Connection-Oriented Networking and ATM

 A Single, Global Network

o Telecom companies invented ATM as a way to "grow instead of die"

-- cope with decreasing profit margins by diversifying into data and
video networking. They aimed to provide:

 Universal Service (everybody, everywhere)

 Support For All Users (voice, video, data)

 Single, Unified Infrastructure (same technology for all services

on any network, be it LAN or WAN)

 Service Guarantees (the same delivery guarantees as other

voice, video and data networks )

 Support For Low-Cost Devices (e.g. cheap ATM phones that

rely on the network to perform well)

 ISDN And ATM

o Integrated Services Digital Network (ISDN) was actually the first

technology developed aimed at achieving the goals above. By the time
it emerged it was obsolescent. ATM is the follow-on to ISDN which
has tried to satisfy the goals.

 ATM Design And Cells

o Packet "jitter" is variance in the delay experienced by packets

traveling in a network. Varying amounts of traffic congestion in
packet-switched networks causes varying delay -- jitter. Jitter
degrades voice and video.

o In data networks, larger packets are considered "good" because the

amount of header overhead tends to be low. However when sending
sound or video, the time it takes to fill a large packet with samples
introduces a delay between the last sample in packet N and the first
sample in packet N+1. Delay in voice or data streams creates "static."
 Also, it is very difficult to implement echo cancellation for voice
communication when there is high delay.

o So that ATM would be able to support voice, video, and data,

designers of ATM standards chose a packet size of 53 octets. In ATM
parlance the
packet is
called a cell.

o Critics doubt
that one
technology
can be adequate
for voice, video,
and data.

Connection-Oriented Service
o Asynchronous Transfer Mode is connection-oriented.

o Host X requests a connection and host Y agrees to accept the

connection.

o The network hardware establishes a data-path/connection and returns

a connection identifier to X and Y. (However, the connection
identifier returned to X is usually not the same as the one returned to
Y.)

o The hosts place the connection identifier in the packets and the ATM
switches use it as a basis for routing packets.

 VPI/VCI

o ATM uses virtual circuits.

o The connection identifier is a 24-bit quantity. The first 8 bits (the VPI)
identify the virtual path, and the last 16 bits (the VCI) identify the
particular virtual channel (aka circuit) using the path.
 Labels and Label Switching

o When an ATM switch forwards a packet it extracts the 24-bit

VPI/VCI connection identifier, uses it as an "index" into a forwarding
table, reads off a port number, sends the packet out the corresponding
switch port, and also replaces the VPI/VCI in the packet with a new
VPI/VCI.

o Each

communicating host sees the connection identifier by which it knows

the connection, and not the other host's connection identifier.

 An Example Trip Through An ATM Network

o The figure refers to a VC from S to R. S refers to the VC as #3 and R
calls it #6. When switch A receives a cell from S containing VPI/VCI
#3, it looks up #3 in its forwarding table. The table tells switch A to
 change the VPI/VCI to #4 and route the cell out interface 2, which it
does. Switch B receives the cell with VPI/VCI equal to #4. Switch B
uses #4 to do its lookup, and the cell goes out interface 2 with
VPI/VCI set to #1. In switch C, it is routed out interface 2 with
VPI/VCI #6, which R recognizes as the identifier of the channel
between S and R.

 Permanent Virtual Circuits

o The virtue of the label-switching idea is that a path can be configured

one hop at a time from ATM switch to ATM switch without any need
for a global value of the VPI/VCI. At each hop it is only necessary to
pick a VPI/VCI that is not in use on the local switch.

o ATM networks offer a service called permanent virtual channel

(PVC). A PVC works like a dedicated point-to-point leased line, but it
is actually a path permanently "programmed" into the network.
Network administrators form the path by first designating a set of
ATM switches to be the "nodes" of the path and then manually
configuring the switches so they will route the connection
appropriately along the path. (This configuring of the switches is
called provisioning in ATM parlance.)

 Switched Virtual Circuits

o ATM also supports dynamic allocation-deallocation of virtual

circuits. This facility is called Switched Virtual Channel (SVC).

o Host X sends a message to a directly connected switch S, requesting a

connection to host Y. S finds a path to Y and sends the request along
the path. At each hop, pairs of switches cooperate to set the VPI/VCI's
to be used, and they make their forwarding table entries for the
connection. When the connection is complete a message travels back
from Y to X notifying X that the connection is ready to use.

o There are reserved virtual channels (similar to PVC's) between

directly connected switches for sending the control messages required
to do the "signaling" that builds the SVC's as just described.

