Naming and Directories

1
Naming and Directories
2
Recall from the last time

• File system components
• Disk management organizes disk blocks into files.
• Many disk blocks management schemes
• A file header associates the file with its data
  blocks
• Naming provides file names and directories to
  users.
• Protection
• Reliability

3
File Header Storage

• Under UNIX, a file header is stored in a data
  structure called i-node
• For early UNIX systems
• I-nodes are stored in a special array
• Fixed number of array entries
• Maximum number of files fixed
• Not stored near data blocks on disk
• Reading a small file involves
• One disk seek to get the i-node
• Other disk seek(s) to get file blocks

4
Reasons for Separate Allocations

• Reliability
• Data corruptions are unlikely to affect i-nodes
• Reduced fragmentation
• File headers are smaller than a whole block
• By packing them in an array, multiple headers can
  be fetched from disk
• File headers are accessed more often
• e.g., ls
• Grouping file headers improves disk efficiency

5
For BSD 4.2

• Portions of file header array stored on each
  cylinder
• For small directories
• All file headers and data stored in the same
  cylinder
• Reduce seek time

6
Naming

• Naming allows users to issue file names instead
  of i-node numbers
• A mapping from names (paths) to I-nodes
• Similar to the DNS in the Internet.

7
Directories

• A table of file names and their i-node numbers
• Under many file systems
• Directories are implemented as normal files
• Containing file names and i_node numbers
• Only the OS is permitted to modify directories
• Is this right?

8
Name Space

• Flat name space
• Hierarchical naming
• Relational name space
• Contextual naming
• Content-based naming

9
Flat Name Space

• All files are stored in a single directory
• Easy to implement
• - Not scalable for large directories
• Name collisions multiple files with the same
  names

10
Hierarchical Naming

• Uses multiple levels of directories
• Most popular name space organization
• Conceptual model maps well into the human model
  of organizing things
• A file cabinet contains many files
• Scalable
• The probability of name collisions decreases
• Spatial locality
• Store all files under a directory within a
  cylinder to avoid disk seeks

11
More on Hierarchical Naming

• Absolute path name consisting the path from the
  root directory / to the file
• e.g., /pets/cat.jpg

root directory
sub directory
file name
12
Drawbacks of Hierarchical Naming

• - Not all files can fit into the hierarchical
  model
• - Accessing a file may involve many levels of
  directory lookups, or a path resolution before
  getting to the file content

13
An Example of Path Resolution

• To access the data content of /pets/cat.jpg
• The system needs to perform the following disk
  I/Os

14
An Example of Path Resolution

• To access the data content of /pets/cat.jpg
• The system needs to perform the following disk
  I/Os
• 1. Read in the file header for the root directory
  /
• Stored at a fixed location on disk

15
An Example of Path Resolution

• To access the data content of /pets/cat.jpg
• The system needs to perform the following disk
  I/Os
• 2. Read the first data block for the root
  directory
• Lookup the directory entry for pets

pets
16
An Example of Path Resolution

• To access the data content of /pets/cat.jpg
• The system needs to perform the following disk
  I/Os
• 3. Read the file header for pets

pets
pets
17
An Example of Path Resolution

• To access the data content of /pets/cat.jpg
• The system needs to perform the following disk
  I/Os
• 4. Read the first data block for the pet
  directory
• Lookup the directory entry for cat.jpg

pets
pets
cat
18
An Example of Path Resolution

• To access the data content of /pets/cat.jpg
• The system needs to perform the following disk
  I/Os
• 5. Read the file header for cat.jpg

pets
cat
pets
cat
19
An Example of Path Resolution

• To access the data content of /pets/cat.jpg
• The system needs to perform the following disk
  I/Os
• 6. Read the data block for cat.jpg

pets
cat
pets
cat
20
Some Performance Optimizations

• Top-level directories are usually cached
• A user inside a directory (e.g., /pets)
• Can issue relative path names (e.g., cat.jpg) to
  refer files within the current directory

21
Relational Name Space

• Hierarchical naming model is largely a tree
• One step beyond is the relational naming model,
  which allows the construction of general graphs
• A file can belong to multiple folders
• According to its attributes
• Files can be accessed in a manner similar to
  relational databases

22
Pros and Cons of Relational Name Space

• More flexible than hierarchical naming
• - May require a long list of attributes to name a
  single piece of data
• e.g., this lecture
• Keywords operating systems, file systems,
  naming, PowerPoint XP
• - Who will create those attributes?

23
Contextual Naming

• Takes advantage of the observation that certain
  attributes can be added automatically
• e.g., when you try to open a file by Word, a
  system will search only the file types supported
  by Word (.doc, .txt, .html)
• Avoids a long list of attributes

24
Content-Based Naming

• Searches a file by its content instead of names
• File contents are extracted automatically
• e.g., I want a photo of a cat taken five years
  ago
• The system returns all files satisfying the
  criteria

25
Content-Based Naming

• - Requires advanced information processing
  techniques
• e.g., image recognition
• Many existing systems use manual indexing
• Automated content-based naming is still an active
  area of research

26
Example The Internet File System

• Can be viewed as a worldwide file system
• What is the naming scheme for the Internet file
  system?

27
The Internet File System

• Contains shades of various naming schemes
• Flat name space
• Each website provides a unique name
• Hierarchical name space
• Within individual websites
• Relational name space
• Can search the Internet via search engines
• Contextual name space
• Page ranked according to relevance
• Content-based name space
• You can find your information without knowing the
  exact file names

28
Example Plan 9

• Modern UNIX has a deep-rooted influence from the
  Plan 9 OS
• Developed by Bell lab
• Major design philosophy everything is a file
• A single hierarchical name space for
• Processes (e.g., /proc)
• Files
• IPC (e.g., pipe)
• Devices (e.g., /dev/fd0)
• Use open/close/read/write for everything
• e.g., /dev/mem