Handwritten Notes
Handwritten Notes
Handwritten Notes
COM
Computer Networks
Notes
UNIT - I
NETWORKS
A network is a set of devices (often referred to as nodes) connected by
communication links. A node can be a computer, printer, or any other device
capable of sending and/or receiving data generated by other nodes on the
network.
“Computer network’’ to mean a collection of autonomous computers
interconnected by a single technology. Two computers are said to be
interconnected if they are able to exchange information.
The connection need not be via a copper wire; fiber optics, microwaves,
m
infrared, and communication satellites can also be used.
Networks come in many sizes, shapes and forms, as we will see later.
.co
They are usually connected together to make larger networks, with the
Internet being the most well-known example of a network of networks.
There is considerable confusion in the literature between a computer
network and a distributed system. The key distinction is that in a distributed
ya
system, a collection of independent computers appears to its users as a single
coherent system. Usually, it has a single model or paradigm that it presents to
the users. Often a layer of software on top of the operating system, called
i
middleware, is responsible for implementing this model. A well-known
un
example of a distributed system is the World Wide Web. It runs on top of the
Internet and presents a model in which everything looks like a document (Web
page).
sD
electronic commerce
entertainment.(game playing,)
3 Mobile Users
Text messaging or texting
Smart phones,
GPS (Global Positioning System)
m-commerce
NFC (Near Field Communication)
4 Social Issues
m
With the good comes the bad, as this new-found freedom brings with it many
unsolved social, political, and ethical issues.
.co
Social networks, message boards, content sharing sites, and a host of
other applications allow people to share their views with like-minded
individuals. As long as the subjects are restricted to technical topics or hobbies
like gardening, not too many problems will arise.
ya
The trouble comes with topics that people actually care about, like politics,
religion, or sex. Views that are publicly posted may be deeply offensive to some
people. Worse yet, they may not be politically correct. Furthermore, opinions
i
un
need not be limited to text; high-resolution color photographs and video clips
are easily shared over computer networks. Some people take a live-and-let-live
view, but others feel that posting certain material (e.g., verbal attacks on
particular countries or religions, pornography, etc.) is simply unacceptable and
sD
that such content must be censored. Different countries have different and
conflicting laws in this area. Thus, the debate rages.
Computer networks make it very easy to communicate. They also make it
l
easy for the people who run the network to snoop on the traffic. This sets up
ria
from a home computer outside working hours. Not all employees agree with
this, especially the latter part.
Another conflict is centered around government versus citizen’s rights.
Tu
A new twist with mobile devices is location privacy. As part of the process of
providing service to your mobile device the network operators learn where you
are at different times of day. This allows them to track your movements. They
may know which nightclub you frequent and which medical center you visit.
m
delivered late are useless. In the case of video and audio, timely delivery means
delivering data as they are produced, in the same order that they are produced,
and without significant delay. This kind of delivery is called real-time
.co
transmission.
4. Jitter. Jitter refers to the variation in the packet arrival time. It is the uneven
delay in the delivery of audio or video packets. For example, let us assume that
ya
video packets are sent every 30 ms. If some of the packets arrive with 30-ms
delay and others with 40-ms delay, an uneven quality in the video is the result.
A data communications system has five components
i
I. Message. The message is the information (data) to be communicated.
un
Popular forms of information include text, numbers, pictures, audio, and video.
2 Sender. The sender is the device that sends the data message. It can be a
computer, workstation, telephone handset, video camera, and so on.
sD
3. Receiver. The receiver is the device that receives the message. It can be a
computer, workstation, telephone handset, television, and so on.
4. Transmission medium. The transmission medium is the physical path by
l
Japanese.
Data Representation
Data Flow
Communication between two devices can be simplex, half-duplex, or full-duplex
m
as shown in Figure.
.co
i ya
un
Simplex In simplex mode, the communication is unidirectional, as on a one-
sD
way street. Only one of the two devices on a link can transmit; the other can
only receive (Figure a). Keyboards and traditional monitors are examples of
simplex devices.
Half-Duplex
l
ria
In half-duplex mode, each station can both transmit and receive, but not at the
same time. When one device is sending, the other can only receive, and vice
versa (Figure b). Walkie-talkies and CB (citizens band) radios are both half-
duplex systems.
to
Full-Duplex
In full-duplex, both stations can transmit and receive simultaneously (Figure c).
Tu
m
the network's robustness in a catastrophe.
Security: Network security issues include protecting data from unauthorized
access, protecting data from damage and development, and implementing
.co
policies and procedures for recovery from breaches and data losses.
Physical Structures
Before discussing networks, we need to define some network attributes.
ya
Type of Connection
A network is two or more devices connected through links. A link is a
communications pathway that transfers data from one device to another.
i
There are two possible types of connections: point-to-point and multipoint.
un
Point-to-Point A point-to-point connection provides a dedicated link between
two devices. The entire capacity of the link is reserved for transmission
between those two devices. Most point-to-point connections use an actual
sD
length of wire or cable to connect the two ends, but other options, such as
microwave or satellite links, are also possible
When you change television channels by infrared remote control, you are
l
connection.
m
.co
Physical Topology
The term physical topology refers to the way in which a network is laid out
physically.
ya
Two or more devices connect to a link; two or more links form a topology. The
topology of a network is the geometric representation of the relationship of all
the links and linking devices (usually called nodes) to one another.
i
un
There are four basic topologies possible: mesh, star, bus, and ring
l sD
ria
MESH:
A mesh topology is the one where every node is connected to every other node
in the network.
to
Tu
m
A failure of one device does not cause a break in the network or transmission
of data.
.co
Adding additional devices does not disrupt data transmission between other
devices.
Disadvantages of a mesh topology
The cost to implement is higher than other network topologies, making it a
ya
less desirable option.
Building and maintaining the topology is difficult and time consuming.
The chance of redundant connections is high, which adds to the high costs
i
un
and potential for reduced efficiency.
STAR:
l sD
ria
A star network, star topology is one of the most common network setups. In
this configuration, every node connects to a central network device, like
a hub, switch, or computer. The central network device acts as a server and the
to
peripheral devices act as clients. Depending on the type of network card used
in each computer of the star topology, a coaxial cable or a RJ-45 network cable
is used to connect computers together.
Tu
m
.co
a line topology, a bus topology is a network setup in which each computer
and network device are connected to a single cable or backbone.
Advantages of bus topology
ya
It works well when you have a small network.
It's the easiest network topology for connecting computers or peripherals
in a linear fashion.
i
un
It requires less cable length than a star topology.
Disadvantages of bus topology
It can be difficult to identify the problems if the whole network goes down.
sD
RING:
Tu
m
All data flows in one direction, reducing the chance of packet collisions.
A network server is not needed to control network connectivity between
.co
each workstation.
Data can transfer between workstations at high speeds.
Additional workstations can be added without impacting performance of
the network.
ya
Disadvantages of ring topology
All data being transferred over the network must pass through each
workstation on the network, which can make it slower than a star topology.
i
un
The entire network will be impacted if one workstation shuts down.
The hardware needed to connect each workstation to the network is more
expensive than Ethernet cards and hubs/switches.
sD
Hybrid Topology A network can be hybrid. For example, we can have a main
star topology with each branch connecting several stations in a bus topology as
shown in Figure
l
ria
to
Tu
m
Design to extend over a large area.
Connecting number of LAN's to form larger network, so that resources can be
shared.
.co
Networks can be up to 5 to 50 km.
Owned by organization or individual.
Data transfer rate is low compare to LAN.
ya
Example: Organization with different branches located in the city.
WAN (Wide Area Network)
Are country and worldwide network.
i
Contains multiple LAN's and MAN's.
un
Distinguished in terms of geographical range.
Uses satellites and microwave relays.
Data transfer rate depends upon the ISP provider and varies over the location.
sD
Other types
l
m
.co
ya
Guided Media: Guided media, which are those that provide a medium from
one device to another, include twisted-pair cable, coaxial cable, and fiber-optic
cable.
i
un
Twisted-Pair Cable: A twisted pair consists of two conductors (normally
copper), each with its own plastic insulation, twisted together. One of the wires
is used to carry signals to the receiver, and the other is used only as a ground
sD
reference.
l
ria
to
Tu
The most common UTP connector is RJ45 (RJ stands for registered jack)
Coaxial Cable
Coaxial cable (or coax) carries signals of higher frequency ranges than those in
twisted pair cable. coax has a central core conductor of solid or stranded wire
(usuallycopper) enclosed in an insulating sheath, which is, in turn, encased in
m
an outer conductor of metal foil, braid, or a combination of the two. The outer
metallic wrapping serves both as a shield against noise and as the second
conductor, which completes the circuit.This outer conductor is also enclosed in
.co
an insulating sheath, and the whole cable is protected by a plastic cover.
i ya
un
l sD
(BNe), connector.
Applications
Coaxial cable was widely used in analog telephone networks,digital telephone
networks
to
Fiber-Optic Cable
A fiber-optic cable is made of glass or plastic and transmits signals in the form
of light. Light travels in a straight line as long as it is moving through a single
uniform substance.
If a ray of light traveling through one substance suddenly enters another
substance(of a different density), the ray changes direction.
Bending of light ray
Optical fibers use reflection to guide light through a channel. A glass or plastic
core is surrounded by a cladding of less dense glass or plastic.
m
Propagation Modes
.co
i ya
Multimode is so named because multiple beams from a light source move
un
through the core in different paths. How these beams move within the cable
depends on the structure of the core, as shown in Figure.
l sD
ria
to
Tu
In multimode step-index fiber, the density of the core remains constant from
the center to the edges. A beam of light moves through this constant density in
a straight line until it reaches the interface of the core and the cladding. The
term step index refers to the suddenness of this change, which contributes to
the distortion of the signal as it passes through the fiber.
A second type of fiber, called multimode graded-index fiber, decreases this
distortion of the signal through the cable. The word index here refers to the
index of refraction.
Single-Mode: Single-mode uses step-index fiber and a highly focused source
of light that limits beams to a small range of angles, all close to the horizontal.
m
Applications
Fiber-optic cable is often found in backbone networks because its wide
.co
bandwidth is cost-effective..
Some cable TV companies use a combination of optical fiber and coaxial
cable,thus creating a hybrid network.
Local-area networks such as 100Base-FX network (Fast Ethernet) and
ya
1000Base-X also use fiber-optic cable
Advantages and Disadvantages of Optical Fiber
Advantages Fiber-optic cable has several advantages over metallic cable
(twisted pair or coaxial). i
un
1 Higher bandwidth.
2 Less signal attenuation. Fiber-optic transmission distance is significantly
greaterthan that of other guided media. A signal can run for 50 km without
sD
than copper cables. Copper cables create antenna effects that can easily be
tapped.
Tu
m
Unguided signals can travel from the source to destination in several ways:
.co
ground propagation, sky propagation, and line-of-sight propagation, as shown in
Figure
i ya
un
l sD
Radio Waves
ria
means that the sending and receiving antennas do not have to be aligned. A
sending antenna sends waves that can be received by any receiving antenna.
The omni directional property has a disadvantage, too. The radio waves
Tu
Applications
m
The Omni directional characteristics of radio waves make them useful for
multicasting, in which there is one sender but many receivers. AM and FM radio,
television, maritime radio, cordless phones, and paging are examples of
.co
multicasting.
ya
Microwaves
Electromagnetic waves having frequencies between 1 and 300 GHz are called
i
microwaves. Microwaves are unidirectional. The sending and receiving antennas
un
need to be aligned. The unidirectional property has an obvious advantage. A
pair of antennas can be aligned without interfering with another pair of aligned
antennas
sD
Unidirectional Antenna
Microwaves need unidirectional antennas that send out signals in one direction.
Two types of antennas are used for microwave communications: the parabolic
l
Applications:
Microwaves are used for unicast communication such as cellular telephones,
satellite networks, and wireless LANs
Infrared
Infrared waves, with frequencies from 300 GHz to 400 THz (wavelengths from 1
mm to 770 nm), can be used for short-range communication. Infrared waves,
having high frequencies, cannot penetrate walls. This advantageous
m
area using line-of-sight propagation.
Switching
.co
A network is a set of connected devices. Whenever we have multiple
devices, we have the problem of how to connect them to make one-to-one
communication possible. One solution is to make a point-to-point connection
ya
between each pair of devices (a mesh topology) or between a central device
and every other device (a star topology). These methods, however, are
impractical and wasteful when applied to very large networks.
i
The number and length of the links require too much infrastructure to be
un
cost-efficient, and the majority of those links would be idle most of the time.
