Table and Index Organization - Relational Database Index Design and the Optimizers - page 11

Chapter 2

Table and Index Organization

ž The physical organization of indexes and tables
ž The structure and use of the index and table pages, index and table rows,
buffer pools, and disk cache
ž The characteristics of disk I/Os, random and sequential
ž Assisted random and sequential reads: skip-sequential, list prefetch, and
data block prefetching
ž The significance of synchronous and asynchronous I/Os
ž The similarities and differences between database management systems
ž Pages and table clustering, index rows, index-only tables, and page
adjacency
ž The very confusing but important issue of the term cluster
ž Alternatives to B-tree indexes
ž Bitmap indexes and hashing

INTRODUCTION

Before we are in a position to discuss the index design process, we need to
understand how indexes and tables are organized and used. Much of this, of
course, will depend on the individual relational DBMS; however, these all rely
on broadly similar general structures and principles, albeit using very different
terminology in the process.
In this chapter we will consider the fundamental structures of the relational
objects in use; we will then discuss the performance-related issues of their use,
such as the role of buffer pools, disks and disk servers, and how they are used
to make the data available to the SQL process.
Once we are familiar with these fundamental ideas, we will be in a position,
in the next chapter, to consider the way these relational objects are processed to
satisfy SQL calls.
This chapter is merely an introduction . Considerably more detail will be
provided throughout the topic at a time when it is more appropriate. At the end