 Quality of Service (QoS)

 o ATM supports connections with quality of service (QoS) assurance.

o For example one could establish an ATM connection that guarantees a

minimum of 64Kbps throughput and a maximum of 500ms delay.

o Some of the available QoS options are:

 Constant Bit Rate (CBR) [good for uncompressed audio or

video]

 Variable Bit Rate (VBR) [good for compressed audio or video]

 Available Bit Rate (ABR) [good for data]

Asynchronous Transfer Mode (ATM) in

Computer Network
1. Driven by the integration of services and performance requirements of both
telephony and data networking: “broadband integrated service vision” (B-
ISON).

2. Telephone networks support a single quality of service and are expensive to

boot.

3. Internet supports no quality of service but is flexible and cheap.

4. ATM networks were meant to support a range of service qualities at a

reasonable cost- intended to subsume both the telephone network and the
Internet.

Asynchronous Transfer Mode (ATM):

It is an International Telecommunication Union- Telecommunications Standards

Section (ITU-T) efficient for call relay and it transmits all information including
multiple service types such as data, video, or voice which is conveyed in small
fixed-size packets called cells. Cells are transmitted asynchronously and the
network is connection-oriented.
ATM is a technology that has some event in the development of broadband ISDN
in the 1970s and 1980s, which can be considered an evolution of packet
switching. Each cell is 53 bytes long – 5 bytes header and 48 bytes payload.
Making an ATM call requires first sending a message to set up a connection.

Subsequently, all cells follow the same path to the destination. It can handle both
constant rate traffic and variable rate traffic. Thus it can carry multiple types of
traffic with end-to-end quality of service. ATM is independent of a transmission
medium, they may be sent on a wire or fiber by themselves or they may also be
packaged inside the payload of other carrier systems. ATM networks use “Packet”
or “cell” Switching with virtual circuits. Its design helps in the implementation of
high-performance multimedia networking.

ATM Cell Format

As information is transmitted in ATM in the form of fixed-size units called cells.
As known already each cell is 53 bytes long which consists of a 5 bytes header and
48 bytes payload.

Asynchronous Transfer Mode can be of two format types which are as follows:
1. UNI Header: This is used within private networks of ATMs for
communication between ATM endpoints and ATM switches. It includes the
Generic Flow Control (GFC) field.

2. NNI Header: is used for communication between ATM switches, and it does
not include the Generic Flow Control(GFC) instead it includes a Virtual Path
Identifier (VPI) which occupies the first 12 bits.

Working of ATM:

ATM standard uses two types of connections. i.e., Virtual path connections (VPCs)
which consist of Virtual channel connections (VCCs) bundled together which is a
basic unit carrying a single stream of cells from user to user. A virtual path can be
created end-to-end across an ATM network, as it does not rout the cells to a
particular virtual circuit. In case of major failure, all cells belonging to a particular
virtual path are routed the same way through the ATM network, thus helping in
faster recovery.

Switches connected to subscribers use both VPIs and VCIs to switch the cells
which are Virtual Path and Virtual Connection switches that can have different
virtual channel connections between them, serving the purpose of creating a virtual
trunk between the switches which can be handled as a single entity. Its basic
operation is straightforward by looking up the connection value in the local
translation table determining the outgoing port of the connection and the new
VPI/VCI value of connection on that link.

ATM vs DATA Networks (Internet) –

 ATM is a “virtual circuit” based: the path is reserved before transmission.

While Internet Protocol (IP) is connectionless and end-to-end resource
reservations are not possible. RSVP is a new signaling protocol on the internet.

 ATM Cells: Fixed or small size and Tradeoff is between voice or data. While IP
packets are of variable size.

 Addressing: ATM uses 20-byte global NSAP addresses for signaling and 32-bit
locally assigned labels in cells. While IP uses 32-bit global addresses in all
packets.