A better solution is switching. A switched network consists of a series of
interlinked nodes, called switches. Switches are devices capable of creating
sD
We can then divide today's networks into three broad categories: circuit-
switched networks, packet-switched networks, and message-switched. Packet-
switched networks can further be divided into two subcategories-virtual-circuit
networks and datagram networks as shown in Figure.
m
.co
CIRCUIT-SWITCHED NETWORKS
A circuit-switched network consists of a set of switches connected by
physical links. A connection between two stations is a dedicated path made of
one or more links. However, each connection uses only one dedicated channel
ya
on each link. Each link is normally divided into n channels by using FDM or TDM.
In circuit switching, the resources need to be reserved during the setup
phase;
i
the resources remain dedicated for the entire duration of data transfer until the
un
teardown phase
l sD
ria
to
Three Phases
The actual communication in a circuit-switched network requires three phases:
Tu
Setup Phase
Before the two parties (or multiple parties in a conference call) can
communicate, a dedicated circuit (combination of channels in links) needs to be
established. Connection setup means creating dedicated channels between the
switches. For example, in Figure, when system A needs to connect to system M,
it sends a setup request that includes the address of system M, to switch I.
Switch I finds a channel between itself and switch IV that can be dedicated for
this purpose. Switch I then sends the request to switch IV, which finds a
m
release the resources.
Efficiency
It can be argued that circuit-switched networks are not as efficient as the other
.co
two types of networks because resources are allocated during the entire
duration of the connection. These resources are unavailable to other
connections.
ya
Delay
Although a circuit-switched network normally has low efficiency, the delay in
this type of network is minimal. During data transfer the data are not delayed
i
at each switch; the resources are allocated for the duration of the connection.
un
The total delay is due to the time needed to create the connection, transfer
data, and disconnect the circuit.
Switching at the physical layer in the traditional telephone network
sD
m
Efficiency
.co
The efficiency of a datagram network is better than that of a circuit-switched
network; resources are allocated only when there are packets to be transferred.
Delay
ya
There may be greater delay in a datagram network than in a virtual-circuit
network. Although there are no setup and teardown phases, each packet may
experience a wait at a switch before it is forwarded. In addition, since not all
i
packets in a message necessarily travel through the same switches, the delay is
un
not uniform for the packets of a message.
Switching in the Internet is done by using the datagram approach to
packet switching at the network layer.
l sD
ria
VIRTUAL-CIRCUIT NETWORKS
A virtual-circuit network is a cross between a circuit-switched
to
m
5. A virtual-circuit network is normally implemented in the data link layer, while
a circuit-switched network is implemented in the physical layer and a datagram
network in the network layer.
.co
Addressing
In a virtual-circuit network, two types of addressing are involved: global and
local (virtual-circuit identifier).
ya
Global Addressing
A source or a destination needs to have a global address-an address that can be
unique in the scope of the network.
Virtual-Circuit Identifier
i
un
The identifier that is actually used for data transfer is called the virtual-circuit
identifier (VCI). A VCI, unlike a global address, is a small number that has only
switch scope; it is used by a frame between two switches. When a frame arrives
sD
Three Phases
Tu
m
We show later how the switches make their table entries, but for the
moment we assume that each switch has a table with entries for all active
.co
virtual circuits. Figure shows such a switch and its corresponding table.
Figure shows a frame arriving at port 1 with a VCI of 14. When the frame
arrives, the switch looks in its table to find port 1 and a VCI of 14. When it is
found, the switch knows to change the VCI to 22 and send out the frame from
ya
port 3.
Figure shows how a frame from source A reaches destination B and how its VCI
changes during the trip.
i
un
l sD
ria
the destination. The procedure at the switch is the same for each frame of a
message. The process creates a virtual circuit, not a real circuit, between the
Tu
Setup Phase
In the setup phase, a switch creates an entry for a virtual circuit. For example,
suppose source A needs to create a virtual circuit to B. Two steps are required:
the setup request and the acknowledgment.
Setup Request A setup request frame is sent from the source to the destination.
Figure shows the process.
m
a. Source A sends a setup frame to switch 1.
.co
b. Switch 1 receives the setup request frame. It knows that a frame going from
A to B goes out through port 3. For the moment, assume that it knows the
output port. The switch creates an entry in its table for this virtual circuit, but it
ya
is only able to fill three of the four columns. The switch assigns the incoming
port (1) and chooses an available incoming VCI (14) and the outgoing port (3). It
does not yet know the outgoing VCI, which will be found during the
i
acknowledgment step. The switch then forwards the frame through port 3 to
un
switch 2.
c. Switch 2 receives the setup request frame. The same events happen here as
at switch 1; three columns of the table are completed: in this case, incoming
sD
from A, it assigns a VCI to the incoming frames that come from A, in this case
77. This VCI lets the destination know that the frames come from A, and not
other sources.
Acknowledgment A special frame, called the acknowledgment frame, completes
to
m
in the table, chosen in the previous step. Switch 1 uses this as the outgoing VCI
in the table.
d. Finally switch 1 sends an acknowledgment to source A that contains its
.co
incoming VCI in the table, chosen in the previous step.
e. The source uses this as the outgoing VCI for the data frames to be sent to
destination B.
ya
Teardown Phase
In this phase, source A, after sending all frames to B, sends a special frame
i
called a teardown request. Destination B responds with a teardown
un
confirmation frame. All switches delete the corresponding entry from their
tables.
sD
Efficiency
In virtual-circuit switching, all packets belonging to the same source and
destination travel the same path; but the packets may arrive at the destination
l
Delay
In a virtual-circuit network, there is a one-time delay for setup and a one-time
to
delay for teardown. If resources are allocated during the setup phase, there is
no wait time for individual packets. Figure shows the delay for a packet
Tu
m
Switching at the data link layer in a switched WAN is normally implemented by
using
.co
virtual-circuit techniques.
ya
Comparison
i
un
l sD
ria
to
Tu
m
.co
i ya
un
l sD
OSI
ria
All layers work together in the correct order to move data around a network
m
.co
ya
Physical Layer
Deals with all aspects of physically moving data from one computer to the next
Converts data from the upper layers into 1s and 0s for transmission over media
i
Defines how data is encoded onto the media to transmit the data
un
Defined on this layer: Cable standards, wireless standards, and fiber optic
standards.
Copper wiring, fiber optic cable, radio frequencies, anything that can be used to
sD
Responsible for moving packets (data) from one end of the network to the
other, called end-to-end communications
Requires logical addresses such as IP addresses
Device example: Router
–Routing is the ability of various network devices and their related software to
move data packets from source to destination
Transport Layer
Takes data from higher levels of OSI Model and breaks it into segments that can
be sent to lower-level layers for data transmission
Conversely, reassembles data segments into data that higher-level protocols
m
and applications can use
Also puts segments in correct order (called sequencing ) so they can be
reassembled in correct order at destination
.co
Concerned with the reliability of the transport of sent data
May use a connection-oriented protocol such as TCP to ensure destination
received segments
ya
May use a connectionless protocol such as UDP to send segments without
assurance of delivery
Uses port addressing
i
un
Session Layer
Responsible for managing the dialog between networked devices
Establishes, manages, and terminates connections
Provides duplex, half-duplex, or simplex communications between devices
sD
Application Layer
Contains all services or protocols needed by application software or operating
system to communicate on the network
Examples
o –Firefox web browser uses HTTP (Hyper-Text Transport Protocol)
o –E-mail program may use POP3 (Post Office Protocol version 3) to read e-mails
and SMTP (Simple Mail Transport Protocol) to send e-mails
m
.co
i ya
un
l sD
ria
SUMMARY:
m
–A protocol suite is a large number of related protocols that work together to
allow networked computers to communicate
.co
i ya
un
l sD
Application Layer
Application layer protocols define the rules when implementing specific network
applications
to
Rely on the underlying layers to provide accurate and efficient data delivery
Typical protocols:
o FTP – File Transfer Protocol
Tu
m
the respective computers
Internet Layer
.co
The network layer, also called the internet layer, deals with packets and
connects independent networks to transport the packets across network
boundaries. The network layer protocols are the IP and the Internet Control
ya
Message Protocol (ICMP), which is used for error reporting.
Host-to-network layer
The Host-to-network layer is the lowest layer of the TCP/IP reference model.
i
It combines the link layer and the physical layer of the ISO/OSI model. At
un
this layer, data is transferred between adjacent network nodes in a WAN or
between nodes on the same LAN.
l sD
ria
to
Tu
m
.co
ya
THE INTERNET
i
un
The Internet has revolutionized many aspects of our daily lives. It has affected
the way we do business as well as the way we spend our leisure time. Count
the ways you've used the Internet recently. Perhaps you've sent electronic
sD
mail (e-mail) to a business associate, paid a utility bill, read a newspaper from
a distant city, or looked up a local movie schedule-all by using the Internet. Or
maybe you researched a medical topic, booked a hotel reservation, chatted
l
A Brief History
to
more networks that can communicate with each other. The most notable
internet is called the Internet (uppercase letter I), a collaboration of more than
hundreds of thousands of interconnected networks. Private individuals as well
as various organizations such as government agencies, schools, research
facilities, corporations, and libraries in more than 100 countries use the
Internet. Millions of people are users. Yet this extraordinary communication
system only came into being in 1969.
In the mid-1960s, mainframe computers in research organizations were
standalone devices. Computers from different manufacturers were unable to
communicate with one another. The Advanced Research Projects Agency
m
the University of California at Los Angeles (UCLA), the University of California
at Santa Barbara (UCSB), Stanford Research Institute (SRI), and the University
of Utah, were connected via the IMPs to form a network. Software called the
.co
Network Control Protocol (NCP) provided communication between the hosts.
In 1972, Vint Cerf and Bob Kahn, both of whom were part of the core
ARPANET group, collaborated on what they called the Internetting Projec1.
ya
Cerf and Kahn's landmark 1973 paper outlined the protocols to achieve end-
to-end delivery of packets. This paper on Transmission Control Protocol (TCP)
included concepts such as encapsulation, the datagram, and the functions of
i
a gateway. Shortly thereafter, authorities made a decision to split TCP into two
un
protocols: Transmission Control Protocol (TCP) and Internetworking Protocol
(lP). IP would handle datagram routing while TCP would be responsible for
higher-level functions such as segmentation, reassembly, and error detection.
sD
The Internet has come a long way since the 1960s. The Internet today is not a
l
service providers, regional service providers, and local service providers. The
Internet today is run by private companies, not the government. Figure 1.13
shows a conceptual (not geographic) view of the Internet.
m
.co
International Internet Service Providers:
At the top of the hierarchy are the international service providers that
ya
connect nations together.
National Internet Service Providers:
The national Internet service providers are backbone networks created
i
and maintained by specialized companies. There are many national ISPs
un
operating in North America; some of the most well known are SprintLink,
PSINet, UUNet Technology, AGIS, and internet Mel. To provide connectivity
between the end users, these backbone networks are connected by complex
sD
switching stations (normally run by a third party) called network access points
(NAPs). Some national ISP networks are also connected to one another by
private switching stations called peering points. These normally operate at a
l
Most end users are connected to the local ISPs. Note that in this sense, a local
ISP can be a company that just provides Internet services, a corporation with
a network that supplies services to its own employees, or a nonprofit
organization, such as a college or a university, that runs its own network.
Each of these local ISPs can be connected to a regional or national service
provider.
UNIT- II
DATA LINK LAYER FUNCTIONS (SERVICES)
1. Providing services to the network layer:
1 Unacknowledged connectionless service.
Appropriate for low error rate and real-time traffic. Ex: Ethernet
2. Acknowledged connectionless service.
Useful in unreliable channels, WiFi. Ack/Timer/Resend
3. Acknowledged connection-oriented service.
m
Guarantee frames are received exactly once and in the right order.
Appropriate over long, unreliable links such as a satellite channel or a long-
.co
distance telephone circuit
2. Framing: Frames are the streams of bits received from the network layer
into manageable data units. This division of stream of bits is done by Data
ya
Link Layer.
