Siksha Sarovar

Siksha Sarovar (sikshasarovar.com) is a free educational web application that helps students in India learn programming and prepare for academic and competitive exams. The platform offers structured coding courses (C, C++, Python, Java, HTML, CSS, PHP, Power BI, AI, Machine Learning, Data Science), complete university curriculum notes for BCA/MCA students with previous year question papers, Class 10 and Class 12 CBSE/HBSE school notes, and dedicated preparation material for SSC, UPSC, Banking, Railway and other government exams. Browsing the site is completely free and requires no account. Users may optionally sign in with Google solely to save their learning progress, quiz scores and personal preferences across devices.

Privacy Policy | Terms of Service | Contact Siksha Sarovar | About Siksha Sarovar

v4.0.9 · PWA
Siksha Sarovar logo
Siksha Sarovar
Your Learning Universe

Siksha Sarovar is a free e-learning platform for coding courses, BCA university notes and competitive exam preparation. Optional Google sign-in saves your learning progress across devices.

Initializing knowledge base…
Compiling modules 0%

Unit 4 — Auxiliary Memory

Lesson 45 of 49 in the free Computer Organization and Architecture notes on Siksha Sarovar, written by Rohit Jangra.

Auxiliary (Secondary) Memory

Auxiliary memory is non-volatile storage that is not directly addressable by the CPU. Its contents must first be transferred into main memory (usually by DMA) before they can be used.
   Key properties:
      - NON-VOLATILE (survives power-off)
      - Very large capacity
      - Very low cost per bit
      - Much slower than main memory
      - Accessed in BLOCKS (sectors), never in single bytes

1. Magnetic Disk

   Physical structure:

      PLATTER  : a rigid circular disk coated with magnetic material
      SURFACE  : each platter has two usable surfaces
      TRACK    : one concentric circle on a surface
      SECTOR   : an arc of a track — the smallest addressable unit
                 (traditionally 512 bytes, now often 4096 bytes)
      CYLINDER : the set of tracks at the same radius on ALL surfaces
      HEAD     : one read/write head per surface, all mounted on a
                 single moving ARM ASSEMBLY
        Top view of one surface           Side view of a disk pack

        .---------------.                     ==== head
       /   .---------.   \                 ---------------- surface 0
      |   /   .---.   \   |                ---------------- surface 1
      |  |   | hub |   |  |     <- tracks  ---------------- surface 2
      |   \   '---'   /   |                ---------------- surface 3
       \   '---------'   /
        '---------------'
              sectors radiate outward

Disk capacity calculation — the standard numerical

   Capacity = surfaces x tracks/surface x sectors/track x bytes/sector

   Q: A disk has 4 platters, 1024 tracks per surface, 128 sectors per
      track, and 512 bytes per sector. What is the total capacity?

      Surfaces = 4 platters x 2 = 8
      Capacity = 8 x 1024 x 128 x 512 bytes
               = 8 x 1024 x 65,536
               = 536,870,912 bytes
               = 512 MB

Disk access time — the second standard numerical

   Total access time = SEEK TIME + ROTATIONAL LATENCY + TRANSFER TIME

   SEEK TIME          : move the arm to the correct cylinder (mechanical,
                        the largest and most variable component)
   ROTATIONAL LATENCY : wait for the required sector to rotate under the head
                        AVERAGE = half a revolution = (60 / RPM) / 2
   TRANSFER TIME      : time to read the sector as it passes under the head
                        = sector size / transfer rate
   Q: A disk rotates at 7200 RPM, has an average seek time of 8 ms,
      512-byte sectors and 128 sectors per track. Find the average
      time to read one sector.

      Time per revolution = 60 / 7200 s = 8.33 ms
      Average rotational latency = 8.33 / 2 = 4.17 ms
      Transfer time for one sector = 8.33 ms / 128 = 0.065 ms

      Total = 8 + 4.17 + 0.065 = 12.24 ms
   Q: What is the data transfer rate of the disk above?

      One track = 128 x 512 = 65,536 bytes, read in one revolution (8.33 ms)
      Rate = 65,536 / 0.00833 = 7,867,000 bytes/s  ≈ 7.5 MB/s

Disk addressing

   CHS addressing (older): Cylinder, Head, Sector
   LBA addressing (modern): Logical Block Address — a single linear
                            number; the drive translates it internally.

