CSE 213

Computer Architecture

Lecture 6: External Memory

Military Institute of Science and Technology

+
Magnetic Disk

◼ A disk is a circular platter constructed of nonmagnetic

material, called the substrate, coated with a magnetizable
material
◼ Traditionally the substrate has been an aluminium or aluminium
alloy material
◼ Recently glass substrates have been introduced

◼ Benefits of the glass substrate:

◼ Improvement in the uniformity of the magnetic film surface to
increase disk reliability
◼ A significant reduction in overall surface defects to help reduce
read-write errors
◼ Ability to support lower fly heights
◼ Better stiffness to reduce disk dynamics
◼ Greater ability to withstand shock and damage
Magnetic Read Data are recorded on and later
retrieved from the disk via a
and Write conducting coil named the head
• In many systems there are two heads, a
Mechanisms read head and a write head
• During a read or write operation the
head is stationary while the platter
rotates beneath it

Electric pulses are sent to the The write mechanism

write head and the resulting exploits the fact that
magnetic patterns are electricity flowing
recorded on the surface below, through a coil produces
+ with different patterns for
positive and negative currents
a magnetic field

An electric current in the

wire induces a magnetic
field across the gap,
which in turn magnetizes
The write head itself is made a small area of the
recording medium
of easily magnetizable
material and is in the shape of Reversing the direction
a rectangular doughnut with a of the current reverses
the direction of the
gap along one side and a few magnetization on the
turns of conducting wire recording medium
along the opposite side
Inductive Write/Magnetoresistive
Read Head
 Disk
Data
Layout
Disk Layout Methods Diagram
Winchester Disk Format
Seagate ST506
+
Physical Characteristics of Disk Systems
+
Characteristics
◼ Fixed-head disk
◼ One read-write head per track ◼ Removable disk
◼ Heads are mounted on a fixed ◼ Can be removed and replaced
ridged arm that extends across with another disk
all tracks ◼ Advantages:
◼ Unlimited amounts of data are
◼ Movable-head disk available with a limited number of
disk systems
◼ One read-write head
◼ A disk may be moved from one
◼ Head is mounted on an arm computer system to another
◼ The arm can be extended or ◼ Floppy disks and ZIP cartridge
retracted disks are examples of
removable disks
◼ Non-removable disk
◼ Permanently mounted in the ◼ Double sided disk
disk drive
◼ The hard disk in a personal ◼ Magnetizable coating is applied
computer is a non-removable to both sides of the platter
disk
+
Multiple
Platters
 Tracks

Cylinders

+
+ Disk
The head mechanism provides
a classification of disks into
Classification
three types

◼ The head must generate or Winchester Heads

sense an electromagnetic field
of sufficient magnitude to write
◼ Used in sealed drive assemblies that
and read properly are almost free of contaminants

◼ The narrower the head, the ◼ Designed to operate closer to the

closer it must be to the platter disk’s surface than conventional rigid
surface to function disk heads, thus allowing greater
◼ A narrower head means
data density
narrower tracks and
therefore greater data ◼ Is actually an aerodynamic foil that
rests lightly on the platter’s surface
density when the disk is motionless
◼ The air pressure generated by a
◼ The closer the head is to the spinning disk is enough to make
disk the greater the risk of the foil rise above the surface
error from impurities or
imperfections
Typical Hard Disk Parameters
+
Timing of Disk I/O Transfer
+
Disk Performance Parameters
◼ When the disk drive is operating the disk is rotating at constant
speed

◼ To read or write the head must be positioned at the desired track

and at the beginning of the desired sector on the track
◼ Track selection involves moving the head in a movable-head
system or electronically selecting one head on a fixed-head
system
◼ Once the track is selected, the disk controller waits until the
appropriate sector rotates to line up with the head
+
Disk Performance Parameters
◼ Seek time
◼ On a movable–head system, the time it takes to position the head at the
track