ATM Layers:

1. ATM Adaption Layer (AAL) –

It is meant for isolating higher-layer protocols from details of ATM processes
and prepares for conversion of user data into cells and segments it into 48-byte
cell payloads. AAL protocol excepts transmission from upper-layer services
and helps them in mapping applications, e.g., voice, data to ATM cells.
2. Physical Layer –
It manages the medium-dependent transmission and is divided into two parts
physical medium-dependent sublayer and transmission convergence sublayer.
The main functions are as follows:
 It converts cells into a bitstream.
 It controls the transmission and receipt of bits in the physical medium.
 It can track the ATM cell boundaries.
 Look for the packaging of cells into the appropriate type of frames.
3. ATM Layer –
It handles transmission, switching, congestion control, cell header processing,
sequential delivery, etc., and is responsible for simultaneously sharing the
virtual circuits over the physical link known as cell multiplexing and passing
cells through an ATM network known as cell relay making use of the VPI and
VCI information in the cell header.

ATM Applications:
1. ATM WANs –
It can be used as a WAN to send cells over long distances, a router serving as
an end-point between ATM network and other networks, which has two stacks
of the protocol.

2. Multimedia virtual private networks and managed services –

It helps in managing ATM, LAN, voice, and video services and is capable of
full-service virtual private networking, which includes integrated access to
multimedia.

3. Frame relay backbone –

Frame relay services are used as a networking infrastructure for a range of data
services and enabling frame-relay ATM service to Internetworking services.

4. Residential broadband networks –

ATM is by choice provides the networking infrastructure for the establishment
of residential broadband services in the search of highly scalable solutions.

5. Carrier infrastructure for telephone and private line networks –

To make more effective use of SONET/SDH fiber infrastructures by building
the ATM infrastructure for carrying the telephonic and private-line traffic.
What is an IP Address?
All the computers of the world on the Internet network communicate with each
other with underground or underwater cables or wirelessly. If I want to download a
file from the internet or load a web page or literally do anything related to the
internet, my computer must have an address so that other computers can find and
locate mine in order to deliver that particular file or webpage that I am requesting.
In technical terms, that address is called IP Address or Internet Protocol
Address.

Working of IP addresses

The working of IP addresses is similar to other languages. It can also use some set
of rules to send information. Using these protocols we can easily send, and receive
data or files to the connected devices. There are several steps behind the scenes.
Let us look at them
 Your device directly requests your Internet Service Provider which then grants
your device access to the web.
 And an IP Address is assigned to your device from the given range available.
 Your internet activity goes through your service provider, and they route it back
to you, using your IP address.
 Your IP address can change. For example, turning your router on or off can
change your IP Address.
 When you are out from your home location your home IP address doesn’t
accompany you. It changes as you change the network of your device.

IP ADDRESS NOTATIONS:
An IP address can be expressed in dotted decimal, binary, octal, or
hexadecimal. While all are correct and mean the same thing, it's most
common to use dotted decimal notation for IPv4 and hexadecimal (hex)
for IPv6.

Types of IP Address
IP Address is of two types:
1. IPv4: Internet Protocol version 4. It consists of 4 numbers separated by the dots.
Each number can be from 0-255 in decimal numbers. But computers do not
understand decimal numbers, they instead change them to binary numbers which
are only 0 and 1. Therefore, in binary, this (0-255) range can be written as
(00000000 – 11111111). Since each number N can be represented by a group of 8-
digit binary digits. So, a whole IPv4 binary address can be represented by 32-bits
of binary digits. In IPv4, a unique sequence of bits is assigned to a computer, so a
total of (2^32) devices approximately = 4,294,967,296 can be assigned with IPv4.
IPv4 can be written as:
189.123.123.90

Classes of IPv4 Address:

There are around 4.3 billion IPv4 addresses and managing all those addresses
without any scheme is next to impossible. Let’s understand it with a simple
example. If you have to find a word from a language dictionary, how long will it
take? Usually, you will take less than 5 minutes to find that word. You are able to
do this because words in the dictionary are organized in alphabetical order. If you
have to find out the same word from a dictionary that doesn’t use any sequence or
order to organize the words, it will take an eternity to find the word. If a dictionary
with one billion words without order can be so disastrous, then you can imagine
the pain behind finding an address from 4.3 billion addresses. For easier
management and assignment IP addresses are organized in numeric order and
divided into the following 5 classes :

IP Class Address Range Maximum number of networks

Class A 1-126 126 (27-2)

Class B 128-191 16384

Class C 192-223 2097152

IP Class Address Range Maximum number of networks

Class D 224-239 Reserve for multitasking

Reserved for Research and

Class E 240-254
development

The 0.0.0.0 is a Non-routable address is that indicates an invalid, or inapplicable

end-user address.