3. Physical Addressing: The Data Link layer adds a header to the frame in
order to define physical address of the sender or receiver of the frame, if
i
the frames are to be distributed to different systems on the network.
un
4. Flow Control: A receiving node can receive the frames at a faster rate
than it can process the frame. Without flow control, the receiver's buffer
can overflow, and frames can get lost. To overcome this problem, the data
sD
link layer uses the flow control to prevent the sending node on one side of
the link from overwhelming the receiving node on another side of the link.
This prevents traffic jam at the receiver side.
l
FRAMING:
To provide service to the network layer, the data link layer must use the
m
service provided to it by the physical layer. What the physical layer does is
accept a raw bit stream and attempt to deliver it to the destination. This bit
stream is not guaranteed to be error free. The number of bits received may be
.co
less than, equal to, or more than the number of bits transmitted, and they
may have different values. It is up to the data link layer to detect and, if
necessary, correct errors. The usual approach is for the data link layer to
ya
break the bit stream up into discrete frames and compute the checksum for
each frame (framing). When a frame arrives at the destination, the
checksum is recomputed. If the newly computed checksum is different from
i
the one contained in the frame, the data link layer knows that an error has
un
occurred and takes steps to deal with it (e.g., discarding the bad frame and
possibly also sending back an error report).We will look at four framing
methods:
sD
1. Character count.
2. Flag bytes with byte stuffing.
3. Starting and ending flags, with bit stuffing.
l
Character count method uses a field in the header to specify the number of
characters in the frame. When the data link layer at the destination sees the
character count, it knows how many characters follow and hence where the end
of the frame is. This technique is shown in Fig. (a) For four frames of sizes 5, 5,
to
m
anymore.
.co
Flag bytes with byte stuffing method gets around the problem of
resynchronization after an error by having each frame start and end with
special bytes. In the past, the starting and ending bytes were different, but in
recent years most protocols have used the same byte, called a flag byte, as
ya
both the starting and ending delimiter, as shown in Fig. (a) as FLAG. In this
way, if the receiver ever loses synchronization, it can just search for the flag
byte to find the end of the current frame. Two consecutive flag bytes indicate
i
un
the end of one frame and start of the next one.
l sD
ria
to
Tu
(a) A frame delimited by flag bytes (b) Four examples of byte sequences
before and after byte stuffing
It may easily happen that the flag byte's bit pattern occurs in the data.
This situation will usually interfere with the framing. One way to solve this
problem is to have the sender's data link layer insert a special escape byte
(ESC) just before each ''accidental'' flag byte in the data. The data link layer
on the receiving end removes the escape byte before the data are given to
the network layer. This technique is called byte stuffing or character stuffing.
m
characters. For example UNICODE uses 16-bit characters, so a new technique
had to be developed to allow arbitrary sized characters
.co
Starting and ending flags, with bit stuffing allows data frames to
contain an arbitrary number of bits and allows character codes with an arbitrary
number of bits per character. It works like this. Each frame begins and ends
with a special bit pattern, 01111110 (in fact, a flag byte). Whenever the
ya
sender's data link layer encounters five consecutive 1s in the data, it
automatically stuffs a 0 bit into the outgoing bit stream. This bit stuffing is
analogous to byte stuffing, in which an escape byte is stuffed into the outgoing
i
un
character stream before a flag byte in the data.
When the receiver sees five consecutive incoming 1 bits, followed by a 0 bit,
it automatically de- stuffs (i.e., deletes) the 0 bit. Just as byte stuffing is
completely transparent to the network layer in both computers, so is bit
sD
stuffing. If the user data contain the flag pattern, 01111110, this flag is
transmitted as 011111010 but stored in the receiver's memory as 01111110.
l
ria
to
Fig:Bit stuffing. (a) The original data. (b) The data as they appear on the line.
(c) The data as they are stored in the receiver's memory after destuffing .
Tu
With bit stuffing, the boundary between two frames can be unambiguously
recognized by the flag pattern. Thus, if the receiver loses track of where it is,
all it has to do is scan the input for flag sequences, since they can only occur
at frame boundaries and never within the data.
Physical layer coding violations method of framing is only applicable to
networks in which the encoding on the physical medium contains some
redundancy. For example, some LANs encode 1 bit of data by using 2 physical
bits. Normally, a 1 bit is a high-low pair and a 0 bit is a low-high pair. The
scheme means that every data bit has a transition in the middle, making it
easy for the receiver to locate the bit boundaries. The combinations high-
m
As a final note on framing, many data link protocols use combination of a
character count with one of the other methods for extra safety. When a frame
arrives, the count field is used to locate the end of the frame. Only if the
.co
appropriate delimiter is present at that position and the checksum is correct is
the frame accepted as valid. Otherwise, the input stream is scanned for the
next delimiter
i ya
un
l sD
Simplest Protocol
to
Tu
It is very simple. The sender sends a sequence of frames without even thinking
about the receiver. Data are transmitted in one direction only. Both sender &
m
receiver always ready. Processing time can be ignored. Infinite buffer space is
available. And best of all, the communication channel between the data link
.co
layers never damages or loses frames. This thoroughly unrealistic protocol,
which we will nickname ‘‘Utopia,’’ .The utopia protocol is unrealistic because it
does not handle either flow control or error correction
Stop-and-wait Protocol
i ya
un
l sD
ria
It is still very simple. The sender sends one frame and waits for feedback from
the receiver. When the ACK arrives, the sender sends the next frame
It is Stop-and-Wait Protocol because the sender sends one frame, stops until it
to
receives confirmation from the receiver (okay to go ahead), and then sends
the next frame. We still have unidirectional communication for data frames,
but auxiliary ACK frames (simple tokens of acknowledgment) travel from the
Tu
NOISY CHANNELS
Although the Stop-and-Wait Protocol gives us an idea of how to add flow
control to its predecessor, noiseless channels are nonexistent. We can ignore
the error (as we sometimes do), or we need to add error control to our
protocols. We discuss three protocols in this section that use error control.
Sliding Window Protocols:
m
and if it is corrupted, it is silently discarded. The detection of errors in this
protocol is manifested by the silence of the receiver.
.co
Lost frames are more difficult to handle than corrupted ones. In our
previous protocols, there was no way to identify a frame. The received frame
could be the correct one, or a duplicate, or a frame out of order. The solution is
to number the frames. When the receiver receives a data frame that is out of
ya
order, this means that frames were either lost or duplicated
The lost frames need to be resent in this protocol. If the receiver does not
respond when there is an error, how can the sender know which frame to
i
resend? To remedy this problem, the sender keeps a copy of the sent frame. At
un
the same time, it starts a timer. If the timer expires and there is no ACK for the
sent frame, the frame is resent, the copy is held, and the timer is restarted.
Since the protocol uses the stop-and-wait mechanism, there is only one
sD
m
.co
i ya
un
Bandwidth Delay Product:
sD
The link utilization is only 1000/20,000, or 5 percent. For this reason, for a link
with a high bandwidth or long delay, the use of Stop-and-Wait ARQ wastes the
capacity of the link.
m
.co
i ya
un
The sender window at any time divides the possible sequence numbers
into four regions.
sD
The first region, from the far left to the left wall of the window, defines
the sequence numbers belonging to frames that are already acknowledged.
The sender does not worry about these frames and keeps no copies of them.
l
The second region, colored in Figure (a), defines the range of sequence
ria
numbers belonging to the frames that are sent and have an unknown status.
The sender needs to wait to find out if these frames have been received or
were lost. We call these outstanding frames.
The third range, white in the figure, defines the range of sequence
to
numbers for frames that can be sent; however, the corresponding data
packets have not yet been received from the network layer.
Tu
Finally, the fourth region defines sequence numbers that cannot be used
until the window slides
The send window is an abstract concept defining an imaginary box of
size 2m − 1 with three variables: Sf, Sn, and Ssize. The variable Sf defines
the sequence number of the first (oldest) outstanding frame. The variable Sn
holds the sequence number that will be assigned to the next frame to be sent.
Finally, the variable Ssize defines the size of the window.
Figure (b) shows how a send window can slide one or more slots to the
right when an acknowledgment arrives from the other end. The
acknowledgments in this protocol are cumulative, meaning that more than one
frame can be acknowledged by an ACK frame. In Figure, frames 0, I, and 2 are
m
sequence number matching the value of Rn is accepted and acknowledged.
The receive window also slides, but only one slot at a time. When a correct
frame is received (and a frame is received only one at a time), the window
.co
slides.( see below figure for receiving window)
ya
with one single variable Rn. The window slides when a correct frame has
arrived; sliding occurs one slot at a time
i
un
l sD
ria
Although there can be a timer for each frame that is sent, in our protocol we
use only one. The reason is that the timer for the first outstanding frame
always expires first; we send all outstanding frames when this timer expires.
Tu
Acknowledgment
The receiver sends a positive acknowledgment if a frame has arrived safe and
sound and in order. If a frame is damaged or is received out of order, the
receiver is silent and will discard all subsequent frames until it receives the
one it is expecting. The silence of the receiver causes the timer of the
unacknowledged frame at the sender side to expire. This, in turn, causes the
sender to go back and resend all frames, beginning with the one with the
expired timer. The receiver does not have to acknowledge each frame
received. It can send one cumulative acknowledgment for several frames.
Below figure is an example(if ack lost) of a case where the forward channel is
reliable, but the reverse is not. No data frames are lost, but some ACKs are
delayed and one is lost. The example also shows how cumulative
m
acknowledgments can help if acknowledgments are delayed or lost
.co
i ya
un
l sD
ria
m
.co
i ya
un
Stop-and-Wait ARQ is a special case of Go-Back-N ARQ in which the size of the
send window is 1.
In Go-Back-N ARQ, The receiver keeps track of only one variable, and there is
no need to buffer out-of- order frames; they are simply discarded. However,
this protocol is very inefficient for a noisy link.
l
In a noisy link a frame has a higher probability of damage, which means the
ria
resending of multiple frames. This resending uses up the bandwidth and slows
down the transmission.
For noisy links, there is another mechanism that does not resend N frames
to
when just one frame is damaged; only the damaged frame is resent. This
mechanism is called Selective Repeat ARQ.
It is more efficient for noisy links, but the processing at the receiver is more
Tu
complex.
Sender Window (explain go-back N sender window concept (before & after
sliding.) The only difference in sender window between Go-back N and Selective
Repeat is Window size)
Receiver window
m
The receiver window in Selective Repeat is totally different from the one in Go
Back-N. First, the size of the receive window is the same as the size of the
.co
send window (2m-1).
The Selective Repeat Protocol allows as many frames as the size of the
receiver window to arrive out of order and be kept until there is a set of in-
ya
order frames to be delivered to the network layer. Because the sizes of the
send window and receive window are the same, all the frames in the send
frame can arrive out of order and be stored until they can be delivered.
i
However the receiver never delivers packets out of order to the network layer.
un
Above Figure shows the receive window. Those slots inside the window that are
colored define frames that have arrived out of order and are waiting for their
neighbors to arrive before delivery to the network layer.
sD
In Selective Repeat ARQ, the size of the sender and receiver window must be at
most one-half of 2m
l
Flow Diagram
Tu
m
.co
Differences between Go-Back N & Selective Repeat
ya
One main difference is the number of timers. Here, each frame sent or
resent needs a timer, which means that the timers need to be numbered (0,
1,2, and 3). The timer for frame 0 starts at the first request, but stops when
i
un
the ACK for this frame arrives.
There are two conditions for the delivery of frames to the network layer:
First, a set of consecutive frames must have arrived. Second, the set starts
from the beginning of the window. After the first arrival, there was only one
sD
frame and it started from the beginning of the window. After the last arrival,
there are three frames and the first one starts from the beginning of the
window.
l
The next point is about the ACKs. Notice that only two ACKs are sent here.
The first one acknowledges only the first frame; the second one acknowledges
three frames. In Selective Repeat, ACKs are sent when data are delivered to
to
the network layer. If the data belonging to n frames are delivered in one shot,
only one ACK is sent for all of them.
Tu
Piggybacking
A technique called piggybacking is used to improve the efficiency of the
bidirectional protocols. When a frame is carrying data from A to B, it can also
carry control information about arrived (or lost) frames from B; when a frame
is carrying data from B to A, it can also carry control information about the
arrived (or lost) frames from A.
