Directory Structure

Introduction

Imagine having a thousand files with no folders — just a flat list of names. Finding anything would be a nightmare, and name conflicts would be constant (you cannot have two files called "report.txt"). Directory structures solve this by organizing files into a hierarchy, much like a physical filing cabinet has drawers, folders, and sub-folders.

A directory (or folder) is a special file that contains entries mapping filenames to their file information (inode numbers in Unix, MFT entries in NTFS). The structure of how directories relate to each other determines how easily users can organize, find, and share files.

Single-Level Directory

The simplest structure: all files in one directory. Every file must have a unique name.

Advantages: Simple to implement, fast file lookup (just scan one list) Disadvantages: Name conflicts when multiple users share the system, no organization, does not scale beyond a few dozen files.

Two-Level Directory

Each user gets their own directory. Solves the name conflict problem but provides no further organization within a user's space.

Advantages: Isolates users, allows same filenames for different users Disadvantages: No sub-directory organization within user space, sharing files between users is difficult.

Tree-Structured Directory (Hierarchical)

The standard approach used by all modern OS. Directories can contain both files AND other directories, creating an arbitrary-depth hierarchy.

Advantages: Natural organization, unlimited depth, easy to manage permissions at directory level, efficient searching (navigate to relevant subtree). Disadvantages: A file can only exist in one directory (no sharing without links), deeper hierarchies mean longer paths.

Path Navigation

Acyclic Graph Directory

Extends tree structure by allowing shared files/directories through links. A file can appear in multiple directories without being duplicated.

Types of Links

Hard link: Multiple directory entries point to the same inode. The file is deleted only when ALL links are removed.

Symbolic link (soft link): A special file containing the path to another file. Like a shortcut. If the target is deleted, the link becomes "dangling."

Challenge: Deletion is complex. If you delete a shared file, do other links become invalid? Hard links solve this with reference counting. Symbolic links can dangle.

General Graph Directory

Allows cycles (directory A contains B which contains A). This is problematic because traversal algorithms can loop infinitely and garbage collection becomes necessary to reclaim space.

Most file systems PREVENT cycles through restrictions:

• Only allow symbolic links (not hard links) to directories
• Check for cycles when creating links
• Garbage collection to find unreachable files (if cycles exist)

Directory Operations

Common operations the OS provides:

• Create directory: mkdir /home/alice/new_folder
• Delete directory: rmdir /home/alice/empty_folder
• List contents: ls /home/alice/ or dir
• Rename: mv old_name new_name
• Search: find / -name "*.pdf"
• Change current directory: cd /home/alice

Real-World Analogy

Directory structures are like address systems. A single-level directory is a small village where everyone has a unique first name. A two-level directory is apartments with unit numbers (Building A, Unit 3). A tree-structured directory is a full mailing address (Country/State/City/Street/Building/Apartment). An acyclic graph is when the same person receives mail at two addresses (home and office). The hierarchy helps postal workers (the OS) efficiently route mail (find files) without checking every address in the country.

Key Takeaways

• Directory structures organize files hierarchically for easy management and retrieval
• Tree-structured (hierarchical) directories are the universal standard in modern OS
• Acyclic graphs extend trees with file sharing through hard and symbolic links
• Cycles are generally prevented because they complicate traversal and deletion
• Paths can be absolute (from root) or relative (from current directory)
• Hard links share the same inode; symbolic links store a path string
• Good directory organization improves both human usability and system performance