◼ Rotational delay (rotational latency)

◼ The time it takes for the beginning of the sector to reach the head

◼ Access time
◼ The sum of the seek time and the rotational delay
◼ The time it takes to get into position to read or write

◼ Transfer time
◼ Once the head is in position, the read or write operation is then
performed as the sector moves under the head
◼ This is the data transfer portion of the operation
Disk Performance Parameters
Disk Performance Parameters
Consider a disk with an advertised average seek time of 4 ms,
rotation speed of 15,000 rpm, and 512-byte sectors with 500
sectors per track. Suppose that we wish to read a file
consisting of 2500 sectors for a total of 1.28 Mbytes. Calculate
the total time for the transfer for both Sequential access
and Random access

Answer:
0.034 seconds for Sequential access
15.02 seconds for Random access
+ ◼ Consists of 7 levels

◼ Levels do not imply a hierarchical

relationship but designate different
RAID design architectures that share three
common characteristics:

1) Set of physical disk drives viewed

by the operating system as a single
logical drive

2) Data are distributed across the

Redundant Array of physical drives of an array in a
scheme known as striping
Independent Disks
3) Redundant disk capacity is used to
store parity information, which
guarantees data recoverability in
case of a disk failure
 Table 6.3 RAID Levels

N = number of data disks; m proportional to log N

RAID Levels
0, 1, 2
 RAID
Levels
3, 4, 5, 6
Data Mapping for a RAID Level 0 Array
+ R
RAID
◼ Addresses the issues of request patterns of
the host system and layout of the data
a
Level 0 ◼ Impact of redundancy does not interfere
with analysis i
d
RAID 0 for High Data Transfer
Capacity
RAID 0 for High I/O Request Rate 0
◼ For applications to experience
a high transfer rate two
◼ For an individual I/O request for a
requirements must be met:
small amount of data the I/O time
1. A high transfer capacity must
is dominated by the seek time and
exist along the entire path
rotational latency
between host memory and the
individual disk drives ◼ A disk array can provide high I/O
execution rates by balancing the
2. The application must make I/O I/O load across multiple disks
requests that drive the disk
array efficiently ◼ If the strip size is relatively large
multiple waiting I/O requests can
be handled in parallel, reducing
the queuing time for each request
 R
+
RAID a
Level 1 i
d
Characteristics Positive Aspects

◼ Differs from RAID levels 2 through 6 ◼ A read request can be serviced by

in the way in which redundancy is either of the two disks that contains
achieved the requested data

◼ Redundancy is achieved by the ◼ There is no “write penalty”

simple expedient of duplicating all
the data ◼ Recovery from a failure is simple,
when a drive fails the data can be
◼ Data striping is used but each logical accessed from the second drive
strip is mapped to two separate
physical disks so that every disk in ◼ Provides real-time copy of all data
the array has a mirror disk that
contains the same data
◼ Can achieve high I/O request rates if
the bulk of the requests are reads
◼ RAID 1 can also be implemented
without data striping, although this is
◼ Principal disadvantage is the cost
less common
+ R
RAID a
Level 2 i
d
Characteristics Performance
◼ An error-correcting code is
◼ Makes use of a parallel access calculated across corresponding
technique bits on each data disk and the bits 2
of the code are stored in the
corresponding bit positions on
◼ In a parallel access array all multiple parity disks
member disks participate in the
execution of every I/O request ◼ Typically a Hamming code is used,
which is able to correct single-bit
◼ Spindles of the individual drives errors and detect double-bit
are synchronized so that each errors
disk head is in the same position
on each disk at any given time ◼ The number of redundant disks is
proportional to the log of the
number of data disks
◼ Data striping is used
◼ Strips are very small, often as ◼ Would only be an effective choice
small as a single byte or word in an environment in which many
disk errors occur
 R
+
RAID a
Level 3 i
d
Redundancy Performance
Requires only a single In the event of a drive failure, the
◼
redundant disk, no matter how
◼
parity drive is accessed and data is 3
reconstructed from the remaining
large the disk array devices

