MEMORY HIERARCHY

Memory in a computer system is required for storage and subsequent retrieval of the instructions and
data. A computer system uses a variety of devices for storing these instructions and data that are
required for its operation. Normally we classify the information to be stored into two basic categories:
Data and Instructions.

The storage devices along with the algorithm or information on how to control and manage these
storage devices constitute the memory system of a computer.

A typical storage hierarchy is shown in Figure

As we move up the hierarchy, we encounter storage elements that have faster access time, higher cost
per bit stored, and slower access time as a result of moving down the hierarchy. Thus, cache memory
generally has the fastest access time, the smallest storage capacity, and the highest cost per bit stored.
The primary memory (main memory) falls next in the storage hierarchy list. On-line, direct-access
secondary storage devices such as magnetic hard disks make up the level of hierarchy just below the
main memory. Off-line, direct-access and sequential access secondary storage devices such as magnetic
tape, floppy disk, zip disk, WORM disk, etc. fall next in the storage hierarchy. Mass storage devices, often
referred to as archival storage, are at the bottom of the storage hierarchy. They are cost-effective for
the storage of very large quantities of data when fast access time is not necessary.

RAM (Random Access Memory)

The main memory is a random access memory. It is normally organized as words of fixed length. The
length of a word is called word length. Each of these memory words has an independent address and
each has the same number of bits. Normally the total numbers of words in memory are some power of
2. Some typical memory word sizes are 8 bits, 16 bits, 32 bits etc. The main memory can be both read
and written into, therefore it is called read-write memory.
The access time and cycle time in RAMs are constant and independent of the location accessed.
DRAM (Dynamic Random Access Memory)

RAM technology is divided into two technologies: dynamic and static. DRAM is a technical term for a
type of random access memory (RAM) that can retain its contents only for a very brief period (measured
in milliseconds) and must, therefore, be continually refreshed by reading its contents at short intervals.

DRAM makes use of a capacitor and stores every bit of data on the different-2 capacitor. It is the most
efficient way to store data as the space required for the storage of data is less as compared to the static
RAM. The power requirement of the DRAM is more as we have to keep the capacitors refreshed so that
they can continue their charge as the leakage of charge is there within the capacitor.
Despite the fact that DRAM requires extra REFRESH circuitry to design into the computer's PROCESSOR
BUS, it has until recently been the most widely used type of memory because it is so much cheaper to
manufacture than non-volatile STATIC RAM. DRAM is the most common type of memory found on
SIMMs.
The CELLS that store the bits in a DRAM chip are simple single-transistor CAPACITORS (whose charge
leaks away, hence the refresh requirement) and more of them can pack onto a single chip than the
multi-transistor cells used in STATIC RAM, reducing the cost per megabyte. DRAM is only now falling out
of favor - on the grounds of speed rather than price because it cannot read fast enough to keep up with
the latest generations of the microprocessor.
Difference between SRAM and DRAM
SRAM DRAM

• SRAM is Static RAM. • DRAM is dynamic RAM.

• DRAM can do byte-level writing and multiple-byte

• SRAM can do byte-level read and write.
level reading.

• There is no need of refreshing in SRAM as it • DRAM needs refreshing as it operates on the

operates on the principle of switching the current. principle of charging of the capacitor.

• SRAM is expensive. • It is less expensive than SRAM.

• SRAM is faster. • DRAM is slower.

• SRAM used in the cache memory. • DRAM used in main memory.

• The structure of DRAM module needs a transistor

• The structure of SRAM needs a lot of transistors.
and a capacitor for storing every bit of data.

• SRAM consumes less power than DRAM in sleep • DRAM requires less power than SRAM inactive
mode. state.

Advantages of DRAM
• Its design is straightforward.
• It has high reliability.
• Its cost is low.
• There is less power dissipation in this.
• It requires less area.
• There is a high integration density in DRAM.
• It has simple memory cell structure.
• It is very dense.
Disadvantages of DRAM
• Inter-signal coupling exists between DRAMS
• There is high power consumption in DRAM.
• Volatile memory of DRAM.
• The data in DRAM require refreshing.
• Its operational speed is relatively low.
• Its manufacturing process is complicated.
 ROM (Read-Only Memory)

ROM is essentially a memory or storage device in which a fixed set of binary information is stored. A
block diagram of ROM is as shown in Figure. It consists of n input lines and m output lines. Each bit
combination of the input variables is called an
address. Each bit combination that comes out of the
output lines is called a word. The number of bits per word is equal to the number of output lines m. The
n.
number of distinct addresses possible with n input variables is 2

n
A ROM is characterised by the number of words (2 ) and the number of bits (m) per word. For example,
5
a 32 × 8 ROM which can be written as 2 × 8 consists of 32 words of 8 bit each, which means there are 8
5
output lines and 32 distinct words stored in the unit. There are only 5 input lines because 32 = 2 and
with 5 binary variables, we can specify 32 addresses.