A loopback address is a distinct reserved IP address range that starts from

127.0.0.0 ends at 127.255.255.255 though 127.255.255.255 is the broadcast
address for 127.0.0.0/8. The loopback addresses are built into the IP domain
system, enabling devices to transmit and receive the data packets.
The loopback address 127.0.0.1 is generally known as localhost.

3. IPv6:
But, there is a problem with the IPv4 address. With IPv4, we can connect
only the above number of 4 billion devices uniquely, and apparently, there
are much more devices in the world to be connected to the internet. So,
gradually we are making our way to IPv6 Address which is a 128-bit IP
address. In human-friendly form, IPv6 is written as a group of 8
hexadecimal numbers separated with colons(:). But in the computer-friendly
form, it can be written as 128 bits of 0s and 1s. Since, a unique sequence of
binary digits is given to computers, smartphones, and other devices to be
connected to the internet. So, via IPv6 a total of (2^128) devices can be
assigned with unique addresses which are actually more than enough for
upcoming future generations.

IPv6 can be written as:

2011:0bd9:75c5:0000:0000:6b3e:0170:8394
Classification of IP Address
An IP address is classified into the following types:
1. Public IP Address: This address is available publicly and it is assigned by
your network provider to your router, which further divides it to your
devices. Public IP Addresses are of two types:

 Dynamic IP Address: When you connect a smartphone or computer to the

internet, your Internet Service Provider provides you an IP Address from the
range of available IP Addresses. Now, your device has an IP Address and you
can simply connect your device to the Internet and send and receive data to and
from your device. The very next time when you try to connect to the internet
with the same device, your provider provides you with different IP Addresses to
the same device and also from the same available range. Since IP Address
keeps on changing every time when you connect to the internet, it is called a
Dynamic IP Address.

 Static IP Address: Static address never changes. They serve as a permanent

internet address. These are used by DNS servers. What are DNS servers?
Actually, these are computers that help you to open a website on your
computer. Static IP Address provides information such as device is located on
which continent, which country, which city, and which Internet Service
Provider provides internet connection to that particular device. Once, we know
who is the ISP, we can trace the location of the device connected to the internet.
Static IP Addresses provide less security than Dynamic IP Addresses because
they are easier to track.

2. Private IP Address: This is an internal address of your device which are not
routed to the internet and no exchange of data can take place between a
private address and the internet.

3. Shared IP addresses: Many websites use shared IP addresses where the

traffic is not huge and very much controllable, they decide to rent it to other
similar websites so to make it cost-friendly. Several companies and email
 sending servers use the same IP address (within a single mail server) to cut
down the cost so that they could save for the time the server is idle.

4. Dedicated IP addresses:
A dedicated IP Address is an address used by a single company or an
individual which gives them certain benefits using a private Secure Sockets
Layer (SSL) certificate which is not in the case of a shared IP address. It
allows to access the website or log in via File Transfer Protocol (FTP) by IP
address instead of its domain name. It increases the performance of the
website when the traffic is high. It also protects from a shared IP address that
is black-listed due to spam.

IP SUBNETTING:
A subnetwork or subnet is a logical subdivision of an IP network. The
practice of dividing a network into two or more networks is called
subnetting. Computers that belong to the same subnet are addressed with
an identical group of its most-significant bits of their IP addresses.
When a bigger network is divided into smaller networks, to maintain security,
then that is known as Subnetting. So, maintenance is easier for smaller networks.
For example, if we consider a class A address, the possible number of hosts is
224 for each network, it is obvious that it is difficult to maintain such a huge
number of hosts, but it would be quite easier to maintain if we divide the network
into small parts.

USES OF SUBNETTING

1. Subnetting helps in organizing the network in an efficient way which helps in

expanding the technology for large firms and companies.
2. Subnetting is used for specific staffing structures to reduce traffic and
maintain order and efficiency.
3. Subnetting divides domains of the broadcast so that traffic is routed
efficiently, which helps in improving network performance.
4. Subnetting is used in increasing network security.
The network can be divided into two parts: To divide a network into two parts,
you need to choose one bit for each Subnet from the host ID part.

In the above diagram, there are two Subnets.