RANDOM ACCESS PROTOCOLS
We can consider the data link layer as two sub layers. The upper sub layer is
responsible for data link control, and the lower sub layer is responsible for
resolving access to the shared media
The upper sub layer that is responsible for flow and error control is called the
logical link control (LLC) layer; the lower sub layer that is mostly responsible
for multiple access resolution is called the media access control (MAC) layer.
When nodes or stations are connected and use a common link, called a
m
multipoint or broadcast link, we need a multiple-access protocol to coordinate
access to the link.
.co
i ya
un
Taxonomy of multiple-access protocols
RANDOM ACCESS
In random access or contention methods, no station is superior to another
sD
why these methods are called random access. Second, no rules specify which
ria
station should send next. Stations compete with one another to access the
medium. That is why these methods are also called contention methods.
ALOHA
1 Pure ALOHA
to
The original ALOHA protocol is called pure ALOHA. This is a simple, but elegant
protocol. The idea is that each station sends a frame whenever it has a frame
Tu
to send. However, since there is only one channel to share, there is the
possibility of collision between frames from different stations. Below Figure
shows an example of frame collisions in pure ALOHA.
m
In pure ALOHA, the stations transmit frames whenever they have data to send.
When two or more stations transmit simultaneously, there is collision and the
frames are destroyed.
.co
In pure ALOHA, whenever any station transmits a frame, it expects the
acknowledgement from the receiver.
If acknowledgement is not received within specified time, the station assumes
ya
that the frame (or acknowledgement) has been destroyed.
If the frame is destroyed because of collision the station waits for a random
amount of time and sends it again. This waiting time must be random
i
otherwise same frames will collide again and again.
un
Therefore pure ALOHA dictates that when time-out period passes, each station
must wait for a random amount of time before resending its frame. This
randomness will help avoid more collisions.
sD
Vulnerable time Let us find the length of time, the vulnerable time, in
which there is a possibility of collision. We assume that the stations send fixed-
length frames with each frame taking Tfr S to send. Below Figure shows the
l
Station A sends a frame at time t. Now imagine station B has already sent a
frame between t - Tfr and t. This leads to a collision between the frames from
station A and station B. The end of B's frame collides with the beginning of A's
frame. On the other hand, suppose that station C sends a frame between t and
t + Tfr . Here, there is a collision between frames from station A and station C.
The beginning of C's frame collides with the end of A's frame
m
.co
i ya
un
sD
Example
ria
Average frame transmission time Tfr is 200 bits/200 kbps or 1 ms. The
vulnerable time is 2 x 1 ms =2 ms. This means no station should send later
than 1 ms before this station starts transmission and no station should start
Tu
sending during the one I-ms period that this station is sending.
The throughput for pure ALOHA is S = G × e −2G . The maximum
throughput Smax = 0.184 when G= (1/2).
PROBLEM
A pure ALOHA network transmits 200-bit frames on a shared channel of 200
kbps. What is the throughput if the system (all stations together) produces a.
1000 frames per second b. 500 frames per second c. 250 frames per second .
The frame transmission time is 200/200 kbps or 1 ms.
a. If the system creates 1000 frames per second, this is 1 frame per
m
percent). This means that the throughput is 250 × 0.152 = 38. Only 38 frames
out of 250 will probably survive.
.co
2 Slotted ALOHA
ya
rule that defines when the station can send. A station may send soon after
another station has started or soon before another station has finished. Slotted
ALOHA was invented to improve the efficiency of pure ALOHA.
i
un
In slotted ALOHA we divide the time into slots of Tfr s and force the station to
send only at the beginning of the time slot. Figure 3 shows an example of
frame collisions in slotted ALOHA
l sD
ria
to
FIG:3
Because a station is allowed to send only at the beginning of the synchronized
Tu
time slot, if a station misses this moment, it must wait until the beginning of
the next time slot. This means that the station which started at the beginning
of this slot has already finished sending its frame. Of course, there is still the
possibility of collision if two stations try to send at the beginning of the same
time slot. However, the vulnerable time is now reduced to one-half, equal to
Tfr Figure 4 shows the situation
Below fig shows that the vulnerable time for slotted ALOHA is one-half that of
pure ALOHA. Slotted ALOHA vulnerable time = Tfr
m
A slotted ALOHA network transmits 200-bit frames using a shared channel with
.co
a 200- Kbps bandwidth. Find the throughput if the system (all stations together)
produces
a. 1000 frames per second b. 500 frames per second c. 250
ya
frames per second
Solution
This situation is similar to the previous exercise except that the network is
i
using slotted ALOHA instead of pure ALOHA. The frame transmission time is
un
200/200 kbps or 1 ms.
a. In this case G is 1. So S =G x e-G or S =0.368 (36.8 percent). This means
that the throughput is 1000 x 0.0368 =368 frames. Only 368 out of 1000
sD
frames will probably survive. Note that this is the maximum throughput case,
percentagewise.
b. Here G is 1/2 In this case S =G x e-G or S =0.303 (30.3 percent). This
l
means that the throughput is 500 x 0.0303 =151. Only 151 frames out of
ria
m
.co
ya
Carrier Sense Multiple Access (CSMA)
To minimize the chance of collision and, therefore, increase the
performance, the CSMA method was developed. The chance of collision can be
i
reduced if a station senses the medium before trying to use it. Carrier sense
un
multiple access (CSMA) requires that each station first listen to the medium (or
check the state of the medium) before sending. In other words, CSMA is based
on the principle "sense before transmit" or "listen before talk."
sD
CSMA can reduce the possibility of collision, but it cannot eliminate it. The
reason for this is shown in below Figure. Stations are connected to a shared
channel (usually a dedicated medium).
l
The possibility of collision still exists because of propagation delay; station may
ria
sense the medium and find it idle, only because the first bit sent by another
station has not yet been received.
At time tI' station B senses the medium and finds it idle, so it sends a
frame. At time t2 (t2> tI)' station C senses the medium and finds it idle
to
m
Space/time model of the collision in CSMA
Vulnerable Time
.co
The vulnerable time for CSMA is the propagation time Tp . This is the time
needed for a signal to propagate from one end of the medium to the other.
When a station sends a frame, and any other station tries to send a frame
ya
during this time, a collision will result. But if the first bit of the frame reaches
the end of the medium, every station will already have heard the bit and will
refrain from sending
i
un
l sD
ria
Persistence Methods
to
What should a station do if the channel is busy? What should a station do if the
channel is idle? Three methods have been devised to answer these questions:
Tu
m
.co
ya
1-Persistent: In this method, after the station finds the line idle, it sends its
frame immediately (with probability 1). This method has the highest chance of
i
collision because two or more stations may find the line idle and send their
un
frames immediately.
Non-persistent: a station that has a frame to send senses the line. If the line
is idle, it sends immediately. If the line is not idle, it waits a random amount of
sD
time and then senses the line again. This approach reduces the chance of
collision because it is unlikely that two or more stations will wait the same
amount of time and retry to send simultaneously. However, this method
reduces the efficiency of the network because the medium remains idle when
l
p-Persistent: This is used if the channel has time slots with a slot duration
equal to or greater than the maximum propagation time. The p-persistent
approach combines the advantages of the other two strategies. It reduces the
to
2. With probability q = 1 - p, the station waits for the beginning of the next
time slot and checks the line again.
a. If the line is idle, it goes to step 1.
b. If the line is busy, it acts as though a collision has occurred and uses the
backoff procedure.
m
.co
a.
Carrier Sense Multiple Access with Collision Detection (CSMA/CD)
The CSMA method does not specify the procedure following a collision.
ya
Carrier sense multiple access with collision detection (CSMA/CD) augments the
algorithm to handle the collision.
In this method, a station monitors the medium after it sends a frame to
i
see if the transmission was successful. If so, the station is finished. If, however,
un
there is a collision, the frame is sent again.
To better understand CSMA/CD, let us look at the first bits transmitted by
the two stations involved in the collision. Although each station continues to
sD
send bits in the frame until it detects the collision, we show what happens as
the first bits collide. In below Figure, stations A and C are involved in the
collision.
l
ria
to
Tu
m
reach the second, and the effect of the collision takes another time Tp to reach
the first. So the requirement is that the first station must still be transmitting
.co
after 2Tp .
i ya
un
l sD
ria
to
Tu
m
.co
i ya
un
Flow diagram for the CSMA/CD
PROBLEM
sD
SOL
The frame transmission time is Tfr = 2 × Tp = 51.2 μs. This means, in the
worst case, a station needs to transmit for a period of 51.2 μs to detect the
to
collision. The minimum size of the frame is 10 Mbps × 51.2 μs = 512 bits or 64
bytes. This is actually the minimum size of the frame for Standard Ethernet.
Tu
m
.co
acknowledgments, as shown in Figure
Timing in CSMA/CA
ya
Inter frame Space (IFS)
First, collisions are avoided by deferring transmission even if the channel
is found idle. When an idle channel is found, the station does not send
i
immediately. It waits for a period of time called the inter frame space or IFS.
un
Even though the channel may appear idle when it is sensed, a distant
station may have already started transmitting. The distant station's signal has
not yet reached this station. The IFS time allows the front of the transmitted
sD
signal by the distant station to reach this station. If after the IFS time the
channel is still idle, the station can send, but it still needs to wait a time equal
to the contention time. The IFS variable can also be used to prioritize stations
l
or frame types. For example, a station that is assigned shorter IFS has a higher
ria
priority.
In CSMA/CA, the IFS can also be used to define the priority of a station or a
frame.
to
Contention Window
The contention window is an amount of time divided into slots. A station
Tu
that is ready to send chooses a random number of slots as its wait time. The
number of slots in the window changes according to the binary exponential
back-off strategy. This means that it is set to one slot the first time and then
doubles each time the station cannot detect an idle channel after the IFS time.
This is very similar to the p-persistent method except that a random outcome
defines the number of slots taken by the waiting station.
One interesting point about the contention window is that the station
needs to sense the channel after each time slot. However, if the station finds
the channel busy, it does not restart the process; it just stops the timer and
restarts it when the channel is sensed as idle. This gives priority to the station
with the longest waiting time.
m
.co
i ya
un
l sD
ria
to
strategies.
As soon as it finds the line to be idle, the station waits for an IFS (Inter frame
space) amount of time.
If then waits for some random time and sends the frame.
After sending the frame, it sets a timer and waits for the acknowledgement
from the receiver.
If the acknowledgement is received before expiry of the timer, then the
transmission is successful.
But if the transmitting station does not receive the expected
acknowledgement before the timer expiry then it increments the back off
In controlled access, the stations seek information from one another to find
which station has the right to send. It allows only one node to send at a time, to
avoid collision of messages on shared medium.
The three controlled-access methods are:
m
.co
Reservation
ya
sending data.
The time line has two kinds of periods:
1. Reservation interval of fixed time length
i
2. Data transmission period of variable frames.
un
If there are M stations, the reservation interval is divided into M slots, and
each station has one slot.
Suppose if station 1 has a frame to send, it transmits 1 bit during the slot
sD
The stations which have reserved their slots transfer their frames in that
order.
After data transmission period, next reservation interval begins.
to
Since everyone agrees on who goes next, there will never be any
collisions.
Tu
The following figure shows a situation with five stations and a five slot
reservation frame. In the first interval, only stations 1, 3, and 4 have made
reservations. In the second interval, only station 1 has made a reservation.
Polling
m
Polling process is similar to the roll-call performed in class. Just like the
teacher, a controller sends a message to each node in turn.
.co
In this, one acts as a primary station(controller) and the others are
secondary stations. All data exchanges must be made through the controller.
The message sent by the controller contains the address of the node
being selected for granting access.
ya
Although all nodes receive the message but the addressed one responds
to it and sends data, if any. If there is no data, usually a “poll reject”(NAK)
message is sent back.
i
un
Problems include high overhead of the polling messages and high
dependence on the reliability of the controller.
l sD
ria
to
Tu
Token Passing
In token passing scheme, the stations are connected logically to each
other in form of ring and access of stations is governed by tokens.
A token is a special bit pattern or a small message, which circulate from
one station to the next in the some predefined order.
In Token ring, token is passed from one station to another adjacent station
in the ring whereas incase of Token bus, each station
uses the bus to send the token to the next station in some predefined order.
In both cases, token represents permission to send. If a station has a
frame queued for transmission when it receives the token, it can send that
frame before it passes the token to the next station. If it has no queued frame,
it passes the token simply.
