Practical UNIX and Internet Security, 3rd Edition

Understanding Filesystems

As the name implies, filesystems store information in files. A file is a block of information that is given a single name and can be acted upon with a single operation. For example, on a Unix system this block of data can be copied with the cp command and erased with the rm command.^[62] Contiguous portions of the data can be read or written under program control.

In addition to the data that is stored in files, filesystems store a second kind of data called metadata, which is information about files. The metadata in a typical filesystem includes the names of the files, the date that the files were created, and information that is used to group the files into manageable categories.

Other Filesystems

Research into the design of filesystems continues to progress. Much work has gone into the development of so-called “journaling” filesystems that provide faster and more reliable recovery when a system is improperly shut down, replicated (RAID) filesystems that are highly resistant to failures and outages, network filesystems that present the user with a unified view of files actually stored on many different computers, and cryptographic filesystems that encipher all the data stored on them to protect confidentiality.

UFS and the Fast File System

The original Unix File System (UFS) pioneered many of the concepts that are widespread in filesystems today. UFS allowed files to contain any number of bytes, rather than forcing the file to be blocked into ...