ROM Types
There are five basic ROM types:

 ROM
 PROM
 EPROM
 EEPROM
 Flash memory

 ROM: Classic or “mask-programmed” ROM chips contain integrated circuits. A ROM chip
sends a current through a specific input-output pathway determined by the location of
fuses among the rows and columns on the chip. The current can only travel along a fuse-
enabled pathway and thus can only return via the output the manufacturer chooses.
Rewiring is functionally impossible, and so there’s no way to modify these types of ROM
chips. While producing a template for an original ROM chip is laborious, chips made
according to an existing template can be much more affordable.
 PROM: Programmable ROM, or PROM , is essentially a blank version of ROM that you can
purchase and program once with the help of a special tool called a programmer. A blank
PROM chip allows current to run through all possible pathways; the programmer chooses
a pathway for the current by sending a high voltage through the unwanted fuses to
“burn” them out. Static electricity can create the same effect by accident, so PROMs are
more vulnerable to damage than conventional ROMs.
 EPROM: Erasable Programmable ROM chips allow you to write and rewrite them many
times. These chips feature a quartz window through which a specialized EPROM
programmer emits a specific frequency of ultraviolet light. This light burns out all the tiny
charges in the EPROM to reopen its circuits. This exposure effectively renders the chip
blank again, after which you can reprogram it according to the same process as a PROM.
EPROM chips will eventually wear out, but they frequently have lifetimes of over 1000
erasures.
 EEPROM: To modify an Electrically Erasable Programmable ROM chip, apply localized
electrical fields to erase and rewrite the data. EEPROMs have several advantages over
other types of ROM. Unlike the earlier forms, you can rewrite EEPROM without dedicated
equipment, without removing it from the hardware, and in specifically designated
increments. You don’t have to erase and rewrite everything to make a single edit.

 Flash Memory: This memory is another form of semiconductor memory, which was first
introduced in the mid-1980. These memories can be reprogrammed at high speed and hence
the name flash. This is a type of non-volatile, electronic random access memory.
Basically this memory falls in between EPROM and EEPROM. In flash memory the entire memory
can be erased in a few seconds by using electric erasing technology. Flash memory is used in
many I/O and storage devices. Flash memory is also used to store data and programming
algorithms in cell phones, digital cameras and MP3 music players.
Flash memory serves as a hard drive for consumer devices. Music, phone lists, applications,
operating systems and other data are generally stored on stored on flash chips. Unlike the
computer memory, data are not erased when the device is turned off.
There are two basic kinds of flash memory:

Code Storage Flash made by Intel, AMD, Atmel, etc. It stores programming algorithms and is
largely found in cell phones.
Data Storage Flash made by San Disk, Toshiba, etc. It stores data and comes in digital cameras
and MP3 players.
SECONDARY MEMORY
It is desirable that the operating speed of the primary storage of a computer system be as fast as
possible because most of the data transfer to and from the processing unit is via the main memory. For
this reason, storage devices with fast access times, such as semiconductors, are generally used for the
design of primary storage. These
high-speed storage devices are expensive and hence the cost per bit of storage is also high for a primary
storage. But the primary memory has the following limitations:

a) Limited capacity: The storage capacity of the primary storage of today ’s computers is not sufficient to
store the large volume of data handled by most of the data processing organisations.
b) Volatile: The primary storage is volatile and the data stored in it is lost when the electric power is
turned off. However, the computer systems need to store data on a permanent basis for several
days, months or even several years.

The result is that an additional memory called secondary storage is used with most of the
computer systems. Some popular memories are described in this section.

Magnetic Disk
A disk is circular platter constructed of nonmagnetic material, called the substrate, coated with a
magnetisable material. This is used for storing large amount of data. Traditionally, the substrate has
been an aluminium or aluminium alloy material; more recently, glass substrates have been introduced.
The glass substrate has a number of benefits, including the following:
• Improvement in the uniformity of the magnetic film surface to increase disk reliability.
• A significant reduction in overall surface 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 and Write Mechanisms
Data are recorded on and later retrieved from the disk via a conducting coil named the head; in many
systems there are two heads, a read head and a write head.
During a read or write operation, the head is stationary while the platter rotates beneath it.

Data Organization and Formatting

The head is a relatively small device capable of reading from or writing to a portion
of the platter rotating beneath it. This gives rise to the organization of data on the
platter in a concentric set of rings, called tracks; each track is of the same width as
the head. There are thousands of tracks per surface.

