CSE 401

Computer Engineering (2)

)2( ‫هندسة الحاسبات‬

4th year, Comm. Engineering

Winter 2020
Lecture #1
Dept. of Computer & Systems Engineering
Credits to Dr. Hazem Shehata and Dr. Ahmed Abdul-Monem Ahmed for the slides
Chapter 6. External Memory
Types of External Memory
• Magnetic Disk
—Magnetic Read and Write Mechanisms
—Data Organization and Formatting
—Physical Characteristics
—Disk Performance Parameters

• Redundant Array of Independent Disks (RAID)

• Solid-State Drive (SSD)

• Optical Disk

• Magnetic Tape
Magnetic Disk
• Foundation of external memory on computers.
• Disk substrate: circular platter constructed of
nonmagnetic material.
• Substrate coated with magnetisable material (iron
oxide or cobalt oxide).
• Substrate used to be aluminium or aluminium
alloy. Now glass ➔ ...
—Less surface defects ➔ less read/write errors.
—Better uniformity of magnetic film ➔ more reliability.
—Better stiffness to reduce disk dynamics.
—Better shock/damage resistance.
—Ability to support lower fly heights (see later).
Disk pack
Read and Write Mechanisms (1)
• Head: Small device (conductive coil) that can
read from or write to a portion of the platter
rotating under it.
—May be single read/write head or separate ones.
—During read/write, head is stationary, platter rotates.
• Write Operation
—FACT: Current passing through coil produces a
magnetic field whose direction depends on current
direction.
—Pulses sent to head.
—Magnetic pattern recorded on surface below.
How disk works Cont’d.

