Hard Disk Partitions and File Allocation Methods

Mass Storage

• Disks store data for long periods, even when the computer is off
• Disks are referred to as secondary storage
• The operating system manages storage allocation, free space, disk scheduling, and data I/O
• Optical storage (DVD/CD) and magnetic tape are also used as storage

Mass Storage Management

• Data transfer to disk is slow compared to other computer operations
• Data access speed from fastest to slowest: registers, cache, main memory, solid state disk, magnetic disk

Caching

• Data flows from the disk to the CPU
• Flow: hard disk -> main memory -> cache-> hardware register
• Applications transfer instructions from disk to RAM to cache upon opening
• The cache holds the next instruction for the processor to avoid slower retrieval from main memory

Hard Drive Structure

• Hard disk drives are the most common mass storage media due to high storage capacities and moderate cost
• Hard disk drives come in sizes 1.8, 2.5, and 3.5 inches
• Hard disk platter rotation speeds include 4200 RPM, 5400/7200 RPM, and 10,000/15,000 RPM

Hard Disk Structure - Physical Components

• Hard drives consist of platters coated with magnetic media for data recording
• Platters are mounted on a common spindle for synchronous rotation
• Hard disks are divided into Tracks and sectors
• A track is a circular ring on the platter
• A cylinder is the grouping of the same tracks, on all platters
• The head reads/writes data to a track sector
• A cluster is a grouping of sectors used to store a file
• A sector is the most basic unit of on a track, and divided into sectors
• Sectors can hold 512 bytes of data.
• Early Hard drives used pie shaped sectors
• The standard was set at 17 sectors per track.
• Track 0 is at the outer edge
• Tracks are longer on the outside of the disk
• Tracks are grouped into zones based on their location
• Zones are assigned a number of sectors
• Sector sizes remains constant over the disk surface
• Outer zones have more sectors per track

Hard Disk Structure - Data

• The hard disk surface is a magnetic medium
• The head is a small electromagnet
• Sending an electric current to the head creates a magnetic field that magnetizes a small section of the disk
• Data bits are written as small magnetic areas of North-South and South-North, representing 0 and 1
• Changing the electric current direction in the drive controller changes the magnetic field, flipping the data bit

Physical and Logical Blocks

• Disk drives are addressed as physical blocks created by low-level formatting
• Blocks are numbered sequentially from 0 to N
• A physical block has a disk address such as drive 1, cylinder 73, head 2, sector 10
• Logical blocks are mapped into physical blocks sequentially by track, cylinder, then innermost cylinders

Disk Scheduling

• Time delays occur when transferring data due to moving parts
• Transfer Rate: The rate of transferring data between the drive and the computer
• Positioning Time (Random Access Time):
  • Time required to move the drive to the location to be read or written(Seek Time & Rotational Latency)
  • Seek Time: time it takes to move the head to the correct cylinder)
  • Rotational Latency: Time it takes to rotate the platter so that the correct sector is under the head

Disk Scheduling Algorithms

• Improves efficiency by queuing and scheduling disk access requests
• First Come First Served (FCFS):
  • Handles requests for disk access in the order they are received using a first-in-first-out queue.
  • Inefficient because the disk heads move erratically to service each request
• Shortest Seek Time First (SSTF):
  • Re-orders disk access requests based on the current head position, requires less head movement
  • Faster data read/write
  • Some process requests may be delayed over others
• SCAN (Elevator Algorithm):
  • Disk head starts at theouter edge of the disk and moves inwards, servicing access requests in order
  • Moves from inside to outside servicing disk requests
• C-SCAN:
  • Disk head starts at one edge, moves to the other servicing requests, then returns without servicing
• LOOK and C-LOOK:
  • Head travels only as far as the last pending request before returning, more efficient than SCAN/C-SCAN

File System Types

• Operating systems support multiple file system types
• Each file system determines the number of files, storage bytes, and file metadata
• Windows files systems: FAT, FAT16, FAT32, exFAT, NTFS, RFS
• Linux file systems: ext3, ext4, and xfs
• Specialized file systems are available for specific uses like zfs for SAN (Storage Area Network) devices

Clusters

• A cluster (or File Allocation Unit) is where multiple sectors are grouped together

File Allocation

• Files are written to a track in a sector.
• The size of a cluster depends on the file system and hard disk size (typically 4,096, 8,192, or 16,384 bytes)
• Assigns a file to a complete cluster
• Even if the file has only 99 bytes in the data, the whole cluster is dedicated to it
• Remaining cluster space cannot be allocated to another file
• Deleting a file makes the cluster available for new writes
• Fragmentation occurs when large files spread over many clusters in different disk sections

File Allocation Methods

• Three main ways: contiguous, linked, and indexed
• Contiguous: files are stored fully together in contiguous blocks, minimizes fragmentation
  • All files are in contiguous blocks
  • Minimizes head movement when Reading/Writing
  • Drawback: Difficult finding space when disks are fragmented
• Linked: files are stored as lists of linked blocks that may be scattered, allows efficient disk use
• Files may be located (scattered) in any block
• Each end of an allocation unit points to the start of next unit
• Indexed: each allocation unit is in order in an index file, which is useful to speed up operations
  • Each allocation unit is order in an index file

File Allocation Table

• A File Allocation Table (FAT) contains the first block number of the file and a list of all other sequential blocks

Indexed File Allocation

• Each file owns the array of disk blocks used
• File Directory has the location
• Drawback: requires more space
• Used by NTFS
• A hard drive will still require an entire block to store the index even if it only has one or two disks

