File Systems and Disk Management

1
File Systems and Disk Management

• Andy Wang
• Operating Systems
• COP 4610 / CGS 5765

2
Design Goals of File Systems
Physical reality File system abstraction
Block-oriented Byte-oriented
Physical sectors Named files
No protection Users protected from one another
Data might be corrupted if machine crashes Robust to machine failures
3
File System Components

• Disk management organizes disk blocks into files
• Naming provides file names and directories to
  users, instead of tracks and sector numbers (e.g.
  Takashita)
• Protection keeps information secure from other
  users
• Reliability protects information loss due to
  system crashes

4
User vs. System View of a File

• User level individual files
• System call level collection of bytes
• Operating system level
• A block is a logical transfer unit
• Even for getc() and putc()
• 4 Kbytes under UNIX
• A sector is a physical transfer unit
• 512-byte sectors on disks
• File a named collection of blocks

5
User vs. System View of a File

• A process
• Read bytes 2 to 12
• OS
• Fetch the block containing those bytes
• Return those bytes to the process

6
User vs. System View of a File

• A process
• Write bytes 2 to 12
• OS
• Fetch the block containing those bytes
• Modify those bytes
• Write out the block

7
Ways to Access a File

• People use file systems
• Design of file systems involves understanding how
  people use file systems
• Sequential accessbytes are accessed in order
• Random access (direct access)bytes are accessed
  in any order
• Content-based accessbytes are accessed according
  to constraints on bye contents
• e.g., return 100 bytes starting with aye carumba

8
File Usage Patterns

• Most files are small, and most references are to
  small files
• e.g., .login and .c files
• Large files use up most of the disk space
• e.g., mp3 files
• Large files account for most of the bytes
  transferred between memory and disk
• Bad news for file system designers

9
File System Design Constraints

• High performance
• Efficient access of small files
• Many small files
• Used frequently
• Efficient access of large files
• Consume most disk space
• Account for most of the data movement

10
Some Definitions

• A file contains a file header, which associates
  the file with its disk sectors

File header
11
Some Definitions

• A file system needs a disk allocation bitmap to
  represent free space on the disk, one bit per
  block

12
Disk Allocation Policies

• Contiguous allocation
• Link-list allocation
• Segment-based allocation
• Indexed allocation
• Multi-level indexed allocation
• Hashed allocation

13
Contiguous Allocation

• File blocks are stored contiguously on disk
• To allocate a file,
• Specify the file size
• Search the disk allocation bitmap for consecutive
  free blocks

File header
14
Pros and Cons of Contiguous Allocation

• Fast sequential access
• Ease of computing random file locations
• Adding an offset to the first disk block location
• - External fragmentation
• - Difficulty in growing files

15
Linked-List Allocation

• Each file block on a disk is associated with a
  pointer to the next block
• A special marker to indicate the end of the file
• e.g., MS-DOS file system
• File attribute table (FAT)

File header
16
Pros and Cons of Linked-List Allocation

• Files can grow dynamically with incremental
  allocation of blocks
• - Sequential access may suffer
• Blocks may not be contiguous
• - Horrible random accesses
• May involve multiple sequential searches
• - Unreliable
• A corrupted pointer can lead to loss of the
  remaining file

17
Indexed Allocation

• Uses a preallocated index to directly track the
  file block locations

File header
18
Pros and Cons of Indexed Allocation

• Fast lookups and random accesses
• - File blocks may be scattered all over the disk
• Poor sequential access
• Needs defragmenter
• - Needs to reallocate index as the file size
  increases

19
Segment-Based Allocation

• Needs a segment table to allocate multiple,
  contiguous regions of blocks

File header
20
Pros and Cons of Segment-Based Allocation

• Relax the requirements for large contiguous
  disk regions
• - Fragmentation ? 100
• Segment-based allocation ? Indexed allocation
• - Random accesses not as fast as pure contiguous
  allocation

21
Multilevel Indexed Allocation

• Certain index entries point to index blocks, as
  opposed to data blocks (e.g., Linux ext2)

File header
22
Multilevel Indexed Allocation

• A single indirect block contains pointers to data
  blocks
• A double indirect block contains pointers to
  single indirect blocks
• A triple indirect block contains pointers to
  double indirect blocks

23
Pros and Cons of Multilevel Indexed Allocation

• Optimized for small and large files
• Small files accessed through the first 12
  pointers
• Large files can grow incrementally
• - Multiple disk accesses to fetch a data block
  under triple indirect block
• - Largest file size capped by the number of
  pointers
• - Arbitrary file size boundaries among levels

24
Hashed Allocation

• Allocates a disk block by hashing the block
  content to a disk location

Old file header
New file header
25
Pros and Cons of Hashed Allocation

• File blocks of the same content can share the
  same disk block to save storage
• e.g., empty blocks
• Good for backups and archival
• Small modifications to a large file result in
  only additional storage of the changes
• - Poor disk performance