Adjacent tracks are separated by gaps. This prevents, or at least minimizes, errors due to misalignment
of the head. Data are transferred to and from the disk in sectors. To identify the sector position normally
there may be a starting point of a track and a starting and end point of each sector.
The disk itself is mounted in a disk drive, which consists of the arm, a shaft that rotates the disk, and the
electronics needed for input and output binary data. A non-removable disk is permanently mounted in
the disk drive; the hard disk in a personal computer is a non-removable disk. A removable disk can be
removed and replaced with another disk. The advantage of the latter type is that unlimited amounts of
data are available with a limited number of disk systems. Furthermore, ZIP cartridge disks are examples
of removable disks.
The head mechanism provides a classification of disks into three types. Traditionally, the read-write
head has been positioned at a fixed distance above the platter, allowing an air gap. At the other extreme
is a head mechanism that actually comes into physical contact with the medium during a read or write
operation. This mechanism is used with the floppy disk, which is a small, flexible platter and the least
expensive type of disk.
To understand the third type of disk, we need to comment on the relationship between data density and
the distance of head from the surface. The head generates or senses an electromagnetic field of
sufficient magnitude to write and read properly. The narrower the head is, the closer it must be to the
platter surface to function. A narrower head means narrower tracks and therefore greater data density,
which is desirable. However, the closer the head is to the disk, the greater are the risks of errors from
impurities or imperfections.
To push the technology further, the Winchester disk was developed. Winchester heads are used in
sealed drive assemblies that are almost free of contaminants. They are designed to operate closer to the
disk’s surface than conventional rigid disk heads, thus allowing greater data density. The head is actually
an aerodynamic foil that rests lightly on the platter’s surface when the disk is motionless. The air
pressure generated by a spinning disk is enough to make the foil rise above the surface. The resulting
non-contact system can be engineered to use narrower heads that operate closer to the platter’s
surface than conventional rigid disk heads.
Accessing the Disk Data
Disks operate is semi-random mode of operation and normally are referenced block wise. The data
access time on a disk consists of two main components:
• Seek time: Time to position the head on a specific track. On a fixed head disks

it is the time taken by the electronic circuit to select the required head while in movable head
disks it is the time required to move the head to a particular track.
• Latency time: This is the time required by a sector to reach below the

read/write head. On an average it is half of the time taken for a rotation by the disk.

Optical Memories
This CD was a non-erasable disk that could store more than 60 minutes of audio information on one
side. The huge commercial success of this CD enabled the development of low-cast optical-disk storage
technology that has revolutionised computer data storage. A variety of optical-disk systems has been
introduced.

Compact Disk ROM (CD-ROM)

Both the audio CD and the CD-ROM (compact disk read-only memory) 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 from
disk to computer. Both types of disk are made the same way. The disk is formed
from a resin, such as polycarbonate. Digitally recorded information (either music or
computer data) is imprinted as a series of microscopic pits on the surface of the
polycarbonate. The pitted surface is then coated with a highly reflective surface,
usually aluminium. This shiny surface is protected against dust and scratches by a
topcoat of clear acrylic. Finally, a label can be silk-screened onto the acrylic.
Information is retrieved from a CD or CD-ROM by a low-powered laser housed in an optical-disk player,
or drive unit.
The intensity of the reflected light of the laser changes as it encounters a pit.
Specifically, if the laser beam falls on a pit, the light scatters and a low intensity is
reflected back to the source. The areas between pits are called lands. A land is a
smooth surface, which reflects back at a higher intensity. The change between pits
and lands is detected by a photo sensor and converted into a digital signal. The
sensor tests the surface at regular intervals. The beginning or end of a pit
represents a 1; when no change in elevation occurs between intervals, a 0 is
recorded.
Data on the CD-ROM are organised as a sequence of blocks.
CD-ROM is appropriate for the distribution of large amounts of data to a large number of users. CD-
ROMs are a common medium these days for distributing information. Compared with traditional hard
disks,
the CD-ROM has three advantages:
1. Large data/information storage capability.
2. The optical disk together with information stored on it can be mass replicated inexpensively, unlike a
magnetic disk. The database on a magnetic disk has to be reproduced by copying data from one
disk to second disk, using two disk drives.
3. The optical disk is removable, allowing the disk itself to be used for archival storage. Most magnetic
disks are non-removable. The information on non-removable magnetic disks must first be copied on
tape before the disk drive / disk can be used to store new information.