◼ Employs parallel access, with ◼ Once the failed drive is replaced, the
data distributed in small strips missing data can be restored on the
new drive and operation resumed
◼ Instead of an error correcting ◼ In the event of a disk failure, all of the
code, a simple parity bit is data are still available in what is
computed for the set of referred to as reduced mode
individual bits in the same
position on all of the data disks ◼ Return to full operation requires that
the failed disk be replaced and the
entire contents of the failed disk be
◼ Can achieve very high data regenerated on the new disk
transfer rates
◼ In a transaction-oriented environment
performance suffers
+ R
RAID a
Level 4 i
d
Characteristics Performance

◼ Makes use of an independent ◼ Involves a write penalty when 4

access technique
an I/O write request of small
◼ In an independent access array,
size is performed
each member disk operates
independently so that separate
I/O requests can be satisfied in ◼ Each time a write occurs the
parallel array management software
must update the user data the
◼ Data striping is used corresponding parity bits
◼ Strips are relatively large
◼ Thus each strip write involves
◼ To calculate the new parity the two reads and two writes
array management software
must read the old user strip
and the old parity strip
+ R
RAID RAID a

Level 5 Level 6
i
d

Characteristics Characteristics

◼ Organized in a similar fashion ◼ Two different parity calculations

are carried out and stored in 5
to RAID 4 separate blocks on different 6
disks
◼ Difference is distribution of
the parity strips across all ◼ Advantage is that it provides
disks extremely high data availability

◼ A typical allocation is a round- ◼ Three disks would have to fail

within the mean time to repair
robin scheme (MTTR) interval to cause data to
be lost
◼ The distribution of parity
strips across all drives avoids ◼ Incurs a substantial write
the potential I/O bottleneck penalty because each write
found in RAID 4 affects two parity blocks
 Table 6.4
RAID
Comparison
(page 1 of 2)
 Table 6.4
RAID
Comparison
(page 2 of 2)
 Flash

Memory

+
Figure 6.10
Flash Memory Operation
Solid State Drive (SSD)

A memory device
made with solid state Two distinctive
components that can
be used as a Flash memory types of flash
replacement to a hard memory:
disk drive (HDD) A type of
semiconductor NOR
memory used in many •The basic unit of access is a bit
consumer electronic •Provides high-speed random
products including access
•Used to store cell phone
smart phones, GPS operating system code and on
devices, MP3 players, Windows computers for the
The term solid digital cameras, and BIOS program that runs at
USB devices start-up
state refers to
electronic
circuitry built with
NAND
semiconductors Cost and
• The basic unit is 16 or 32 bits
performance has • Reads and writes in small blocks
evolved to the point • Used in USB flash drives, memory
where it is feasible to cards, and in SSDs
use to replace HDDs • Does not provide a random-
access external address bus so
the data must be read on a block-
wise basis
+
SSD
Organization
 SSD Compared to HDD
SSDs have the following advantages over HDDs:
Table
◼ High-performance input/output operations per second
6.5
(IOPS)

◼ Durability

◼ Longer lifespan

◼ Lower power consumption Comparisons

◼ Quieter and cooler running capabilities

◼ Lower access times and latency rates

+
+ Practical Issues
There are two practical issues peculiar to SSDs
that are not faced by HDDs:
◼ Flash memory becomes
◼ SDD performance has a unusable after a certain
tendency to slow down as the number of writes
device is used ◼ Techniques for prolonging
life:
◼ The entire block must be
◼ Front-ending the flash with a
read from the flash memory cache to delay and group
and placed in a RAM buffer write operations
◼ Using wear-leveling
◼ Before the block can be algorithms that evenly
written back to flash distribute writes across block
memory, the entire block of of cells
flash memory must be ◼ Bad-block management
techniques
erased
◼ Most flash devices estimate
◼ The entire block from the their own remaining lifetimes
buffer is now written back to so systems can anticipate
the flash memory failure and take preemptive
action
Table 6. 6
Optical
Disk
Products
+
Compact Disk Read-Only Memory
(CD-ROM)
◼ Audio CD and the CD-ROM share a similar technology
◼ The main difference is that CD-ROM players are more rugged and
have error correction devices to ensure that data are properly transferred