Number of cylinders = number of tracks

Read and Write Mechanisms (2)
• Read Operation
—Traditional - floppy and old disks
– FACT: Magnetic field moving relative to a coil produces
current in the coil.
– Coil is the same for read and write
– One read/write head.
—Contemporary
– Separate read head, close to write head.
– Partially shielded magnetoresistive (MR) sensor.
– Electrical resistance depends on direction of magnetic field.
– Current passing thru MR and resistance is read as voltage.
– Much greater storage densities and higher operating speeds
could be achieved.
Inductive Write - MR Read Head
How disk works.
Disk Data Layout
Data Organization and Formatting
• Tracks: concentric rings
—Typically thousands per surface
—Track’s width = head’s width.
—Inter-track gaps between adjacent tracks ➔ prevent
errors due to misalignment of the head or interference.
• Sectors: tracks are divided into sectors.
—Typically hundreds per track.
—Fixed (512 bytes) or variable length.
—Inter-sector gaps ➔ relax precision requirements
—Usually data are read from or written to disks in blocks
(of sectors).
Disk Velocity ➔ Disk Layout
• Problem: bits on inner tracks pass under head
slower than bits on outer tracks ➔ How can we
have a constant read/write rate?!
• Solution: increase spacing between bits stored
on outer tracks and rotate disk at a fixed speed.
—This technique is know as Constant Angular
Velocity (CAV).
Constant Angular Velocity (CAV)
• Disk rotates at constant speed
—Constant angular velocity
• All tracks have same capacity
—Constant bits/track
—Assuming constant bits/sector
– Constant sectors/track
– Pie-shaped sectors
• Outer tracks have lower density
—Variable bits/meter
• Pros: Simple to locate sectors ➔ Move head to
a given track and wait for a given sector.
• Cons: Waste of space on outer tracks
Multiple Zone Recording (or Zone Bit Recording)
• Tracks are grouped into zones
• Within a zone, tracks have same
capacity
—Constant sectors/track in each zone
• Tracks in outer zones have
higher capacities
—Sectors/track increase in outer zones
• Pros: Higher capacity for outer tracks ➔ Greater
overall storage capacities.
• Cons: Harder to locate sectors ➔ more complex
circuitry.
Finding Sectors
• Must be able to identify start of track and sector.
• Disk Format
— Extra control data recorded on disk to mark tracks and sectors.
— Used only by the disk drives, not accessible to the user.
• Seagate ST506 format
Physical Characteristics
1. Single or double sided.
2. Removable or not (fixed).
3. Single or multiple platter.
4. Fixed or movable head (w.r.t. radial direction).
5. Head mechanism
—Contact (Floppy).
—Fixed gap.
—Aerodynamic gap (Winchester).
1. Single or double sided
• Single-sided
—Magnetisable coating applied to one side of platter.
—Low expensive disk systems (e.g., floppy disks).
• Double-sided
—Magnetisable coating applied to both sides of platter.
—Each side has its own head.
2. Removable or Non-removable
• Removable disk
—Can be removed from drive and replaced with
another disk.
—e. g., floppy disks.
—Provides unlimited storage capacity.
—Easy data transfer between systems.
—Obsoleted by CD-R and CD-R/W!
• Non-removable disk
—Permanently mounted in the drive.
—e. g., hard disks in personal computers.
3. Single or Multiple Platters
• Multiple platters
mounted on a spindle.
• Multiple heads
—joined and aligned
—One head per side.
• Aligned tracks on each
platter form cylinder.
• Data is striped by
cylinder.
— i.e., Rd/Wr cylinder by cylinder.
— Advantages:
– Reduces head movement.
– Increases transfer rate.
4. Fixed or Movable Head Disk
• Fixed head
—One read write head per track.
—Heads mounted on fixed ridged arm.
—Rare!
• Movable head
—One read/write head per side
—Mounted on a movable arm
5. Head Mechanism (1)
• Contact
—Head comes into physical contact with storage medium
during read/write.
—Platter needs to be flexible.
—Used in floppy disks.
• Fixed gap
—Head positioned a fixed distance above platter.
—Used in early generations of hard disks.
• Aerodynamic gap (Winchester)
—<...next slide...>
Head Mechanism (2)
• Aerodynamic gap (Winchester)
—Developed by IBM in Winchester (USA)
—Head rests lightly on platter surface when disk is
motionless.
—As disk spins, head rises (flies on a thin layer of air)
due its aerodynamic design.
—Allow very small head-to-surface gap
– Smaller heads ➔ narrower tracks ➔ higher density
– Closer head proximity ➔ greater risk of errors due to
impurities ➔ Winchester heads are housed in sealed
containers (free of contaminants)
—Used in contemporary hard disks
Disk Performance Parameters
• Disk operation
— Disk rotates at a constant speed.
— To read/write, head is positioned at the desired track and at
the beginning of the desired sector on that track.
• Seek time: time taken to move the disk arm
(head) to the required track.
— Equals: startup time + traversal time.
— Average: less than 10 ms.
• Disk controller waits till the desired sector rotates
to line up with the head.
• Rotational delay (latency): time taken for the
beginning of the sector to reach the head.
— Average: 1/2r, for r rpm ➔ less than 5ms if r=7200 rpm
Disk Performance Parameters (2)
• Access time = seek time + rotational delay
• Read/write performed as the sector moves
under the head.
• Transfer time: time required for the transfer.

—T = transfer time,
—b = number of bytes to be transferred,
—r = rotational speed (r.p.s.),
—N = number of bytes on a track.
Typical Hard Disk Drive Parameters
Types of External Memory
• Magnetic Disk

• Redundant Array of Independent Disks (RAID)

• Solid-State Drive (SSD)

• Optical Disk

• Magnetic Tape
RAID
• Problem: Improvement rate in secondary storage <
rate for CPU and MM.
• Solution: Can’t improve one-disk perf. ➔ use multiple
in parallel!
• Array of disks
—Operate independently and in parallel.
—Single I/O request can be handled in parallel if the block
is distributed across multiple disks.
—Separate I/O requests can be handled in parallel.
—Performance metrics: depend on request patterns & data layout.
– I/O data transfer rate.
– I/O request rate (response time).
– Recovery from errors & disk failure.
• RAID: Redundant Array of Independent Disks.
—7 levels in common use, not a hierarchy.
—Set of physical disks viewed by OS as single logical drive.
—Data distributed across physical drives.
—Can use redundant capacity to store parity information.
RAID 0 - Stripping without Mirroring or Parity

Stripe

• No disks dedicated for redundancy or parity!