After sending a frame, each station must wait for all N stations (including
itself) to send the token to their neighbors and the other N – 1 stations to send
a frame, if they have one.
m
There exists problems like duplication of token or token is lost or insertion
of new station, removal of a station, which need be tackled for correct and
.co
reliable operation of this scheme.
i ya
un
l sD
ria
Error Detection
to
Error
A condition when the receiver’s information does not matches with the sender’s
Tu
information. During transmission, digital signals suffer from noise that can
introduce errors in the binary bits travelling from sender to receiver. That
means a 0 bit may change to 1 or a 1 bit may change to 0.
Error Detecting Codes (Implemented either at Data link layer or
Transport Layer of OSI Model)
Whenever a message is transmitted, it may get scrambled by noise or data
may get corrupted. To avoid this, we use error-detecting codes which are
additional data added to a given digital message to help us detect if any error
has occurred during transmission of the message.
Basic approach used for error detection is the use of redundancy bits, where
3. Checksum
m
Simple Parity check
Blocks of data from the source are subjected to a check bit or parity bit
generator form, where a parity of : 1 is added to the block if it contains odd
.co
number of 1’s, and
0 is added if it contains even number of 1’s
This scheme makes the total number of 1’s even, that is why it is called even
ya
parity checking.
i
un
l sD
ria
parity check bit. Parity check bits are also calculated for all columns, then both
are sent along with the data. At the receiving end these are compared with the
Tu
m
.co
Checksum
ya
In checksum error detection scheme, the data is divided into k segments
each of m bits.
In the sender’s end the segments are added using 1’s complement
i
arithmetic to get the sum. The sum is complemented to get the checksum.
un
The checksum segment is sent along with the data segments.
At the receiver’s end, all received segments are added using 1’s complement
arithmetic to get the sum. The sum is complemented.
sD
m
.co
i ya
un
Unlike checksum scheme, which is based on addition, CRC is based on binary
sD
division.
In CRC, a sequence of redundant bits, called cyclic redundancy check bits,
are appended to the end of data unit so that the resulting data unit becomes
l
At the destination, the incoming data unit is divided by the same number. If
at this step there is no remainder, the data unit is assumed to be correct and
is therefore accepted.
A remainder indicates that the data unit has been damaged in transit and
to
Error Correction
Error Correction codes are used to detect and correct the errors when data is
transmitted from the sender to the receiver.
Suppose r is the number of redundant bits and d is the total number of the data
bits. The number of redundant bits r can be calculated by using the formula:
r
2 >=d+r+1
The value of r is calculated by using the above formula. For example, if the
value of d is 4, then the possible smallest value that satisfies the above relation
m
would be 3.
.co
R.W Hamming is Hamming code which can be applied to any length of the data
unit and uses the relationship between data units and redundant units.
Hamming Code
ya
Parity bits: The bit which is appended to the original data of binary bits so that
the total number of 1s is even or odd.
Even parity: To check for even parity, if the total number of 1s is even, then the
i
value of the parity bit is 0. If the total number of 1s occurrences is odd, then the
un
value of the parity bit is 1.
Odd Parity: To check for odd parity, if the total number of 1s is even, then the
value of parity bit is 1. If the total number of 1s is odd, then the value of parity
sD
bit is 0.
Algorithm of Hamming code:
An information of 'd' bits are added to the redundant bits 'r' to form d+r.
The location of each of the (d+r) digits is assigned a decimal value.
l
At the receiving end, the parity bits are recalculated. The decimal value of the
parity bits determines the position of an error.
Relationship b/w Error position & binary number.
to
Tu
m
Representation of Data on the addition of parity bits:
.co
Determining the Parity bits
ya
Determining the r1 bit: The r1 bit is calculated by performing a parity check on
the bit positions whose binary representation includes 1 in the first position.
i
un
sD
We observe from the above figure that the bit position that includes 1 in the
l
first position are 1, 3, 5, 7. Now, we perform the even-parity check at these bit
ria
We observe from the above figure that the bit positions that includes 1 in the
second position are 2, 3, 6, 7. Now, we perform the even-parity check at these
m
.co
We observe from the above figure that the bit positions that includes 1 in the
third position are 4, 5, 6, 7. Now, we perform the even-parity check at these bit
positions. The total number of 1 at these bit positions corresponding to r4 is
ya
even, therefore, the value of the r4 bit is 0.
Suppose the 4th bit is changed from 0 to 1 at the receiving end, then parity bits
are recalculated.
l
R1 bit
ria
We observe from the above figure that the binary representation of r1 is 1100.
Now, we perform the even-parity check, the total number of 1s appearing in the
r1 bit is an even number. Therefore, the value of r1 is 0.
R2 bit
The bit positions of r2 bit are 2,3,6,7.
We observe from the above figure that the binary representation of r2 is 1001.
Now, we perform the even-parity check, the total number of 1s appearing in the
m
r2 bit is an even number. Therefore, the value of r2 is 0.
R4 bit
.co
The bit positions of r4 bit are 4,5,6,7.
i ya
We observe from the above figure that the binary representation of r4 is 1011.
un
Now, we perform the even-parity check, the total number of 1s appearing in the
r4 bit is an odd number. Therefore, the value of r4 is 1.
sD
The binary representation of redundant bits, i.e., r4r2r1 is 100, and its
corresponding decimal value is 4. Therefore, the error occurs in a 4th bit
position. The bit value must be changed from 1 to 0 to correct the error.
l
In 1985, the Computer Society of the IEEE started a project, called Project
802, to set standards to enable intercommunication among equipment from a
variety of manufacturers. Project 802 is a way of specifying functions of the
to
physical layer and the data link layer of major LAN protocols.
The relationship of the 802 Standard to the traditional OSI model is shown in
Tu
below Figure. The IEEE has subdivided the data link layer into two sub layers:
logical link control (LLC) and media access control).
IEEE has also created several physical layer standards for different LAN
protocols
m
STANDARD ETHERNET
.co
The original Ethernet was created in 1976 at Xerox’s Palo Alto Research
Center (PARC). Since then, it has gone through four generations.
Standard Ethernet (l0 Mbps), Fast Ethernet (100 Mbps), Gigabit Ethernet (l
ya
Gbps), and Ten-Gigabit Ethernet (l0 Gbps),
We briefly discuss the Standard (or traditional) Ethernet in this section
i
un
sD
MAC Sublayer
l
In Standard Ethernet, the MAC sublayer governs the operation of the access
ria
method. It also frames data received from the upper layer and passes them to
the physical layer.
Frame Format
The Ethernet frame contains seven fields: preamble, SFD, DA, SA, length or
to
type of protocol data unit (PDU), upper-layer data, and the CRC. Ethernet does
not provide any mechanism for acknowledging received frames, making it
Tu
Preamble. The first field of the 802.3 frame contains 7 bytes (56 bits) of
alternating 0s and 1s that alerts the receiving system to the coming frame and
enables it to synchronize its input timing. The pattern provides only an alert
and a timing pulse. The 56-bit pattern allows the stations to miss some bits at
the beginning of the frame. The preamble is actually added at the physical
layer and is not (formally) part of the frame.
Start frame delimiter (SFD). The second field (l byte: 10101011) signals the
beginning of the frame. The SFD warns the station or stations that this is the
m
last chance for synchronization. The last 2 bits is 11 and alerts the receiver
that the next field is the destination address.
Destination address (DA). The DA field is 6 bytes and contains the physical
.co
address of the destination station or stations to receive the packet.
Source address (SA). The SA field is also 6 bytes and contains the physical
address of the sender of the packet.
ya
Length or type. This field is defined as a type field or length field. The original
Ethernet used this field as the type field to define the upper-layer protocol
using the MAC frame. The IEEE standard used it as the length field to define
i
the number of bytes in the data field. Both uses are common today.
un
Data. This field carries data encapsulated from the upper-layer protocols. It is a
minimum of 46 and a maximum of 1500 bytes.
CRC. The last field contains error detection information, in this case a CRC-32
sD
Frame Length
Ethernet has imposed restrictions on both the minimum and maximum lengths
of a frame, as shown in below Figure
l
ria
to
Tu
Addressing
The Ethernet address is 6 bytes (48 bits), normally written in hexadecimal
m
notation, with a colon between the bytes.
Example of an Ethernet address in hexadecimal notation
.co
Unicast, Multicast, and Broadcast Addresses A source address is always a
ya
unicast address-the frame comes from only one station. The destination
address, however, can be unicast, multicast, or broadcast. Below Figure
shows how to distinguish a unicast address from a multicast address.
i
un
If the least significant bit of the first byte in a destination address is 0, the
address is unicast; otherwise, it is multicast.
l sD
ria
The slot time in Ethernet is defined in bits. It is the time required for a station
Slot Time and Maximum Network Length There is a relationship between the
slot time and the maximum length of the network (collision domain). It is
dependent on the propagation speed of the signal in the particular medium.
In most transmission media, the signal propagates at 2 x 10 8 m/s (two-thirds
of the rate for propagation in air).
For traditional Ethernet, we calculate
MaxLength =PropagationSpeedx (SlotTime/2)
m
MaxLength= (2 x 108) X(51.2 X10-6 )/2= 5120m
.co
Of course, we need to consider the delay times in repeaters and interfaces,
and the time required to send the jam sequence. These reduce the maximum-
length of a traditional Ethernet network to 2500 m, just 48 percent of the
theoretical calculation. MaxLength=2500 m
ya
Physical Layer
i
The Standard Ethernet defines several physical layer implementations; four of
un
the most common, are shown in Figure
l sD
ria
to
the sender, data are converted to a digital signal using the Manchester scheme;
at the receiver, the received signal is interpreted as Manchester and decoded
into data. Manchester encoding is self-synchronous, providing a transition at
each bit interval. Figure shows the encoding scheme for Standard Ethernet
m
The first implementation is called 10Base5, thick Ethernet, or Thicknet.
lOBase5 was the first Ethernet specification to use a bus topology with an
.co
external transceiver (transmitter/receiver) connected via a tap to a thick
coaxial cable. Figure shows a schematic diagram of a lOBase5 implementation
i ya
un
10Base5 implementation
sD
10Base2 also uses a bus topology, but the cable is much thinner and more
ria
10Base2 implementation
m
10Base-T implementation
.co
Although there are several types of optical fiber 10-Mbps Ethernet, the most
common is called 10Base-F.10Base-F uses a star topology to connect stations
to a hub. The stations are connected to the hub using two fiber-optic cables, as
ya
shown in Figure
i
un
sD
10Base-F implementation
l
ria
to
Tu
UNIT-III
Network Layer Design Issues
1. Store-and-forward packet switching
2. Services provided to transport layer
3. Implementation of connectionless service
4. Implementation of connection-oriented service
5. Comparison of virtual-circuit and datagram networks
m
1 Store-and-forward packet switching
.co
i ya
un
A host with a packet to send transmits it to the nearest router, either on its own LAN or over a
sD
point-to-point link to the ISP. The packet is stored there until it has fully arrived and the link
has finished its processing by verifying the checksum. Then it is forwarded to the next router
along the path until it reaches the destination host, where it is delivered. This mechanism is
l
interface. The services need to be carefully designed with the following goals in mind:
1. Services independent of router technology.
Tu
If connectionless service is offered, packets are injected into the network individually and
routed independently of each other. No advance setup is needed. In this context, the packets
are frequently called datagrams (in analogy with telegrams) and the network is called a
datagram network.
m
.co
A’s table (initially) A’s table (later) C’s Table E’s Table
i ya
un
Let us assume for this example that the message is four times longer than the maximum
sD
packet size, so the network layer has to break it into four packets, 1, 2, 3, and 4, and send each
of them in turn to router A.
Every router has an internal table telling it where to send packets for each of the possible
l
ria
destinations. Each table entry is a pair(destination and the outgoing line). Only directly
connected lines can be used.
A’s initial routing table is shown in the figure under the label ‘‘initially.’’
At A, packets 1, 2, and 3 are stored briefly, having arrived on the incoming link. Then each
to
packet is forwarded according to A’s table, onto the outgoing link to C within a new frame.
Packet 1 is then forwarded to E and then to F.