File Allocation Table - FAT

• Originally designed in 1977 for floppy disks, currently used on solid state memory devices
• FAT16:16-bit, allows max 65,535 clusters per volume
• FAT32: Default for USB memory drives. Limits File size to 4GB
  • 2^32 is 4GB
  • Uses 28bits for the cluster number
  • 2^28 is 268Mill clusters
  • supports up to 2 terabytes in size

New Technology File System (NTFS)

• A Microsoft development for WinNT 3.1, improving on FAT32
• More support for file metadata
  • Has information about file, owner, date created, permissions
  • Additional security with permissions (DACLS)
  • Journaling
  • NTFS volume up to 16 terabytes using the default Size (4 KB) for large volumes
• Volumes up to 256 terabytes using the maximum cluster of 64 KB

Journaled File Systems

• Ensures integrity if errors occurs
• Replays transaction in order to recover file
• Only protects data structure instead of the file data
• Examples include NTFS, ext3, ext4, and xfs

Disk Management

• A disk must be configured before it can be used by an OS
• The low-level formating is performed by the manufacturer
  • Layout of tracks and sectors
• The operating system then stores its data on the disk
• The disk is then partitioned to be used by the operating system

Disk Partitioning

• Disks can be separated into different partitions, treated as logical disk groups of cylinders called volumes
• System below has 2 disks: Disk 0 (60GB) and Disk 1 (10GB)
• The system or boot partition must be setup to create a bootable disk
• The Master Boot Record (MBR) is located in track (cylinder) 0, side (head) 0, and sector 1
  • It contains the Partition Table and a small amount of executable code
• The BIOS instructions are loaded from ROM into memory when the computer powers on
• POST checks hardware, and loads basic kernel I/O like keyboard and video
• The BIOS checks the MBR after POST looking for instructions to allocate the OS

Master Boot Record

• The MBR contains a small amount of executable code called the master boot code, the disk signature, and the partition table for the disk.

Partition Table

• Master boot code is executed when the MBR program runs and branches to the operating system found
• Holds information about the primary and extended partitions
• Has 64-byte data field located in the same sector as the MBR
• Adheres to an independent standard layout for the operating system
• Each Partition Table entry is 16 bytes long, only 4 parititon names max
• A disk has 3 as primary and 1 as extention partition disk
• The operating system resides in the active operating system
• Pre win7 was always C, it is a seperate partition that created unencrpyted partition of the disk

Windows Disk Management

• Multiple partitions may be created for organizational purposes or for booting into different operating systems.
• Primary partition is a reserved part of Windows disk, can be identified by a drive letter such as the entire "C" drive

Disk Partitioning - Examples

• Disk 0 (60GB) has 1 partition set as the primary or boot
• Disk 0 (60GB) has 1 partition formatted for the NTFS file system
• Multiple partitions can be formatted to have for separate file types/OSs.

Globally Unique Identifier Partition Table (GPT)

• A new partition table layout that is part of the Unified Extensible Firmware Interface (UEFI)
• Replaces BIOS (Basic Input Output System) and the Master Boot Record
• Expands on older standards: MBR limited by 32 bit fields and 512 byte sector = 2TiB versus GPT which uses 64bits and is 232 = 8 ZiB
• Allows for unlimited partitions (windows only allows 128 partitions) @ 18 exabytes each, uses a copy in case the original is corrupted.
• Windows defaults to MBR for any hard drives smaller than 2TB.

Dual Boot

• Set up a partition for one OS and a partition for a diff OS by splitting the drive ex. Windows and Linux
• The BIOS loads the boot loader when the computer powers on. Bootloader will have a menu
• Linux installs the Grub
• If windows is already installed, Grub takes its place and allows the OS to choose at startup

Hard Drive Interfaces

• Hard drives are connected using various technologies:
  • IDE, SATA, SCSI, and Fiber Channel
• Integrated Drive Electronics (IDE) is standardized
• Uses IBM 80206 computer
  • I/O controller moved onto on hard disk
  • 40 pin connector with 16 parallel transmission bits data
• Serial ATA-faster serial data transmission
• Small Computer System Interface (SCSI) is a bus system that connects multiple devices to a device controller
• Up to 8 disks attached to a device controller

Data Security

• With typical formatting, the command just clears the file allocation table
• The data files will remain untouched, can be useful if you want to reset an operating system
• The tracks can be overwritten at a high-level format, but the original data is still retrievable
• A low level format is required to lay down this tracks and is therefore usually done by the manufacturer
• Erase any data before selling
• Use a wipe command: the free Windows Washer program will write binary to all tracks
  • Sophisticated too can read residual magnetic fields from the disks
• "Sdelete" command line wipe
• Best wipe the filesystem
• Erase, GUI program does so by writing over the files 30times
• Hard drive should physically be destroyed to recommend for the best method

File Security

• The windows swap file also known pagefile.sys and hibernation file known as hiberfil.sys contain data transferred from RAM
• Hibernation and swapfiles include: email, passwords, and encryption keys
• Be sure to secure erasure these files from the drive
• Windows BitLocker total drive encryption is an option to protect your drive
• Solid State Drives, SSD does not have magnetic media and uses nonvaltile memory
• Can not delete data

Explore hard disk partitioning schemes including primary vs. extended partitions. Understand the primary function of the boot sector. Learn file allocation methods such as contiguous and linked allocation, disk fragmentation.