• Data is striped and distributed across disks.
• Strip: physical block, sector, or another unit.
• Strips are mapped round robin to consecutive disks.
• Stripe: Set of logically consecutive strips that maps
exactly one strip to each array member.
RAID 0 - Performance
• RAID 0 for high I/O data transfer rate
— System is configured s. t. each (single) I/O request can be processed
by multiple disks in parallel ➔ Less tr. time ➔ Higher tr. rate!
— Two requirements (to experience high data transfer rate):
1. High transfer capacity between memory and disks (internal
controller buses, I/O buses, memory buses, etc.).
2. Small strip ➔ Higher chance that any single I/O request
requires data from multiple strips located on different disks.
• RAID 0 for high I/O request rate
— System configured s. t. multiple I/O requests can be processed by
multiple disks in parallel ➔ Higher I/O req. rate!
— There are typically hundreds of I/O requests per second by multiple
independent applications, or single one.
— Balance I/O load across multiple disks ➔ Achieve high I/O req. rate.
— Large strip ➔ few seeks/disk per request ➔ less I/O queuing time.
RAID 0 - Pros and Cons
RAID 1 - Mirroring without Stripping or Parity
• Duplicate data (without striping) ➔ mirrored disks.
• 2 copies of each block on separate disks.
— Read from either disk (the one with min. access time).
— Write to both (in parallel) ➔ Time = larger access time.
• Recovery is simple: swap faulty disk & re-mirror.
• Expensive ➔ used to store system S/W & critical files.
• I/O transfer rate
— read: > single disk, write: ≈ single disk.
• I/O request rate
— read: ≈ 2x single disk, write: ≈ single disk.

block a0 block block c0 block block a0 block block c0 block

b0 d0 b0 block c1
d0
block a1 block block c1 block block a1 block block
block a2 b
block
1 block c2 d
block
1 block a2 b
block
1 block c2 d
block
1

b2 d2 b2 d2
block a3 block block c3 block block a3 block block c3 block
b3 d3 b3 d3
RAID 1 - Pros and Cons

block a0 block block c0 block block a0 block block c0 block

b0 d0 b0 block c1
d0
block a1 block block c1 block block a1 block block
block a2 b
block
1 block c2 d
block
1 block a2 b
block
1 block c2 d
block
1

b2 d2 b2 d2
block a3 block block c3 block block a3 block block c3 block
b3 d3 b3 d3
RAID 2 - Bit-Level Stripping with Hamming ECC
• Parallel access: all disks participate in every I/O request.
• Disks are synchronized: all heads in same position at any given time.
• Bit-level stripping: Very small strips ➔ single-bit strips!
• Error correction calculated across corresponding bits on disks.
• Multiple parity disks store Hamming code in corresponding bit positions.
• Number of redundant disks is proportional to log number of data disks.
• Read: data & parity delivered to controller ➔ error corrected instantly.
• Write: all data and parity disks accessed.
• Contemporary disks are highly reliable ➔ RAID 2 not used in practice!!
• I/O transfer rate: very high due to small strip size.
• I/O request rate: only one at a time ➔ ≈ single disk!
RAID 2 - Pros and Cons
RAID 3 - Byte-Level Stripping with Parity
• Similar to RAID 2: synchronized disks, small strips. Byte-
level stripping ➔ single-byte strips.
• Only 1 redundant disk, no matter how large the array.
• Simple parity bit for each set of corresponding bits.
• Drive fails ➔ replace it and reconstruct data from surviving
disks and parity info. For instance, in a 5-disk array:
— Disk #1 fails ➔ X1(i) = X4(i)  X3(i)  X2(i)  X0(i)
• I/O transfer rate: very high due to small strip size.
• I/O request rate: only one at a time ➔ ≈ single disk.
RAID 3 - Pros and Cons
RAID 4 - Block-Level Stripping with Parity
• Independent access: each disk operates independently
(not synchronized) ➔ separate I/O requests in parallel.
• Relatively large strips (block-level).
• Bit-by-bit parity calculated across strips on each disk.
• Parity stored on parity disk.
• Write: read old data and old parity, update both.
— Write to disk #1 ➔ X4’(i) = X4(i)  X1(i)  X1’(i)
— Not the case in RAID 2 and RAID 3 due to parallel access.
• I/O transfer rate: read: ≈ RAID 0, write: < single disk.
• I/O request rate: read: ≈ RAID 0, write: < single disk.
RAID 4 - Pros and Cons
RAID 5 - Block-Level Stripping with Distributed Parity
• Independent access, relatively large strips (block-level).
• Like RAID 4, except parity distributed across all disks.
• Round robin allocation for parity strips.
—n-disk array: parity strip is on a different disk for the first
n stripes, and the pattern repeats.
• Avoids RAID 4 bottleneck at parity disk.
• Commonly used in network servers.
• N.B. Does not mean 5 disks!
• I/O transfer rate: read: ≈ RAID 0, write: < single disk.
• I/O request rate: read: ≈ RAID 0, write: < single disk.
RAID 5 - Pros and Cons
RAID 6 - Block-level striping with double distributed parity

• Independent access, relatively large strips (block-level).