The disadvantages of CD- ROM are as follows:

1. It is read-only and cannot be updated.
2. It has an access time much longer than that of a magnetic disk drive
Compact Disk Recordable (CD-R)

To accommodate applications in which only one or a small number of copies of a set

data is needed, the write-once read-many CD, known as the CD Recordable (CD-R),
has been developed. For CD-R a disk is prepared in such a way that it can be
subsequently written once with a laser beam of modest intensity. Thus, with a
somewhat more expensive disk controller than for CD-ROM, the customer can write
once as well as read the disk.
The CD-R medium is similar to but not identical to that of a CD or CD-ROM. For CDs
and CD-ROMs, information is recorded by the pitting of the surface of the medium,
which changes reflectivity. For a CD-R, the medium includes a dye layer. The
resulting disk can be read on a CD-R drive or a CD-ROM drive.
The CD-R optical disk is attractive for archival storage of documents and files. It
provides a permanent record of large volumes of user data.
Compact Disk Rewritable (CD-RW)
The CD-RW optical disk can be repeatedly written and overwritten, as with a
magnetic disk. Although a number of approaches have been tried, the only pure
optical approach that has proved attractive is called phase change. The phase
change disk uses a material that has two significantly different reflectivities in two
different phase states. There is an amorphous state, in which the molecules exhibit
a random orientation and which reflects light poorly; and a crystalline state, which
has a smooth surface that reflects light well. A beam of laser light can change the
material from one phase to the other. The primary disadvantage of phase change
optical disks is that the material eventually and permanently loses its desirable
properties. Current materials can be used for between 500,000 and 1,000,000 erase
cycles.

The CDRW has the obvious advantage over CD-ROM and CD-R that it can be
rewritten and thus used as a true secondary storage. As such, it competes with
magnetic disk. A key advantage of the optical disk is that the engineering
tolerances for optical disks are much less severe than for high-capacity magnetic
disks. Thus, they exhibit higher reliability and longer life.
Digital Versatile Disk (DVD)
With the capacious digital versatile disk (DVD), the electronics industry has at last found an acceptable
replacement for the videotape used in videocassette recorders (VCRs) and, more important for this
discussion, replace the CD-ROM in personal computers and servers. The DVD has taken video into
the digital age. It delivers movies with impressive picture quality, and it can be randomly accessed
like audio CDS, which DVD machines can also play. Vast volumes of data can be crammed onto the
disk, several times as much as a CD-ROM. With DVD’s huge storage capacity and vivid quality, PC
games will become more realistic and educational software will incorporate more video.

CCDs, Bubble Memories

Charge-coupled Devices (CCDs)

CCDs are used for storing information. They have arrays of cells that can hold
charge packets of electron. A word is represented by a set of charge packets, the
presence of each charge packet represent the bit-value 1. The charge packets do
not remain stationary and the cells pass the charge to the neighbouring cells with
the next clock pulse. Therefore, cells are organized in tracks with a circuitry for
writing the data at the beginning and a circuitry for reading the data at the end.
Logically the tracks (one for each bit position) may be conceived as loops since the
read circuitry passes the information back to the write circuit, which then re-creates
the bit values in the track unless new data is written to the circuit.
These devices come under the category of semi-random operation since the devices
must wait till the data has reached the circuit for detection of charge packets. The
access time to these devices is not very high. At present this technology is used
only in specific applications and commercial products are not available.
Magnetic Bubble Memories

In certain material such as garnets on applying magnetic fields certain cylindrical

areas whose direction of magnetization is opposite to that of magnetic field are
created. These are called magnetic bubbles. The diameter of these bubbles is found
to be in the range of 1 micrometer. These bubbles can be moved at high speed by
applying a parallel magnetic field to the plate surface. Thus, the rotating field can
be generated by an electromagnetic field and no mechanical motion is required.
In these devices deposition of a soft magnetic material called Perm alloy is made as
a predetermined path, thus making a track. Bubbles are forced to move
continuously in a fixed direction on these tracks. In these memories the presence of
a bubble represents a 1 while absence represents a 0 state. For writing data into a
cell, a bubble generator to introduce a bubble or a bubble annihilator to remove a
bubble, are required. A bubble detector performs the read operation. Magnetic
bubble memories having capacities of 1M or more bits per chip have been
manufactured. The cost and performance of these memories fall between semi-
conductor RAMs and magnetic disks.

These memories are non-volatile in contrast to semi-conductor RAMs. In addition,

since there are no moving parts, they are more reliable than a magnetic disk.
But these memories are difficult to manufacture and difficult to interface with in
conventional processors. These memories at present are used in specialized
applications, e.g., as a secondary memory of air or space borne computers,
where extremely high reliability is required.