабр. — Режим доступа: http://habrahabr.ru/post/126580/, свободный. — Загл. с экрана.
  12. Аноприенко А.Я., Святный В.А. Универсальные моделирующие среды // Сборник трудов факультета вычислительной техники и информатики. Вып.1. – Донецк: ДонГТУ. – 1996. С.8-23.
  13. Фельдман Л.П., Назарова И.А. Оценка масштабируемости параллельных вычислений для одношаговых многоточечных методов решения задачи Коши [Электронный ресурс]. — Режим доступа: http://ea.donntu.ru:8080/jspui/handle/123456789/8337, свободный. — Загл. с экрана.
  14. Святний, В.А., Молдованова, О.В., Чут, А.М. Стан розробок та перспективи інтеграції паралельних моделюючих середовищ з Grid-технологіями [Электронный ресурс]. — Режим доступа: http://ea.donntu.ru:8080/jspui/handle/123456789/7339, свободный. — Загл. с экрана.
  15. Солонин А.Н. Подсистема обмена данными в распределенной параллельной моделирующей среде [Электронный ресурс]. — Режим доступа: http://ea.donntu.ru:8080/jspui/handle/123456789/1584, свободный. — Загл. с экрана.
  16. Надеев Д.В. Балансирование загрузки ресурсов в распределенной параллельной моделирующей среде [Электронный ресурс]. — Режим доступа: http://ea.donntu.ru:8080/jspui/handle/123456789/17429, свободный. — Загл. с экрана.
  17. Шило А.В. Разработка подсистемы баз данных распределенной параллельной моделирующей среды [Электронный ресурс]. — Режим доступа: http://masters.donntu.ru/2011/fknt/shilo/diss/index.htm, свободный. — Загл. с экрана.
  18. Навоев А.С. Разработка и исследования средств оброботки массивов данных в распределенной параллельной моделирующей среды (РПМС) [Электронный ресурс]. — Режим доступа: http://masters.donntu.ru/2011/fknt/navoev/diss/index.htm, свободный. — Загл. с экрана.
  19. Технология CUDA: Что такое CUDA? [Электронный ресурс]. — Режим доступа: http://www.nvidia.ru/object/what_is_cuda_new_ru.html, свободный. — Загл. с экрана.