(2021 Revision) Chapter 3: Essential Aspects of Physical Design and Implementation of Relational Databases
Tatiana Malyuta, Ashwin Satyanarayana · CUNY Academic Works (City University of New York) · 2021
The distributed solution, which is more expensive and more complex than the centralized one, is justified by the following expectations: Improving performance of local applications in each city and the global New York application Achieving better reliability and availability of the data Ensuring that the database is scalable and will be able to function with an increased number of users and larger amounts of data. Prerequisites of the Distributed DesignDistribution of data is one of the considerations in performing the physical design of the database.Decisions related to physical design are based on the logical data model and user requirements that specify: The structure of the company and the architecture of the company's network How data will be used by database applications, including what data is needed by each application, the amounts of processed data, and the frequencies of applications' execution.This chapter discusses how to use these requirements for designing appropriate distributions of data.When performing the design and implementation of the distributed database, it is important to understand that the physical data model has to correspond to the logical model and preserve all semantics of the latter. Review of Basic Features of the Distributed DatabaseThe basic features of distributed databases which were discussed in Chapter 2 on the physical model of data are as follows: The distributed database is a collection of logically interrelated shared data stored across several physically independent databases. Participating databases are linked by a network. Each participating database is controlled by a DBMS autonomously.