Tu
m
.co
A’s table C’s Table E’s Table
i ya
If connection-oriented service is used, a path from the source router all the way to the
un
destination router must be established before any data packets can be sent. This connection is
called a VC (virtual circuit), and the network is called a virtual-circuit network
sD
When a connection is established, a route from the source machine to the destination
machine is chosen as part of the connection setup and stored in tables inside the routers. That
l
route is used for all traffic flowing over the connection, exactly the same way that the
ria
telephone system works. When the connection is released, the virtual circuit is also
terminated. With connection-oriented service, each packet carries an identifier telling which
virtual circuit it belongs to.
to
As an example, consider the situation shown in Figure. Here, host H1 has established
Tu
connection 1 with host H2. This connection is remembered as the first entry in each of the
routing tables. The first line of A’s table says that if a packet bearing connection identifier 1
comes in from H1, it is to be sent to router C and given connection identifier 1. Similarly, the
first entry at C routes the packet to E, also with connection identifier 1.
Now let us consider what happens if H3 also wants to establish a connection to H2. It chooses
connection identifier 1 (because it is initiating the connection and this is its only connection)
and tells the network to establish the virtual circuit.
This leads to the second row in the tables. Note that we have a conflict here because although
A can easily distinguish connection 1 packets from H1 from connection 1 packets from H3, C
cannot do this. For this reason, A assigns a different connection identifier to the outgoing
traffic for the second connection. Avoiding conflicts of this kind is why routers need the ability
to replace connection identifiers in outgoing packets.
In some contexts, this process is called label switching. An example of a connection-oriented
network service is MPLS (Multi Protocol Label Switching).
m
5 Comparison of virtual-circuit and datagram networks
.co
i ya
un
l sD
ria
Routing Algorithms
to
The main function of NL (Network Layer) is routing packets from the source machine to the
destination machine.
There are two processes inside router:
Tu
a) One of them handles each packet as it arrives, looking up the outgoing line to use for it in
the routing table. This process is forwarding.
b) The other process is responsible for filling in and updating the routing tables. That is where
the routing algorithm comes into play. This process is routing.
Regardless of whether routes are chosen independently for each packet or only when new
connections are established, certain properties are desirable in a routing algorithm
correctness, simplicity, robustness, stability, fairness, optimality
m
Adaptive algorithm, in contrast, change their routing decisions to reflect changes in the
.co
topology, and usually the traffic as well.
Adaptive algorithms differ in
1) Where they get their information (e.g., locally, from adjacent routers, or from all routers),
ya
2) When they change the routes (e.g., every ∆T sec, when the load changes or when the
topology changes), and
3) What metric is used for optimization (e.g., distance, number of hops, or estimated transit
i
un
time).
This procedure is called dynamic routing
sD
• Flooding
ria
One can make a general statement about optimal routes without regard to network topology
or traffic. This statement is known as the optimality principle.
It states that if router J is on the optimal path from router I to router K, then the optimal path
from J to K also falls along the same
As a direct consequence of the optimality principle, we can see that the set of optimal routes
from all sources to a given destination form a tree rooted at the destination. Such a tree is
called a sink tree. The goal of all routing algorithms is to discover and use the sink trees for all
routers
m
Shortest Path Routing (Dijkstra’s)
.co
The idea is to build a graph of the subnet, with each node of the graph representing a router
and each arc of the graph representing a communication line or link.
To choose a route between a given pair of routers, the algorithm just finds the shortest path
ya
between them on the graph
1. Start with the local node (router) as the root of the tree. Assign a cost of 0 to this node and
make it the first permanent node.
i
un
2. Examine each neighbor of the node that was the last permanent node.
3. Assign a cumulative cost to each node and make it tentative
4. Among the list of tentative nodes
sD
a. Find the node with the smallest cost and make it Permanent
b. If a node can be reached from more than one route then select the route with the
shortest cumulative cost.
l
ria
m
.co
i ya
un
Flooding
• Another static algorithm is flooding, in which every incoming packet is sent out on every
sD
• One such measure is to have a hop counter contained in the header of each packet, which
ria
is decremented at each hop, with the packet being discarded when the counter reaches
zero. Ideally, the hop counter should be initialized to the length of the path from source to
destination.
to
• A variation of flooding that is slightly more practical is selective flooding. In this algorithm
the routers do not send every incoming packet out on every line, only on those lines that
Tu
Routing between autonomous systems is referred to as inter domain routing. (PATH VECTOR)
Each autonomous system can choose one or more intra domain routing protocols to handle
routing inside the autonomous system. However, only one inter domain routing protocol
handles routing between autonomous systems.
m
.co
Distance Vector Routing i ya
un
In distance vector routing, the least-cost route between any two nodes is the route with
minimum distance. In this protocol, as the name implies, each node maintains a vector (table)
sD
Initialization
Sharing
Updating
to
Initialization
Each node can know only the distance between itself and its immediate neighbors, those
Tu
directly connected to it. So for the moment, we assume that each node can send a message to
the immediate neighbors and find the distance between itself and these neighbors. Below fig
shows the initial tables for each node. The distance for any entry that is not a neighbor is
marked as infinite (unreachable).
m
.co
Sharing
ya
The whole idea of distance vector routing is the sharing of information between neighbors.
Although node A does not know about node E, node C does. So if node C shares its routing
table with A, node A can also know how to reach node E. On the other hand, node C does not
i
know how to reach node D, but node A does. If node A shares its routing table with node C,
un
node C also knows how to reach node D. In other words, nodes A and C, as immediate
neighbors, can improve their routing tables if they help each other.
sD
NOTE: In distance vector routing, each node shares its routing table with its immediate
neighbors periodically and when there is a change
l
Updating
ria
When a node receives a two-column table from a neighbor, it needs to update its routing
table. Updating takes three steps:
1. The receiving node needs to add the cost between itself and the sending node to each value
to
the route.
3. The receiving node needs to compare each row of its old table with the corresponding row
of the modified version of the received table.
a. If the next-node entry is different, the receiving node chooses the row with the
smaller cost. If there is a tie, the old one is kept.
b. If the next-node entry is the same, the receiving node chooses the new row.
For example, suppose node C has previously advertised a route to node X with distance 3.
Suppose that now there is no path between C and X; node C now advertises this route with a
distance of infinity. Node A must not ignore this value even though its old entry is smaller. The
old route does not exist anymore. The new route has a distance of infinity.
m
.co
i ya
un
sD
Final Diagram
l
ria
to
Tu
When to Share
The question now is, When does a node send its partial routing table (only two columns) to all
its immediate neighbors? The table is sent both periodically and when there is a change in the
table.
Periodic Update A node sends its routing table, normally every 30 s, in a periodic update. The
period depends on the protocol that is using distance vector routing.
Triggered Update A node sends its two-column routing table to its neighbors anytime there is
a change in its routing table. This is called a triggered update. The change can result from the
m
following.
1. A node receives a table from a neighbor, resulting in changes in its own table after updating.
.co
2. A node detects some failure in the neighboring links which results in a distance change to
infinity.
ya
Two-node instability
i
un
l sD
ria
Three-node instability
to
Tu
be 1 and define 16 as infinity. However, this means that the distance vector routing cannot
be used in large systems. The size of the network, in each direction, cannot exceed 15 hops.
2. Split Horizon: In this strategy, instead of flooding the table through each interface, each
node sends only part of its table through each interface. If, according to its table, node B
thinks that the optimum route to reach X is via A, it does not need to advertise this piece of
information to A; the information has come from A (A already knows). Taking information
from node A, modifying it, and sending it back to node A creates the confusion. In our
m
scenario, node B eliminates the last line of its routing table before it sends it to A. In this
case, node A keeps the value of infinity as the distance to X. Later when node A sends its
.co
routing table to B, node B also corrects its routing table. The system becomes stable after
the first update: both node A and B know that X is not reachable.
ya
3. Split Horizon and Poison Reverse Using the split horizon strategy has one drawback.
Normally, the distance vector protocol uses a timer, and if there is no news about a route,
the node deletes the route from its table. When node B in the previous scenario eliminates
i
un
the route to X from its advertisement to A, node A cannot guess that this is due to the split
horizon strategy (the source of information was A) or because B has not received any news
about X recently. The split horizon strategy can be combined with the poison reverse
sD
strategy. Node B can still advertise the value for X, but if the source of information is A, it
can replace the distance with infinity as a warning: "Do not use this value; what I know
about this route comes from you."
l
ria
m
.co
ya
Building Routing Tables
i
1. Creation of the states of the links by each node, called the link state packet (LSP).
un
2. Dissemination of LSPs to every other router, called flooding, in an efficient and reliable way.
3. Formation of a shortest path tree for each node.
4. Calculation of a routing table based on the shortest path tree
sD
I. Creation of Link State Packet (LSP) A link state packet can carry a large amount of
l
information. For the moment, we assume that it carries a minimum amount of data: the
ria
node identity, the list of links, a sequence number, and age. The first two, node identity and
the list of links, are needed to make the topology. The third, sequence number, facilitates
flooding and distinguishes new LSPs from old ones. The fourth, age, prevents old LSPs from
to
2. A node that receives an LSP compares it with the copy it may already have. If the
newly arrived LSP is older than the one it has (found by checking the sequence number),
it discards the LSP. If it is newer, the node does the following:
a. It discards the old LSP and keeps the new one.
b. It sends a copy of it out of each interface except the one from which the packet
arrived. This guarantees that flooding stops somewhere in the domain (where a node
has only one interface).
III. Formation of Shortest Path Tree: Dijkstra Algorithm
m
A shortest path tree is a tree in which the path between the root and every other node is the
shortest.
.co
The Dijkstra algorithm creates a shortest path tree from a graph. The algorithm divides the
nodes into two sets: tentative and permanent. It finds the neighbors of a current node, makes
them tentative, examines them, and if they pass the criteria, makes them permanent.
i ya
un
l sD
ria
to
Tu
m
.co
i ya
un
IV. Calculation of a routing table
sD
huge amount of resources to calculate routing tables. It also creates heavy traffic because of
flooding. There is a need for a third routing protocol which we call path vector routing.
Path vector routing proved to be useful for inter domain routing. The principle of path vector
routing is similar to that of distance vector routing. In path vector routing, we assume that
there is one node (there can be more, but one is enough for our conceptual discussion) in
each AS that acts on behalf of the entire AS. Let us call it the speaker node. The speaker node
in an AS creates a routing table and advertises it to speaker nodes in the neighboring ASs. The
m
idea is the same as for distance vector routing except that only speaker nodes in each AS can
communicate with each other. However, what is advertised is different. A speaker node
.co
advertises the path, not the metric of the nodes, in its autonomous system or other
autonomous systems
ya
Initialization
Initial routing tables in path vector routing
i
un
l sD
ria
to
Tu
Sharing
Just as in distance vector routing, in path vector routing, a speaker in an autonomous system
shares its table with immediate neighbors. In Figure, node A1 shares its table with nodes B1
and C1. Node C1 shares its table with nodes D1, B1, and A1. Node B1 shares its table with C1
and A1. Node D1 shares its table with C1.
m
.co
i ya
un
Updating When a speaker node receives a two-column table from a neighbor, it updates its
own table by adding the nodes that are not in its routing table and adding its own autonomous
sD
system and the autonomous system that sent the table. After a while each speaker has a table
and knows how to reach each node in other Ass
a) Loop prevention. The instability of distance vector routing and the creation of loops can be
l
avoided in path vector routing. When a router receives a message, it checks to see if its AS
ria
is in the path list to the destination. If it is, looping is involved and the message is ignored.
b) Policy routing. Policy routing can be easily implemented through path vector routing.
When a router receives a message, it can check the path. If one of the AS listed in the path
to
is against its policy, it can ignore that path and that destination. It does not update its
routing table with this path, and it does not send this message to its neighbors.
Tu
c) Optimum path. What is the optimum path in path vector routing? We are looking for a
path to a destination that is the best for the organization that runs the AS. One system may
use RIP, which defines hop count as the metric; another may use OSPF with minimum delay
defined as the metric. In our previous figure, each AS may have more than one path to a
destination. For example, a path from AS4 to ASI can be AS4-AS3-AS2-AS1, or it can be AS4-
AS3-ASI. For the tables, we chose the one that had the smaller number of ASs, but this is
not always the case. Other criteria, such as security, safety, and reliability, can also be
applied
Hierarchical Routing
As networks grow in size, the router routing tables grow proportionally. Not only is router
memory consumed by ever-increasing tables, but more CPU time is needed to scan them and
more bandwidth is needed to send status reports about them.