• Two parity calculations.
—P and Q are two different data check algorithms.
• Stored in separate blocks on different disks.
• N data disks ➔ N+2 disks required to build the array.
• High data availability
—Three disks need to fail for data loss.
—Significant write penalty.
• I/O transfer rate: read: ≈ RAID 0, write: < RAID 5.
• I/O request rate: read: ≈ RAID 0, write: < RAID 5.
RAID 6 - Pros and Cons
 RAID Levels - Summary

Similar to single
disk

Similar to single
disk

* N = number of data disks. N must be greater than 1 in all RAID configurations except RAID 1 where N could be equal to 1.
* m = number of ECC disks. N is proportional to log m.
 Project : Task 3
Literature Review:
Find 5 papers related to your suggested system:
description Analysis Advantag Design …. ……
&tools es&disadv &impleme
. ntation
Paper 1
title
Paper 2
title
Paper 3
title
Paper 4
title
Paper 5
title
Your
title
Types of External Memory
• Magnetic Disk

• Redundant Array of Independent Disks (RAID)

• Solid-State Drive (SSD)

• Optical Disk

• Magnetic Tape
Solid-State Drive (SSD)
• Purpose: complement or even replace HDDs!
• The term “solid-state” refers to electronic
circuitry built with semiconductors.
• SSD’s store data in flash memory cells.
• Each flash memory cell is built using a single
transistor: floating-gate MOSFET (FG-MOSFET).
CG

FG

D S
States of FG-MOSFET
Id “1” “0”
programming

erasing

VT1 VT0 Vgs

Logic 1 Logic 0
• No electrons trapped • Electrons trapped on
on the FG. the FG.
• Smaller threshold • Higher threshold
voltage (VT1) voltage (VT0)
—Forming the channel —Forming the channel
(i.e., turning transistor (i.e., turning transistor
on) requires applying on) requires applying
a relatively smaller a slightly higher
voltage to CG. voltage to CG.
Read/Program/Erase Flash Memory Cell
• To read: Id “1” “0”

—Apply voltage Vr to CG s.t. VT1<Vr<VT0.

—Measure (sense) drain current (Id). VT1 VT0 Vgs
– Id > 0 ➔ transistor on ➔ logic 1. Vr
– Id = 0 ➔ transistor off ➔ logic 0.
• To program (write “0”):
Only Applicable
to NOR flash!!

—Apply high +ve voltage to CG & D.

—Electrons jump from channel through
insulating layer onto FG ➔ hot-electron injection.
• To erase (write “1”):
—Apply high voltage diff. between D & CG.
—Electrons pulled off FG through insul-
ating layer to D ➔ quantum tunneling.
 Layouts of Flash Memory
• Faster read.
— Transistors directly
NOR Flash
connected to BL.
• Simpler Interface.
— Random access.
• Lower error rate.
• Data units
— Bytes/words.
• Applications
— ROM replacement.
• Capacity: < 2 Gb.

• Faster program/erase.
• Higher Density.
NAND Flash
— Less connections to
GND/BL.
• Longer life span.
• Data units
— Pages/blocks.
• Applications
— SSDs, USB drives,
SD Cards.
• Capacity: < 2 Tb.
 NAND/NOR Flash – Reading
NOR Flash
Voff Vr Voff Voff Voff Voff
Id “1” “0”

VT1 VT0 Vgs

Voff Vr

NAND Flash
Id “1” “0” Von Von Vr Von Von Von Von Von Von Von

VT1 VT0 Vgs

Vr Von
NAND/NOR Flash – Programming (X ➔ 0)
NOR Flash
• Technique: Vhigh
— Hot electron Voff Vhigh Voff Voff Voff Voff
injection.
• Explanation: 1 1 1 1 1 1
— Electrons 0
jump from
channel to
FG.