   Conversion:
      LBA = (C x heads_per_cylinder + H) x sectors_per_track + (S - 1)

2. Magnetic Tape

   A long plastic strip coated with magnetic material, wound on reels.

   Structure: data is written in parallel TRACKS across the tape width
              (typically 9 tracks: 8 data bits + 1 parity bit),
              organised into RECORDS separated by INTER-RECORD GAPS (IRG).

   ACCESS METHOD: strictly SEQUENTIAL. To read record 500 you must
                  pass over records 1 to 499.
PropertyValue
AccessSequential only
Access timeSeconds to minutes
CapacityVery high (LTO-9: 18 TB native)
Cost per GBLowest of all media
Durability30 years archival
UseBackup, archival, cold storage
   Tape capacity numerical:

   Q: A tape has a density of 1600 bytes per inch, records of 800 bytes,
      and an inter-record gap of 0.5 inch. How much tape does one record
      occupy, and what is the storage efficiency?

      Record length on tape = 800 / 1600 = 0.5 inch
      Total per record      = 0.5 + 0.5 = 1.0 inch
      Efficiency            = 0.5 / 1.0 = 50 %

      (Blocking several records together between gaps raises this
       efficiency dramatically — this is why tapes use large blocks.)

3. Optical Storage

MediumCapacityWavelengthNotes
CD-ROM700 MB780 nm (infrared)Data in a single spiral track
DVD4.7 GB (SL), 8.5 GB (DL)650 nm (red)Smaller pits, tighter tracks
Blu-ray25 GB (SL), 50 GB (DL)405 nm (blue-violet)Shortest wavelength → densest pits
   How optical discs work:
      Data is stored as PITS and LANDS along a spiral track.
      A laser is reflected off the surface; the transition between a pit
      and a land scatters the light and is read as a 1; no transition = 0.

   Variants:  -ROM (read only), -R (write once), -RW (rewritable,
              uses a phase-change material)

4. Solid State Drives (SSD)

   Built from NAND FLASH memory — floating-gate transistors that hold
   charge without power.

   NO moving parts:
      + no seek time, no rotational latency
      + shock resistant, silent, low power
      - limited write endurance (each cell survives 1,000 - 100,000
        program/erase cycles)
      - erases happen in large BLOCKS, writes in smaller PAGES
        -> WRITE AMPLIFICATION
      - needs WEAR LEVELLING (spread writes evenly across all cells)
        and GARBAGE COLLECTION, managed by the drive's controller
Cell typeBits per cellEnduranceCost
SLC1~100,000 cyclesHighest
MLC2~10,000High
TLC3~3,000Medium
QLC4~1,000Lowest

5. HDD vs SSD

BasisHard disk (HDD)Solid state (SSD)
TechnologyMagnetic, rotating plattersNAND flash, no moving parts
Access time5–10 ms50–100 µs
Random accessSlow (seek + latency)Fast and uniform
Sequential rate100–250 MB/s500 MB/s – 7 GB/s (NVMe)
Noise / powerAudible, higherSilent, lower
Shock resistancePoorExcellent
Cost per GBLowHigher
EnduranceUnlimited writes (mechanical wear instead)Limited write cycles
Best forBulk/archival storageOS, applications, databases

6. RAID (bonus — commonly asked)

   RAID = Redundant Array of Independent Disks

   RAID 0 : STRIPING     — data split across disks. Fast, NO redundancy.
                           Capacity = n x disk;  one failure loses everything.
   RAID 1 : MIRRORING    — identical copies. Capacity = disk (50%).
                           Survives one failure per mirror pair.
   RAID 5 : STRIPING + DISTRIBUTED PARITY — survives ONE disk failure.
                           Capacity = (n - 1) x disk.
   RAID 6 : DOUBLE PARITY — survives TWO simultaneous failures.
                           Capacity = (n - 2) x disk.
   RAID 10: MIRROR of STRIPES — fast and redundant, 50% capacity.

Summary

   Auxiliary memory : non-volatile, block-accessed, not CPU-addressable
   Disk capacity    = surfaces x tracks x sectors x bytes/sector
   Disk access time = seek + rotational latency + transfer
   Avg rotational latency = (60 / RPM) / 2
   Tape             : sequential only, cheapest per GB, archival
   SSD              : no mechanical delay, limited write endurance

Next: the one memory in the hierarchy that is searched by content rather than by address.