At a certain point, the network may grow to the point where it is no longer feasible for every
router to have an entry for every other router, so the routing will have to be done
hierarchically, as it is in the telephone network.
When hierarchical routing is used, the routers are divided into what we will call regions. Each
m
router knows all the details about how to route packets to destinations within its own region
but knows nothing about the internal structure of other regions.
.co
For huge networks, a two-level hierarchy may be insufficient; it may be necessary to group the
regions into clusters, the clusters into zones, the zones into groups, and so on, until we run out
of names for aggregations
i ya
un
l sD
ria
to
Tu
When a single network becomes very large, an interesting question is ‘‘how many levels
should the hierarchy have?’’
For example, consider a network with 720 routers. If there is no hierarchy, each router needs
720 routing table entries.
If the network is partitioned into 24 regions of 30 routers each, each router needs 30 local
entries plus 23 remote entries for a total of 53 entries.
m
degrades performance. This situation is called congestion.
The network and transport layers share the responsibility for handling congestion. Since
.co
congestion occurs within the network, it is the network layer that directly experiences it and
must ultimately determine what to do with the excess packets.
However, the most effective way to control congestion is to reduce the load that the transport
ya
layer is placing on the network. This requires the network and transport layers to work
together. In this chapter we will look at the network aspects of congestion.
i
un
l sD
ria
When too much traffic is offered, congestion sets in and performance degrades sharply
to
Above Figure depicts the onset of congestion. When the number of packets hosts send into
Tu
the network is well within its carrying capacity, the number delivered is proportional to the
number sent. If twice as many are sent, twice as many are delivered. However, as the offered
load approaches the carrying capacity, bursts of traffic occasionally fill up the buffers inside
routers and some packets are lost. These lost packets consume some of the capacity, so the
number of delivered packets falls below the ideal curve. The network is now congested. Unless
the network is well designed, it may experience a congestion collapse
m
computer that is capable of handling only 1 Gbps. Although there is no congestion (the
network itself is not in trouble), flow control is needed to force the supercomputer to stop
.co
frequently to give the personal computer a chance to breathe.
At the other extreme, consider a network with 1-Mbps lines and 1000 large computers, half of
which are trying to transfer files at 100 kbps to the other half. Here, the problem is not that of
ya
fast senders overpowering slow receivers, but that the total offered traffic exceeds what the
network can handle.
i
un
The reason congestion control and flow control are often confused is that the best way to
handle both problems is to get the host to slow down. Thus, a host can get a ‘‘slow down’’
message either because the receiver cannot handle the load or because the network cannot
sD
handle it.
1. Warning bit
ria
2. Choke packets
3. Load shedding
4. Random early discard
to
5. Traffic shaping
The first 3 deal with congestion detection and recovery. The last 2 deal with congestion
Tu
avoidance
Warning Bit
1. A special bit in the packet header is set by the router to warn the source when congestion
is detected.
2. The bit is copied and piggy-backed on the ACK and sent to the sender.
3. The sender monitors the number of ACK packets it receives with the warning bit set and
adjusts its transmission rate accordingly.
Choke Packets
1. A more direct way of telling the source to slow down.
2. A choke packet is a control packet generated at a congested node and transmitted to
restrict traffic flow.
3. The source, on receiving the choke packet must reduce its transmission rate by a certain
percentage.
4. An example of a choke packet is the ICMP Source Quench Packet.
Hop-by-Hop Choke Packets
m
1. Over long distances or at high speeds choke packets are not very effective.
2. A more efficient method is to send to choke packets hop-by-hop.
.co
3. This requires each hop to reduce its transmission even before the choke packet arrive at
the source
ya
Load Shedding
1. When buffers become full, routers simply discard packets.
2. Which packet is chosen to be the victim depends on the application and on the error
i
un
strategy used in the data link layer.
3. For a file transfer, for, e.g. cannot discard older packets since this will cause a gap in the
received data.
sD
4. For real-time voice or video it is probably better to throw away old data and keep new
packets.
5. Get the application to mark packets with discard priority.
l
ria
2. Each time a packet arrives, the RED algorithm computes the average queue length, avg.
3. If avg is lower than some lower threshold, congestion is assumed to be minimal or non-
Tu
Traffic Shaping
1. Another method of congestion control is to “shape” the traffic before it enters the
network.
2. Traffic shaping controls the rate at which packets are sent (not just how many). Used in
ATM and Integrated Services networks.
3. At connection set-up time, the sender and carrier negotiate a traffic pattern (shape).
m
Leaky Bucket
Token Bucket
.co
The Leaky Bucket Algorithm used to control rate in a network. It is implemented as a single -
server queue with constant service time. If the bucket (buffer) overflows then packets are
ya
discarded.
i
un
l sD
ria
to
(a) A leaky bucket with water. (b) a leaky bucket with packets.
1. The leaky bucket enforces a constant output rate (average rate) regardless of the
Tu
1. In contrast to the LB, the Token Bucket Algorithm, allows the output rate to vary,
depending on the size of the burst.
2. In the TB algorithm, the bucket holds tokens. To transmit a packet, the host must capture
and destroy one token.
3. Tokens are generated by a clock at the rate of one token every t sec.
4. Idle hosts can capture and save up tokens (up to the max. size of the bucket) in order to
m
send larger bursts later.
.co
i ya
un
l sD
ria
2. With TB, a packet can only be transmitted if there are enough tokens to cover its length in
bytes.
3. LB sends packets at an average rate. TB allows for large bursts to be sent faster by speeding
up the output.
4. TB allows saving up tokens (permissions) to send large bursts. LB does not allow saving.
TRANSPORT LAYER
The network layer provides end-to-end packet delivery using datagrams
m
or virtual circuits.
.co
The transport layer builds on the network layer to provide data
transport from a process on a source machine to a process on a
destination machine with a desired level of reliability that is
ya
independent of the physical networks currently in use.
i
Services Provided to the Upper Layers
un
The ultimate goal of the transport layer is to provide efficient, reliable,
lsD
and cost-effective data transmission service to its users, normally
processes in the application layer.
To achieve this, the transport layer makes use of the services
ria
provided by the network layer. The software and/or hardware within
the transport layer that does the work is called the transport entity.
to
Tu
m
.co
i ya
un
lsD
ria
The network, transport, and application layers.
to
Tu
m
Addressing and flow control
The connectionless transport service .
.co
i ya
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
23.7
m
.co
i ya
un
lsD
ria
to
Tu
23.8
m
.co
i ya
un
lsD
ria
to
Tu
23.9
m
.co
i ya
un
lsD
ria
The primitives for a simple transport
to
service.
Tu
m
client turns up.
.co
When a client wants to talk to the server, it executes a CONNECT
primitive. The transport entity carries out this primitive by blocking the
ya
caller and sending a packet to the server. The client’s CONNECT call
causes a CONNECTION REQUEST segment to be sent to the server.
i
When it arrives, the transport entity checks to see that the server is
un
blocked on a LISTEN (i.e., is interested in handling requests). If so, it then
unblocks the server and sends a CONNECTION ACCEPTED segment
lsD
back to the client. When this segment arrives, the client is unblocked and
the connection is established.
ria
Data can now be exchanged using the SEND and RECEIVE primitives. In
the simplest form, either party can do a (blocking) RECEIVE to wait for the
other party to do a SEND. When the segment arrives, the receiver is
to
unblocked. It can then process the segment and send a reply. As long as
Tu
both sides can keep track of whose turn it is to send, this scheme works
fine.
m
variants: asymmetric and symmetric.
.co
In the asymmetric variant, either transport user can issue a DISCONNECT
primitive, which results in a DISCONNECT segment being sent to the
ya
remote transport entity. Upon its arrival, the connection is released.
i
In the symmetric variant, each direction is closed separately, independently
un
of the other one. When one side does a DISCONNECT, that means it has
no more data to send but it is still willing to accept data from its partner. In
lsD
this model, a connection is released when both sides have done a
DISCONNECT
ria
to
Tu
m
.co
i ya
un
lsD
ria
solid lines show the client's state sequence. The dashed lines show
the server's state sequence.
Download FREE Computer Science Notes at TutorialsDuniya.com
Download FREE Computer Science Notes at TutorialsDuniya.com
Elements of Transport
m
Protocols
.co
• Addressing
ya
• Connection Establishment
i
• Connection Release
un
• Flow Control and Buffering
• lsD
Multiplexing
ria
• Crash Recovery
to
Tu
Transport Protocol
m
.co
i ya
un
lsD
ria
to
m
explicit addressing of destinations is required.
.co
2 The process of establishing a connection over the wire of Fig(a) is simple:
the other end is always there (unless it has crashed, in which case it is not
ya
there). Either way, there is not much to do. Even on wireless links the
process is not much different. Just sending a message is sufficient to have it
i
reach all other destinations. If the message is not acknowledged due to an
un
error, it can be resent. In the transport layer, initial connection establishment
is complicated, as we will see.
lsD
3 Another (exceedingly annoying) difference between the data link layer and
the transport layer is the potential existence of storage capacity in the
ria
network. The consequences of the network’s ability to delay and duplicate
packets can sometimes be disastrous and can require the use of special
protocols to correctly transport information.
to
4. Buffering and flow control are needed in both layers, but the
Tu
Addressing
m
When an application (e.g., a user) process wishes to set up a connection to
.co
a remote application process, it must specify which one to connect to.
(Connectionless transport has the same problem: to whom should each
message be sent?) The method normally used is to define transport
ya
addresses to which processes can listen for connection requests. In the
Internet, these endpoints are called ports. We will use the generic term
i
TSAP (Transport Service Access Point) to mean a specific endpoint in the
un
transport layer. The analogous endpoints in the network layer (i.e., network
layer addresses) are not-surprisingly called NSAPs (Network Service
lsD
Access Points). IP addresses are examples of NSAPs.
ria
to
Tu
Addressing
m
.co
i ya
un
lsD
ria
to
connections.
Download FREE Computer Science Notes at TutorialsDuniya.com
Download FREE Computer Science Notes at TutorialsDuniya.com
m
an incoming call. A call such as our LISTEN might be used, for example.
2. An application process on host 1 wants to send an email message, so it
.co
attaches itself to TSAP 1208 and issues a CONNECT request. The
request specifies TSAP 1208 on host 1 as the source and TSAP 1522 on
ya
host 2 as the destination. This action ultimately results in a transport
connection being established between the application process and the
i
server.
un
3. The application process sends over the mail message.
4. The mail server responds to say that it will deliver the message.
lsD
5. The transport connection is released.
ria
CONNECTION ESTABLISHMENT
Establishing a connection sounds easy, but it is actually surprisingly tricky. At
m
first glance, it would seem sufficient for one transport entity to just send a
CONNECTION REQUEST segment to the destination and wait for a
.co
CONNECTION ACCEPTED reply. The problem occurs when the network
can lose, delay, corrupt, and duplicate packets. This behavior causes serious
ya
complications
i
un
introduced the three-way handshake. This establishment protocol
involves one peer checking with the other that the connection request is
lsD
indeed current. The normal setup procedure when host 1 initiates is shown
in Fig. (a). Host 1 chooses a sequence number, x, and sends a
CONNECTION REQUEST segment containing it to host 2. Host 2 replies
ria
with an ACK segment acknowledging x and announcing its own initial
sequence number, y. Finally, host 1 acknowledges host 2’s choice of an
initial sequence number in the first data segment that it sends.
to
Tu
Connection Establishment
m
.co
i ya
un
lsD
Three protocol scenarios for establishing a connection using a
three-way handshake. CR denotes CONNECTION REQUEST.
ria
(a) Normal operation,
(b) Old CONNECTION REQUEST appearing out of nowhere.
to
m
ACK segment, in effect asking for verification that host 1 was indeed trying
.co
to set up a new connection. When host 1 rejects host 2’s attempt to
establish a connection, host 2 realizes that it was tricked by a delayed
duplicate and abandons the connection. In this way, a delayed duplicate
ya
does no damage
i
The worst case is when both a delayed CONNECTION REQUEST and an
un
ACK are floating around in the subnet. This case is shown in Fig. (c). As in
the previous example, host 2 gets a delayed CONNECTION REQUEST
lsD
and replies to it. At this point, it is crucial to realize that host 2 has proposed
using y as the initial sequence number for host 2 to host 1 traffic, knowing
full well that no segments containing sequence number y or
ria
acknowledgements to y are still in existence. When the second delayed
segment arrives at host 2, the fact that z has been acknowledged rather
than y tells host 2 that this, too, is an old duplicate. The important thing to
to
realize here is that there is no combination of old segments that can cause
Tu
the protocol to fail and have a connection set up by accident when no one
wants it.
