Physical Storage and Distribution

Abstract

In this chapter we consider how we actually store data. Turning information into a series of 1s and 0s is at the heart of every current database system and so an understanding of issues like physical storage and distribution are important concepts to understand. We explore the ways in which database architectures are built up from the building blocks of bytes and data blocks, and the way that physical elements such as RAM and Disk memory are used by the database system. We illustrate the way that these elements interact by walking through a simple read operation. For most databases permanency is of critical importance and we examine how data is stored and how memory structures are used to attempt to minimise the comparatively slow process of reading data from disk. Most users, however, interact with a modern database system by using logical concepts such as tables and rows, and so we also explain how physical storage can map to logical objects. Database professionals deal with data on a daily basis. We look at the ways that this raw material can be moved between disparate systems using tools like SQL Loader and Import/Expotr. In addition we look at how professionals can use the supporting files, such as redo logs and alert logs. Finally we discuss placement of the data, and when it may be appropriate to distribute the data stored amongst a number of servers.