Note: It is a class C IP so, there are 24 bits in the network id part and 8 bits in the
host id part.
How Does Subnetting Work?
The working of subnets starts in such a way that firstly it divides the subnets into
smaller subnets. For communicating between subnets, routers are used. Each
subnet allows its linked devices to communicate with each other. Subnetting for a
network should be done in such a way that it does not affect the network bits.

In class C the first 3 octets are network bits so it remains as it is.

 For Subnet-1: The first bit which is chosen from the host id part is zero and
the range will be from (193.1.2.00000000 till you get all 1’s in the host ID part
i.e, 193.1.2.01111111) except for the first bit which is chosen zero for subnet
id part.

Thus, the range of subnet 1 is: 193.1.2.0 to 193.1.2.127

Subnet id of Subnet-1 is : 193.1.2.0
The direct Broadcast id of Subnet-1 is: 193.1.2.127
The total number of hosts possible is: 126 (Out of 128,
2 id's are used for Subnet id & Direct Broadcast id)
The subnet mask of Subnet- 1 is: 255.255.255.128

 For Subnet-2: The first bit chosen from the host id part is one and the range
will be from (193.1.2.100000000 till you get all 1’s in the host ID part i.e,
193.1.2.11111111).
Thus, the range of subnet-2 is: 193.1.2.128 to 193.1.2.255
Subnet id of Subnet-2 is : 193.1.2.128
The direct Broadcast id of Subnet-2 is: 193.1.2.255
The total number of hosts possible is: 126 (Out of 128,
2 id's are used for Subnet id & Direct Broadcast id)
The subnet mask of Subnet- 2 is: 255.255.255.128
The best way to find out the subnet mask of a subnet
is to set the fixed bit of host-id to 1 and the rest to 0.
Finally, after using the subnetting the total number of usable hosts is reduced
from 254 to 252.

Example 1:
An organization is assigned a class C network address of 201.35.2.0. It uses a
netmask of 255.255.255.192 to divide this into sub-networks. Which of the
following is/are valid host IP addresses?

1. 201.35.2.129
2. 201.35.2.191
3. 201.35.2.255
4. Both (A) and (C)

Solution:
Converting the last octet of the
netmask into the binary form: 255.255.255.11000000
Converting the last octet of option 1
into the binary form: 201.35.2.10000001
Converting the last octet of option 2
into the binary form: 201.35.2.10111111
Converting the last octet of option 3
into the binary form: 201.35.2.11111111

From the above, we see that Options 2 and 3 are not valid host IP addresses (as
they are broadcast addresses of a subnetwork), and OPTION 1 is not a broadcast
address and it can be assigned to a host IP.

Example 2:

An organization has a class C network address of 201.32.64.0. It uses a subnet

mask of 255.255.255.248. Which of the following is NOT a valid broadcast
address for any subnetworks?
1. 201.32.64.135
2. 201.32.64.240
3. 201.32.64.207
4. 201.32.64.231

Solution:
Converting the last octet of the netmask
into the binary form: 255.255.255.11111000
Converting the last octet of option 1
into the binary form: 201.32.64.10000111
Converting the last octet of option 2
into the binary form: 201.32.64.11110000
Converting the last octet of option 3
into the binary form: 201.32.64.11001111
Converting the last octet of option 4
into the binary form: 201.32.64.11100111

From the above, we can see that in OPTION 1, 3, and 4, all the host bits are 1 and
give the valid broadcast address of subnetworks and OPTION 2, the last three
bits of the Host address are not 1 therefore it’s not a valid broadcast address.

Advantages of Subnetting
The advantages of Subnetting are mentioned below:
1. It provides security to one network from another network. eg) In an
Organisation, the code of the Developer department must not be accessed by
another department.
2. It may be possible that a particular subnet might need higher network priority
than others. For example, a Sales department needs to host webcasts or video
conferences.
3. In the case of Small networks, maintenance is easy.

Disadvantages of Subnetting
The disadvantages of Subnetting are mentioned below:
1. In the case of a single network, only three steps are required to reach a Process
i.e Source Host to Destination Network, Destination Network to Destination
Host, and then Destination Host to Process.
2. In the case of a Single Network only two IP addresses are wasted to represent
Network Id and Broadcast address but in the case of Subnetting two IP
addresses are wasted for each Subnet.
3. The cost of the overall Network also increases. Subnetting requires internal
routers, Switches, Hubs, Bridges, etc. which are very costly.