1. File Management
Chase Goehring

2. What is a File?
“From a user’s perspective, a file is the smallest allotment of logical secondary storage; that is, data cannot be written to secondary storage unless they are within a file.”
- Silberschatz et al.
In other words…
It’s what everything on your hard drive is stored in.

3. Properties of Files
Name - what the human user can identify the file by
Identifier - what the computer can identify the file by
Type - what kind of file it is
Location - where it is on the storage device
Size - number of bytes
Protection - read/write permissions, execution, etc.
Time, Date - shows when a file was created, last modified, or used

4. File Operations
Create a file - Space allocated on disk, then file entry is placed on directory
Write to a file - information is added to a file in the location of the write pointer
Read a file - get information from a file from the read pointer’s location
Reposition - change the location of the read/write pointer
Delete a file - release all file space and erase directory entry
Truncate a file - keep properties but erase contents

5. File Organization
Extremely wide variety of file organization implementations
Stallings’ top 5:
Pile
Sequential File
Indexed Sequential File
Indexed File
B-Tree

6. The Pile Least complicated form of file organization
Data collected in the order it arrives
Each record is a single “burst” of data
Each record must be self-describing (include field name and value, length of field must be delimited or included as a subfield)
Records can only be accessed by exhaustive search
Works well when stored data varies in size and structure
Typically used when data is not easy to organize
Unsuitable for most applications

7. The Sequential File Most common form of file structure
Fixed length records w/ fixed length fields
Key field stores record identifier
Records ordered by key sequence
Typically used in batch applications
Poor performance for applications that involve queries/ updates of individual records
Access requires key match or sequential search

8. Indexed Sequential File
Index added to sequential file structure to support random access
Index provides for quick searches
Includes Overflow file
Multiple levels of indexing possible

9. Indexed File Multiple indexes, one for each type of searchable field
No sequentiality, no single key
Record access only through indexes
Two indexes: Main, Partial
Main index - sequential exhaustive index
Partial index - entries to records with field of interest
Useful for situations where information is needed quickly
Not very good for exhaustive searching

10. B-Trees B-Tree: Balanced-Tree structure of an index file
Sequential file of indexes is not fast enough for large databases
All branches are of equal length
Structure found in Mac OS X, Windows, and many Linux file systems

11. Properties of B-Trees B-Trees have a minimum degree d
Every node has at most 2d-1 keys and 2d children
Every node has at least d-1 keys and d pointers (is at least half full)
Root has at least one key and 2 children
All leaves are on the same level
Nonleaf nodes with k pointers contain k-1 keys

12. B-Tree insertions

13. Directories
A directory is a file accessible by file management routines
Directories store information on other files
Basic information - File name, type, organization
Address information - Volume, Starting address, size used/allocated
Access Control information - permissions
Usage information - date created, last modified, etc.

14. Directory Structures
List of entries Entire structure is a sequential file with the filename being the key Bad for systems with many files or multiple users
Hierarchical Tree Allows for directories and subdirectories

15. Structure of Tree Directory

16. File System Properties
Users can create, read, write, delete, modify files
Users should have controlled access to other user’s files
Each user should be able to move data between files
Each user should be able to back up and recover user’s files in case of damage
Users should see files by filename instead of identifier

17. Basic File System Structure
Device drivers - communicate with devices, start I/O operations, process I/O completion (considered part of OS)
Basic file system - physical i/o, deals with blocks of data exchanged with disk/tape (considered part of OS)
Basic I/O supervisor - responsible for all I/O initiation and termination, is part of OS
Logical I/O - enables users and apps to access records
Access method - provides interface between applications and file systems and devices

19. FAT 32 File System Introduced in 1996 to Windows 95
FAT = File allocation table: holds start of file cluster and file size
Has a backup file allocation table stored after the first one in the volume
Max filesize = 4GB
Max volume size = 2TB
Does not support file permissions
Still used today in portable drives, like USB or SDHC cards

20. NTFS file system NTFS = New technology file system
Introduced in July 1993 with Windows NT 3.1
Uses master file table, which is duplicated much like FAT
MFT allows for faster access to files than FAT
Uses B+ Tree instead of table for directory contents
Maximum file size = 16 exabytes (exabyte = 1,000 GB)
JOURNALING - changes to volume are saved in a log before they are committed = less corrupted data

21. HSF+ Introduced January 1998 with Mac OS 8.1
HSF+ = The Mac OS Extended Volume Hard Drive Format
Volume header keeps information about volume, has a backup at end of volume
Catalog file holds all folder and file hierarchy information, organized as B-Tree
Max volume size = 8 exabytes
Journaling as of 2002

22. References
“A Short History of the BTree.” A Short History of the BTree | Perforce, Perforce, 28 Sept. 2011,
Hoffman, Chris. “Why Do Removable Drives Still Use FAT32 Instead of NTFS?” How-To Geek,
How-To Geek, 5 July 2017,
“NTFS: FAT File System.” NTFS.com. The FAT File System,
Silberschatz, Abraham, et al. Operating System Concepts, 8th Edition. 8th ed., John Wiley &
Sons, 2008.
Stallings, William. Operating systems: Internals and Design Principles. 8th ed., Pearson, 2015.
Tanenbaum, Andrew S. Modern Operating Systems. 3rd ed., Pearson Education, 2009.
“Technical Note TN1150 HFS Plus Volume Format.” Technical Note TN1150: HFS Plus Volume
Format, dubeiko.com/development/FileSystems/HFSPLUS/tn1150.html.