NAND Flash
• Technique:
— Quantum Vhigh
tunneling
(injection).
1 1 1 1 1 1 1 1
• Explanation: 0
— Electrons
tunnel from
B to FG.
NAND/NOR Flash – Block Erasure (X ➔ 1)
NOR Flash
• Technique:
— Quantum -Vhigh -Vhigh -Vhigh -Vhigh -Vhigh -Vhigh
tunneling
(ejection).
1 1 1 1 1 1
• Explanation:
— Electrons
tunnel from
FG to S/B.

NAND Flash
• Technique:
— Quantum -Vhigh -Vhigh -Vhigh -Vhigh -Vhigh -Vhigh -Vhigh -Vhigh
tunneling
(ejection).
1 1 1 1 1 1 1 1
• Explanation:
— Electrons
tunnel from
FG to B.
NAND Flash Block Organization
• NAND flash memory is organized as a set of
independent blocks. Each block has a set of pages.
• Blocks are the smallest erasable units.
• Pages are the smallest programmable units.
— Partial pages can be programmed in some devices!

** The numbers of bits-per-page and pages-per-block here are typical for a 2Gb NAND flash device **
NAND Flash Memory Architecture
SSD Architecture
• Controller: provides SSD device
level interfacing & firmware
execution.
• Addressing: logic to select one of
the flash memory components.
• Data buffer/cache: High-speed
RAM for speed matching and
increasing data throughput.
• Error correction: Logic for error
detection and correction.
• Flash memory components:
Individual NAND flash chips.
SSD vs. HDD
• SSDs have the following advantages over HDDs:
—Higher input/output operations per second (IOPS)
—Longer lifespan: no mechanical wear.
—Lower power consumption.
—Quieter and cooler running capabilities.
—Lower access times & latency rates: >10x faster.
Optical Storage - History
• 1983 CD (Compact Disk, audio CD)
CD-ROM
.........

• 1996 DVD-ROM
DVD-R
.........

• 2002 Blu-ray
.........
CD-ROM
• Same technology used for audio (CD).
—Difference: CD-ROM drives support error-correction.
• Capacity
—Data: 650-700MB, or Audio: 74-80 minute.
• Material
—Polycarbonate coated with highly reflective coat,
usually aluminium.
• Data stored as sequence of pits engraved along
a spiral track on top of polycarbonate layer.
• Read by reflecting laser.
• Constant packing density ➔ Constant Linear
Velocity (CLV) ➔ variable angular velocity.
CD-ROM - Fabrication
(3) (4)

(2)
(1)

• Master disk: Info printed as pits on the polycarbonate

surface using a high-intensity laser.
• Master is used to make a die to stamp out copies.
• Pitted surface coated with a reflective material (Aluminium).
• Coat of acrylic to protect against dust and scratches.
CD-ROM - Operation

• CD drive transmits a low-power laser beam towards disk.

— beam falls on a pit (rough surface) ➔ low intensity reflected.
— beam falls on a land (smooth surface) ➔ high intensity reflected.
• Photo-sensor senses surface at regular intervals.
— Change in elevation ➔ logic 1, otherwise ➔ logic 0.
CD-ROM Drive Speeds
• Audio is single speed (1x=150kB/s)
—Constant linear velocity
—1.2 m/s
—Track (spiral) is 5.27km long
—Gives 4391 seconds = 73.2 minutes
• Other speeds are quoted as multiples.
—e.g. 24x (24x150kB/s=3.6MB/s)
• Quoted figure is maximum drive can achieve.
 CD-ROM Block/Sector Format

• Mode 0=blank data field

• Mode 1=2048 byte data+error correction
• Mode 2=2336 byte data
Mag. disks: ID(track, sector, head, CRC), Data(512+2)
CD: ID(minute, second, sector, mode), Data(2048+288)
Access on CD-ROM
• Difficult!!
1. Move head to a rough position.
2. Set correct speed.
3. Read address.
4. Adjust to required location.
CD-ROM For & Against
• Pros:
—Large capacity (cf. floppy disks).
—Easy to mass produce (cf. magnetic disks).
—Removable (cf. magnetic disks).
—Robust.
• Cons:
—Expensive for small quantities.
—Slow (access time ≈ 0.5 sec).
—Read only.
Other Optical Storage
• CD-Recordable (CD-R)
—Write Once Read Many (WORM).
—Quite affordable.
—Compatible with CD-ROM drives.
—Medium includes a dye layer. Reflectivity is activated
by a high-intensity laser.
• CD-ReWritable (CD-RW)
—Erasable.
—Inexpensive.
—Mostly CD-ROM drive compatible.
—Phase change by a laser beam
– Material has two different reflectivity’s in two different phase
states (crystalline or amorphous).
– Eventually, the material loses its desirable properties
(500000~1000000 erase cycles).
DVD
• Digital Video Disk
—Used to indicate a player for movies
– Only plays video disks.
• Digital Versatile Disk
—Used to indicate a computer drive
– Will read computer disks and play video disks.
• Very high capacity (4.7G per layer).
—Small spacing between spiral loops (tracks). Pits too.
Shorter wavelength.
—Double layer: Semi-reflective layer on top of the
reflective layer. Read by adjusting focus.
—Double sided.
• Full length movie on single disk (MPEG compression).
CD vs. DVD
 Optical Storage - Types
Optical Disk

