File Systems and Storage

• Files and file systems are examined
• The chapter covers file and file system implementation, including file supports, file system implementation on disks, directories and links, and examples of file systems like Ext2 and NTFS.

Memory

• Electrotechnical means of storing information.

Storage Support Types

• Semi-conductor: Includes RAM, flash memory (USB sticks), and SSD (Solid-State Drives).
• Magnetic: Includes hard disks, magnetic tapes, and floppy disks.
• Optical: Includes CD-ROM, DVD-ROM, and Blu-ray discs.
• Paper: Includes punched cards.
• Hybrid: Includes SSHD hard drives (SSD + HDD).

Storage Levels

• Primary Storage: Directly accessible by the CPU (RAM).
• Secondary Storage: Requires specific channels to transfer information to primary storage.
• Tertiary Storage: Involves robotic mechanisms to mount and dismount storage, with data copied from secondary storage.
• Offline Storage (Removable): Requires manual insertion of the storage medium by the user.

Memory Characteristics

• Volatility:

  • Volatile memory loses data when power is off
  • Non-volatile memory retains data even when power is off

• DRAM vs SRAM:

  • DRAM (Dynamic Random Access Memory) information needs periodic refreshing.
  • SRAM (Static Random Access Memory) does not require refreshing.

• Access:

  • Read and write access
  • Read-only access
  • Fast read, slow write access (CD-ROM example)

• Type of Access:

  • Random access (RAM)
  • Sequential access(CD-ROM, DVD)

Memory Types

• RAM (Random Access Memory): Modifiable and accessible at any time.
• ROM (Read Only Memory): Information cannot be modified.
• Mass Storage: Slower access speed but larger storage capacity.
• Cache Memory: Short-term memory designed to accelerate hardware performance.
• Flash Memory: Can be entirely erased in a single operation.

Logical File

• Is the view of data perceived by users and programs.
• Each file has a determined name.
• Possible operations:
  • Creation/Destruction: Linking of logical to physical file
  • Open/Close: Rupture of the link
• Involves logical unit of processing
• Involves data structure understandable by the program
• Is accessible via access functions (read and write).

File Types

• Ordinary Files: Contains information, programs, or libraries.
• Directories: System files maintaining the file system's structure.
• Character Special Files:
  • Linked to inputs/outputs
  • Transfer data via peripherals like terminals, printers, and networks
  • No timing consideration (character by character send).
• Block Special Files:
  • Linked to inputs/outputs
  • Transfer data via peripherals like hard drives, USB drives, and CD-ROMs
  • Buffering is employed to accelerate transfers.

Ordinary Files

• Text/ASCII Files: Contains lines of text that can be directly read and printed.
• Binary Files: Contains incomprehensible byte sequences, except for applications.
• File type is determined by:
  • Extension (e.g., .exe in MSDOS)
  • Attribute (in MAC-OS)
  • Defined only by the application recognizing the extension.

File Attributes

• Contains complementary information about a file.
• The number of attributes varies depending on the system.
• Examples include:
  • Read, Write, and Execute permissions (r, w, x in Unix)
  • File creator/owner
  • File type (ASCII or binary)
  • Creation date and time.

Logical File Access Modes

• Sequential Access:

  • Records are processed one after another
  • A "pointer" indicates the current record
  • Reading returns the current record and moves to the next one
  • Writing occurs at the end of the file
  • Access can be read-only or write-only

• Indexed/Random Access:

  • Direct access to a record
  • A key is used (common field among records)
  • Maintained index or access structure
  • Access can be read-only, write-only, or read/write

• Direct/Relative Access:

  • Specifies the relative position of the record
  • Relative to the beginning, end, or current position

File Implantation Challenges

• Allocating entity on the support.
• Contains the physical data.
• Address questions on:
  • How to place a file on the storage support, as blocks or not.
  • Where to place the file.
  • How to find the file later.

Hard Disk Structure

• May have up to 20 platters plates inside an average hard disk
• has 2 faces per platter and 10 to 1000 tracks per face
• has one cylinder per track and four to 32 sectors per track

Example Hard Disk Structure

• Three platters make up six faces with six read heads. Sectors typically hold 512 bytes (ranging from 32 to 4096 bytes). The first addressable sector is at cylinder 0, head 0, and sector 0.

Blocks

• Optimization of read/write operations.
• Groups several consecutive sectors
• Exchange unit between the disk and the memory.

Partitions

• Sector 0 of the disk contains the MBR (Master Boot Record)
  • Used to boot the machine.
• MBR followed by the partition table.
  • Includes start and end addresses of each partition.
• Each partition Contains:
  • Boot Block
  • Superblock: contains info on the file system

Machine Startup

• Query the MBR to get the active partition.
• Read the first block of that partition.
• The program in the boot block loads the system.

File Implantation: Contiguous

• Files are stored in consecutive blocks.

