Introduction
Topicality
Goals and tasks
Scientific novelty
Expected practical results
The degree of elaboration of the scientific and unsolved problems
Description of studies
Conclusion
Introduction
Topicality
Formation of data base management systems (DBMS) 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 have become the dominant tool for creating data-intensive applications. Through the integration of workstations in a distributed environment it is possible to more efficient allocation of functions in it, when applications are executed on the workstations, called application servers and databases maintained a dedicated computer called a server database. This is the source of such distributed architectures, where the role of nodes serve not only general-purpose computers and specialized servers. In accordance with the above is an actual task of developing a new architecture of the database system that is more responsive to the requirements of performance, security and availability of data for comparison with existing architectures. The practical goal of this work was to develop and study new methods of communication, buffering, and the management of requests in relation to the new architecture.
Goals and tasks
To achieve these goals it is necessary to solve the following tasks:
1) Develop an optimal architecture of DBMS;
2) Implement a backup and unzipping the data;
3) Perform client identification and verification of compliance with the rights claimed request;
4) Provide other subsystems high-level langrage API for working with databases;
5) Ensure that store information about the request before and after.
Scientific novelty
Scientific novelty of this work is to create the optimal structure of the database system uses RPMS. Given the architectural features of distributed computing and parallel processing is required to propose a new structure that not only met all the requirements specified in the "Goals and tasks", but also to optimize the technical costs.
Expected practical results
Is assumed to choose the best, according to a match, the DBMS. Then create an API for handling it the other subsystems that provide a comfortable fit, protection against unauthorized access and detailed adequately logged data for the early detection and error recovery.
Search on this issue
According to the results search for approaches to the task was found only work of Merenkov AV, which had a similar problem but the solution took place in the context of a technical architecture.
Description of studies
Figure 1 shows the architecture of database management system that was designed specifically for operation in a parallelized system. The need to work with several distinct sub-systems (sending and receiving data) has demanded the introduction of power "identification" and "determination of rights." So in the block "identification" is the definition of the client (the subsystem) and assigning it a unique number that includes a code number and a unique subsystem id treatment. Further there is a determination of rights and defamation comparison with the incoming request, which excludes the erroneous entry or data retrieval. Block "registration" allows you to store information about incoming requests after authentication is complete the procedure and immediately after the completion of the work with the database. This allows you to monitor the work of the DBMS, as well as to timely detect and correct errors in the performance of queries. Subblock "backup" (performed only when data is received on account) allows you to save space and speed up reading data record in the database. To provide an opportunity for other subsystems high-level API for working with database unit was introduced "standard translation", which frees developers from writing other subsystems sql queries. In the block "request queue" built-in ability to assign priority, which will turn out to fulfill requests that require immediate treatment. Block " CMBD " works directly with the database. Its main task is to send the request, as well as receive data from the database. Once the database returns the data on demand, "CMBD" sends a parallel record in the block "registration" and the sub-block "decompressing." After which the final information on the selected id previously received by the user
Conclusion
In this work described the architecture of the database subsystem is designed to work in a parallelized system. The architecture implements all the basic features common database management system (input, output, registration and data management), as well as take into account some features unique to this model (identification and definition of client's rights, a high-level API for working with databases, the priority queue). The presented structure allows you to perform all the tasks, as well as meets the essential requirements for the implementation of the databases in threaded systems.
-
Development and research of topological analysis subsystem of network dynamic systems as objects of simulation
Автор: Мирошниченко Кирилл Владимирович
Научный руководитель: Святный Владимир Андреевич
-
Анализ
существующих решений и задачи разработки топологического анализатора
сетевых динамических объектов в составе распределенной параллельной
моделирующей среды
Автор: Мирошниченко Кирилл Владимирович
Научный руководитель: Святный Владимир Андреевич
-
Подсистема топологического анализа в распределенной параллельной моделирующей среде
Автор: Мирошниченко Кирилл Владимирович
Научный руководитель: Святный Владимир Андреевич
-
Development and research of the client interface of the subsystem of the visualization of distributed parallel simulation environment
Автор: Терентьев Вячеслав Юрьевич
Научный руководитель: Святный Владимир Андреевич
-
Функции и системная организация клиентского интерфейса подсистемы визуализации распределенной параллельной моделирующей среды
Автор: Терентьев Вячеслав Юрьевич
Научный руководитель: Святный Владимир Андреевич
-
Organisation of dialog subsystem of distributed parallel simulation environment
Автор: Макогон Владимир Николаевич
Научный руководитель: Святный Владимир Андреевич