◼ Production:
◼ The disk is formed from a resin such as polycarbonate
◼ Digitally recorded information is imprinted as a series of microscopic pits on
the surface of the polycarbonate
◼ This is done with a finely focused, high intensity laser to create a master disk
◼ The master is used, in turn, to make a die to stamp out copies onto
polycarbonate
◼ The pitted surface is then coated with a highly reflective surface, usually
aluminum or gold
◼ This shiny surface is protected against dust and scratches by a top
coat of clear acrylic
◼ Finally a label can be silkscreened onto the acrylic
+
CD Operation
+
CD-ROM Block Format
+
◼ CD-ROM is appropriate for the distribution of large CD-ROM
amounts of data to a large number of users

◼ Because the expense of the initial writing process it

is not appropriate for individualized applications
◼ The CD-ROM has two advantages:

◼ The optical disk together with the information stored

on it can be mass replicated inexpensively

◼ The optical disk is removable, allowing the disk itself

to be used for archival storage

◼ The CD-ROM disadvantages:

◼ It is read-only and cannot be updated

◼ It has an access time much longer than that of a

magnetic disk drive
+
CD Recordable CD Rewritable
(CD-R) (CD-RW)
◼ Write-once read-many ◼ Can be repeatedly written and
overwritten
◼ Accommodates applications in ◼ Phase change disk uses a material that
which only one or a small has two significantly different
number of copies of a set of data reflectivities in two different phase states
is needed ◼ Amorphous state
◼ Disk is prepared in such a way ◼ Molecules exhibit a random
that it can be subsequently orientation that reflects light poorly
written once with a laser beam ◼ Crystalline state
of modest-intensity ◼ Has a smooth surface that reflects light
well
◼ Medium includes a dye layer
which is used to change ◼ A beam of laser light can change the
reflectivity and is activated by a material from one phase to the other
high-intensity laser ◼ Disadvantage is that the material
eventually and permanently loses its
◼ Provides a permanent record of desirable properties
large volumes of user data ◼ Advantage is that it can be rewritten
+
Digital
Versatile Disk
(DVD)
High-Definition
Optical Disks
+
Magnetic Tape
◼ Tape systems use the same reading and recording techniques as
disk systems

◼ Medium is flexible polyester tape coated with magnetizable

material

◼ Coating may consist of particles of pure metal in special binders

or vapor-plated metal films

◼ Data on the tape are structured as a number of parallel tracks

running lengthwise

◼ Serial recording
◼ Data are laid out as a sequence of bits along each track

◼ Data are read and written in contiguous blocks called physical

records

◼ Blocks on the tape are separated by gaps referred to as inter-

record gaps
+
Magnetic
Tape
Features
+
Table 6.7
LTO Tape Drives
+ Summary
External Memory

Chapter 6
◼ RAID
◼ Magnetic disk
◼ RAID level 0
◼ Magnetic read and write
mechanisms ◼ RAID level 1
◼ Data organization and ◼ RAID level 2
formatting
◼ RAID level 3
◼ Physical characteristics
◼ Disk performance parameters
◼ RAID level 4
◼ RAID level 5
◼ Solid state drives ◼ RAID level 6
◼ Flash memory
◼ SSD compared to HDD ◼ Optical memory
◼ SSD organization ◼ Compact disk
◼ Practical issues
◼ Digital versatile disk

◼ Magnetic tape ◼ High-definition optical disks