Contiguous File Storage Advantages:

• Simple to implement. Offers excellent performance due to only one initial movement to the file's start.

Contiguous File Storage Disadvantage

• Suffers from space fragmentation.

File Allocation Strategies

• first-fit
• next-fit
• best-fit
• worst-fit

First-Fit Algorithm

• Begins searching from the start for file placement. It advances to the next free segment until one is found. If the segment size is insufficient, advances to the next free segment until the current segment is adequate and selected. The algorithm is fast due to its limited search range.

Next-Fit Algorithm

• Similar to First-Fit but starts searching from the position of the previously found free segment. Simulation indicates it's faster than First-Fit.

Best-Fit Algorithm

• Chooses the most optimal space in the available disk space. Space can be lost because available disk space is typically very small

Worse-Fit Algorithm

• Space is allocated in the largest block on the disk. Performance is not efficient based on simulations

File Implantation: Chained List

• Each file is a linked list of blocks that may be distributed across the disk.

Linked List File Storage Advantages:

• No space fragmentation.

Linked List File Storage Disadvantages

• Slower and more complex file reading. Requires using octets within each block to indicate the next block.

File Implantation: Chained List and FAT

• Pointers to the block are in stored on a table (file allocation table)
• advantage: a physical block fully available for data
• disadvantage: the table can often require large spaces.

File Implantation - Indexed

• Each file has a structure called an i-node structure with attributes and addresses for the disk blocks that make up the file
• benefit, only the i-nodes that are open get loaded

Free Space Management

• For each block a bit is associated with a table or a chain/
  • 0=free block
  • 1=not free
  • there could be problems with size of table of chain though

Free Space Management: Chain List

• Each entry indicates:
  • Status of a segment
  • The address where the segment starts
  • The size of the segment

Directories

• Correspondence between logical and physical objects is done using a directory
• a directory is an entry that represents a file and contains
  • name of the file
  • the type of the file
  • the location of the name
  • and the protections

Directory Level Types

• Single-Level Directory Systems: All files reside in one directory.
• Two-Level Directory Systems: A root directory includes a directory for each user and a subdirectory for each file.
• Hierarchical Directory Systems: Organized like a tree structure with possible child subdirectories.

Important note

• Most Modern OS are built of hierarchical directory systems
• Make sure the maximum size isn't too big, for system limits

Links

• It's use to share files between several users.
• Physical and material links
  • Additional file in a directory, pointing to the same inode
  • Allows multiple names on a single file
  • The inode has a counter for listing thanks to the is -I command Symbolic Links
  • It's a text file that contains the path and the name of the symbolic link
  • It is identified with a specified type (in linux it appears with '@s' on the end)
  • Associated to a short cut

File System Examples

• Ext2 (Extended File System version 2)
• NTFS (NT File System)

EXT2 vs EXT3 vs EXT4

• Ext2: The initial block is the Boot Block. Subsequent block groups each contain a super block, group descriptor, inode table, and data blocks. Block sizes range from 512 to 4096 bytes.
• Ext3: Aimed to incorporate journaling into Ext2, maintains compatibility and is good for personal use but not suitable for servers.
• Ext4: Handles disks up to 1024 x 250 bytes, contiguously allocates files to reduce fragmentation, and has limited compatibility with Ext3.

Ext2 i-nodes

• At 128 bytes, each inode has an associated "name" and "type" for file-rights. In addition inodes has how the last time you accessed, modifed etc, for each file. Also what permissions are required
• "There are many different list controls for files", "and table addresses."
• Table has 15 entries 12, of the first are logical blocks with the data, it's 4 octect in size
• The Last 3 are level of direction

Key Study Statistics:

• A single I-node has a lot of smaller files and can locate/store 12 blocks of data
• Block 13 is related to, 1st indirection
• Block 14 is related to, 2nd indirection
• Block 15 is related to, 3rd indirection
• Maximum sized file is around 12 X Block Table's size
• if the table is greater, then it uses an indirection address with blocks that, in return, link together other Block's addresses

Journaling

• Motivation includes that when the machine's in a unstable state when power is loss, all new tools require the system to boot in emergency tools/mode
  • there are automatically commands running on boot up that verify the files on a specific structure, but this takes time
• A single block containing "journal" can store a ton of new information
  • Allows system to quickly re run that block if the system crashes

NTFS (NT File System)

• Developed for Windows NT, also Known as New Technology File System.

Key features:

• Addresses Disks on 64 bits, which can create massive sized directories
• File names limited to 255, path limited to 32,767 characters
• Possible use with unicode
• Has some limitations on API's
• Blocks Size can range between 512 -64,000 bytes

Explore file systems, storage implementation, and different memory and storage types. Learn about semi-conductor, magnetic, and optical storage. Understand storage levels like primary, secondary, and tertiary storage.