8.1 File Systems Overview
8.1.1 Partitions
When the UNIX operating system was first developed, hard disks could store what is now considered a very small amount of data. When disks larger than 300 Mbytes arrived on the market, the operating system could not address such a large amount of space. This necessitated the development of logical partitions of the hard drive to allow the kernel to access smaller, addressable parts of the drive. Partitions can be either physical or logical. If a partition takes up the entire physical hard disk then it is a physical partition. If the hard disk is divided into smaller sections, these are considered logical partitions.

Today, with hard drives being sold with up to a Terabyte (Tbyte or a million million bytes) or more of space on them, the kernel can access all available space. Despite this, the convention of partitioning a hard drive into logical partitions has persisted for various reasons. Partitioning allows an administrator to functionally organize data so that user files are on a different logical partition than operating system, executables or applications.

One of the main advantages to this is that daily backups can be done only to those partitions on which the data changes frequently, without having to back up all information on the disk. Logical partitioning also cuts down on seek time. If the disk is partitioned, when a user specifies a file to search for or a directory to change to, the name of the directory gives the kernel information about where the information is stored. This prevents the system from having to seek the information over the entire disk. Other benefits of partitioning include security and dynamic size allocation. For increased security, partitions on servers can be shared to the network or not which means that a system administrator can make them unavailable for workstations to access when desired. Partitions or file systems can increase or decrease in size dynamically.

Logical partitions are referred to as file systems. They are transparent to a user and therefore appear to be just part of the directory hierarchy. The most common file systems on a UNIX system are /usr, where binary and executables are stored, /opt, where third-party applications are usually located, and root (/), where the files that pertain to the operation of the system are kept. A UNIX file system is similar to a drive letter in the PC world or a Volume with Novell NetWare. Most hard disks in today's PCs are one physical partition which is referred to as drive C:. A PC hard disk can also be divided into smaller logical partitions, which are referred to as drive C:, drive D: and drive E:, etc.