CD DVD
Capacity 650M~750MB 4.7G~17GB
Drive Compatibility Incomp. with DVD Comp. with CD
Structure One 1.2mm base Two 0.6mm bases
Gap Track gap=1.6µm 0.74µm
Pit gap=0.834µm 0.4µm
Side & layer Single layer/side Double layer/side
Products CD --
CD-ROM DVD-ROM
CD-R DVD-R
CD-RW DVD-RW
High-Definition Optical Disks
• Designed for high-definition (HD) videos
— Resolution>standard-definition (SD) videos.
– e.g., 1280 x 720 pixels or 1920 x 1080 pixels.
• Much higher capacity than DVD
— Shorter wavelength laser in the blue-violet range.
— Smaller pits ➔ higher bit density.
• HD-DVD
— 15GB single side single layer.
• Blu-ray Disk (BD)
— 25GB single side single layer.
— Data layer closer to laser.
– Tighter focus, less distortion, smaller pits.
— Types: BD-ROM (read only), BD-R (recordable), and BD-RE (re-
recordable).
Optical Memory Characteristics

Track
Summary of Optical Disks
Types of External Memory
• Magnetic Disk

• Redundant Array of Independent Disks (RAID)

• Solid-State Drive (SSD)

• Optical Disk

• Magnetic Tape
Magnetic Tape
• Features:
—Oldest secondary memory (1951).
—Slowest-speed & lowest-cost in memory hierarchy.
• Usage: offline storage (i.e., backup).
• Access method: sequential ➔ slow!
• Medium: flexible polyester tape coated with
magnetizable material (metal particles).
• Tape width: 0.38cm - 1.27cm.
• Tape package:
—Used to be open reels.
—Now housed in cartridges.
Data Layout and Recording Techniques
• Data are laid out on tape along parallel tracks
running lengthwise ➔ linear recording.
• Tape moves underneath magnetic head in drive.
• Head has multiple (8-32) read/write elements ➔
read/write multiple (8-32) tracks simultaneously.
• There are two types of linear recording:
1. Parallel recording: Each byte/word is stored on
multiple tracks in parallel, .e.g., 9-track tapes.
Serial Recording
• … types of linear recording ... (cont.):
2. Serial Recording: data laid out as sequence of
bits along each track, similar to magnetic disks.
— Data read/written in contagious blocks called records.
— Records are separated by inter-record gaps.
— Tape is formatted to assist in locating records.
Serpentine Recording
• Typical recording technique in serial tapes.
— First set of bits is recorded on tracks in forward direction.
— When tape reaches end, rd/wr elements are repositioned to new
tracks & recording is resumed in opposite direction.
— Process continues, back and forth, until tape is full.
— Data still recorded serially along individual tracks, but blocks in
sequence are stored in adjacent tracks.
— Adv.: more tracks than head/elements ➔ increasing capacity.
..
.
B(6n+2) B(6n+5) B(9n-1)
B(6n-1) B(3n+5) B(3n+2)
B(2) B(5) B(3n-1)
..
B(6n+1) B(6n+4) B(9n-2)
Head

B(6n-2) B(3n+4) B(3n+1)

B(1) B(4) B(3n-2)
..
B(6n) B(6n+3) B(9n-3)
B(6n-3) B(3n+3) B(3n)
B(0) B(3) B(3n-3)
Linear Tape-Open (LTO)
• Dominant magnetic tape technology today.
• Developed late 1990s as an open source
alternative to proprietary tape systems.