Connection Release
m
.co
i ya
un
lsD
ria
to
Tu
m
the connection as two separate unidirectional connections and requires each
.co
one to be released separately
Asymmetric release is abrupt and may result in data loss. Consider the
ya
scenario of Fig. After the connection is established, host 1 sends a segment
that arrives properly at host 2. Then host 1 sends another segment.
i
Unfortunately, host 2 issues a DISCONNECT before the second segment
un
arrives. The result is that the connection is released and data are lost.
lsD
Clearly, a more sophisticated release protocol is needed to avoid data loss.
One way is to use symmetric release, in which each direction is released
independently of the other one. Here, a host can continue to receive data even
ria
after it has sent a DISCONNECT segment.
Symmetric release does the job when each process has a fixed amount of data
to send and clearly knows when it has sent it. One can envision a protocol in
to
which host 1 says ‘‘I am done. Are you done too?’’ If host 2 responds: ‘‘I am
Tu
Connection Release
m
.co
The two-army problem.
i ya
un
lsD
ria
to
Tu
Connection Release
m
.co
i ya
un
6-14, a, b
lsD
ria
to
Connection Release
m
.co
i ya
un
6-14, c,d
lsD
ria
to
In Fig. (a), we see the normal case in which one of the users
sends a DR (DISCONNECTION REQUEST) segment to initiate the
m
connection release. When it arrives, the recipient sends back a
DR segment and starts a timer, just in case its DR is lost. When
.co
this DR arrives, the original sender sends back an ACK segment
and releases the connection. Finally, when the ACK segment
ya
arrives, the receiver also releases the connection.
i
If the final ACK segment is lost, as shown in Fig.(b), the situation
un
is saved by the timer. When the timer expires, the connection is
released anyway. Now consider the case of the second DR being
lsD
lost. The user initiating the disconnection will not receive the
expected response, will time out, and will start all over again.
ria
In Fig.(c), we see how this works, assuming that the second time
no segments are lost and all segments are delivered correctly
and on time.
to
Last scenario, Fig.(d), is the same as Fig. (c) except that now we
Tu
TCP
TCP is a connection oriented protocol; it creates a virtual connection
m
between two TCPs to send data. In addition, TCP uses flow and error
.co
control mechanisms at the transport level. In brief, TCP is called a
connection-oriented, reliable transport protocol. It adds connection-oriented
and reliability features to the services of IP.
ya
Topics discussed in this section:
i
un
TCP Services
TCP Features
Segment lsD
A TCP Connection
ria
Flow Control
Error Control
to
Tu
TCP Services
m
1 Process-to-Process Communication
TCP provides process-to-process communication using port
.co
numbers. Below Table lists some well-known port numbers used by
TCP.
i ya
un
lsD
ria
to
Tu
m
stream of bytes and allows the receiving process to obtain data as a
stream of bytes. TCP creates an environment in which the two processes
.co
seem to be connected by an imaginary "tube“ that carries their data
across the Internet. This imaginary environment is showed in below
ya
Figure. The sending process produces (writes to) the stream of bytes, and
the receiving process consumes (reads from) them
i
un
lsD
ria
to
Tu
3 Sending and Receiving Buffers Because the sending and the receiving
processes may not write or read data at the same speed, TCP needs
m
buffers for storage. There are two buffers, the sending buffer and the
receiving buffer, one for each direction. One way to implement a buffer is
.co
to use a circular array of I-byte locations as shown in Figure. For simplicity,
we have shown two buffers of 20 bytes each. Normally the buffers are
ya
hundreds or thousands of bytes, depending on the implementation. We
also show the buffers as the same size, which is not always the case.
i
un
lsD
ria
to
Tu
Figure shows the movement of the data in one direction. At the sending
m
site, the buffer has three types of chambers. The white section contains
empty chambers that can be filled by the sending process (producer). The
.co
gray area holds bytes that have been sent but not yet acknowledged. TCP
keeps these bytes in the buffer until it receives an acknowledgment. The
colored area contains bytes to be sent by the sending TCP.
ya
However, as we will see later in this chapter, TCP may be able to send
only part of this colored section. This could be due to the slowness of the
i
un
receiving process or perhaps to congestion in the network. Also note that
after the bytes in the gray chambers are acknowledged, the chambers are
recycled and available for use by the sending process.
lsD
This is why we show a circular buffer.
The operation of the buffer at the receiver site is simpler. The circular
ria
buffer is divided into two areas (shown as white and colored). The white
area contains empty chambers to be filled by bytes received from the
to
network. The colored sections contain received bytes that can be read by
the receiving process. When a byte is read by the receiving process, the
Tu
4 TCP segments
m
.co
i ya
un
At the transport layer, TCP groups a number of bytes together into a packet
lsD
called a segment. TCP adds a header to each segment (for control
purposes) and delivers the segment to the IP layer for transmission. The
ria
segments are encapsulated in IP datagrams and transmitted.
This entire operation is transparent to the receiving process. Later we will
see that segments may be received out of order, lost, or corrupted and
to
resent. All these are handled by TCP with the receiving process unaware of
any activities. Above fig shows how segments are created from the bytes in
Tu
the buffers
23.34
5 Full-Duplex Communication
TCP offers full-duplex service, in which data can flow in both directions at
m
the same time. Each TCP then has a sending and receiving buffer, and
segments move in both directions
.co
6 Connection-Oriented Service
ya
TCP is a connection-oriented protocol. When a process at site A wants to
send and receive data from another process at site B, the following occurs:
i
1. The two TCPs establish a connection between them.
un
2. Data are exchanged in both directions.
3. The connection is terminated.
TCP Features
1 Numbering System
m
There are two fields called the sequence number and the
.co
acknowledgment number. These two fields refer to the byte number and
not the segment number.
Byte Number The bytes of data being transferred in each connection
ya
are numbered by TCP. The numbering starts with a randomly generated
number. For example, if the random number happens to be 1057 and the
i
total data to be sent are 6000 bytes, the bytes are numbered from 1057
un
to 7056. We will see that byte numbering is used for flow and error
control.
lsD
Sequence Number After the bytes have been numbered, TCP assigns a
sequence number to each segment that is being sent. The sequence
number for each segment is the number of the first byte carried in that
ria
segment.
Acknowledgment Number The value of the acknowledgment field in a
segment defines the number of the next byte a party expects to receive.
to
2 Flow Control
TCP, provides flow control. The receiver of the data controls the amount
m
of data that are to be sent by the sender. This is done to prevent the
.co
receiver from being overwhelmed with data. The numbering system
allows TCP to use a byte-oriented flow control.
ya
3 Error Control
To provide reliable service, TCP implements an error control mechanism.
i
Although error control considers a segment as the unit of data for error
un
detection (loss or corrupted segments), error control is byte-oriented, as
we will see later.
m
.co
i ya
un
lsD
ria
to
Tu
23.38
m
application program in the host that is sending the segment.
Destination port address. This is a 16-bit field that defines the port number of the
.co
application program in the host that is receiving the segment.
Sequence number. This 32-bit field defines the number assigned to the first byte
of data contained in this segment. As we said before, TCP is a stream transport
ya
protocol. To ensure connectivity, each byte to be transmitted is numbered. The
sequence number tells the destination which byte in this sequence comprises the
i
first byte in the segment. During connection establishment, each party uses a
un
random number generator to create an initial sequence number (ISN), which is
usually different in each direction.
lsD
Acknowledgment number. This 32-bit field defines the byte number that the
receiver of the segment is expecting to receive from the other party. If the
receiver of the segment has successfully received byte number x from the other
ria
party, it defines x + I as the acknowledgment number. Acknowledgment and data
can be piggybacked together.
Header length. This 4-bit field indicates the number of 4-byte words in the TCP
to
header. The length of the header can be between 20 and 60 bytes. Therefore,
the value of this field can be between 5 (5 x 4 =20) and 15 (15 x 4 =60).
Tu
m
.co
ya
These bits enable flow control, connection establishment and termination,
i
connection abortion, and the mode of data transfer in TCP.
un
Window size. This field defines the size of the window, in bytes, that the other
party must maintain. Note that the length of this field is 16 bits, which means
lsD
that the maximum size of the window is 65,535 bytes. This value is normally
referred to as the receiving window (rwnd) and is determined by the receiver.
The sender must obey the dictation of the receiver in this case.
ria
Checksum. This 16-bit field contains the checksum. The calculation of the
checksum for TCP follows the same procedure as the one described for UDP.
to
pseudoheader, serving the same purpose, is added to the segment. For the
TCP pseudoheader, the value for the protocol field is 6.
Urgent pointer. This l6-bit field, which is valid only if the urgent flag is set, is
used when the segment contains urgent data. It defines the number that
m
must be added to the sequence number to obtain the number of the last
urgent byte in the data section of the segment. This will be discussed later
.co
in this chapter.
Options. There can be up to 40 bytes of optional information in the TCP
ya
header. We will not discuss these options here; please refer to the
reference list for more information.
i
un
lsD
ria
to
Tu
m
A TCP Connection
.co
TCP is connection-oriented. A connection-oriented transport protocol
establishes a virtual path between the source and destination. All the
segments belonging to a message are then sent over this virtual path.
ya
Using a single virtual pathway for the entire message facilitates the
acknowledgment process as well as retransmission of damaged or lost
i
frames.
un
In TCP, connection-oriented transmission requires three phases:
3. connection termination.
ria
to
Tu
m
1 The client sends the first segment, a SYN segment, in which only the SYN
flag is set.
.co
NOTE:A SYN segment cannot carry data, but it consumes one
sequence number.
ya
2. The server sends the second segment, a SYN +ACK segment, with 2 flag
bits set: SYN and ACK. This segment has a dual purpose. It is a SYN
i
un
segment for communication in the other direction and serves as the
acknowledgment for the SYN segment. It consumes one sequence number.
NOTE:A SYN+ACK segment cannot carry data, but does consume
one sequence number lsD
3. The client sends the third segment. This is just an ACK segment. It
ria
acknowledges the receipt of the second segment with the ACK flag and
acknowledgment number field. Note that the sequence number in this
to
segment is the same as the one in the SYN segment; the ACK segment does
not consume any sequence numbers.
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
segments to a server, pretending that each of them is corning from a
different client by faking the source IP addresses in the datagram's.
.co
The server, assuming that the clients are issuing an active open, allocates
the necessary resources, such as creating communication tables and
ya
setting timers. The TCP server then sends the SYN +ACK segments to the
fake clients, which are lost. During this time, however, a lot of resources
i
are occupied without being used. If, during this short time, the number of
un
SYN segments is large, the server eventually runs out of resources and
may crash. This SYN flooding attack belongs to a type of security attack
lsD
known as a denial-of-service attack, in which an attacker monopolizes a
system with so many service requests that the system collapses and
denies service to every request.
ria
SOLUTIONS:
1 Some have imposed a limit on connection requests during a
specified period of
to
time.
2 Others filter out datagrams coming from unwanted source
Tu
addresses.
3 One recent strategy is to postpone resource allocation until the
entire connection is set up, using what is called a cookie.
Download FREE Computer Science Notes at TutorialsDuniya.com
Download FREE Computer Science Notes at TutorialsDuniya.com
Data Transfer
After connection is established, bidirectional data transfer can take place.
m
The client and server can both send data and acknowledgments. Data
traveling in the same direction as an acknowledgment are carried on the
.co
same segment. The acknowledgment is piggybacked with the data
ya
In this example, after connection is established (not shown in the figure),
the client sends 2000 bytes of data in two segments. The server then
i
sends 2000 bytes in one segment. The client sends one more segment.
un
The first three segments carry both data and acknowledgment, but the last
segment carries only an acknowledgment because there are no more data
to be sent.
lsD
Note the values of the sequence and acknowledgment numbers. The data
segments sent by the client have the PSH (push) flag set so that the
ria
server TCP knows to deliver data to the server process as soon as they
are received.
to
Tu
Data transfer
m
.co
i ya
un
lsD
ria
to
Tu
m
TCP can handle such a situation. The application program at the sending
site can request a push operation. This means that the sending TCP must
.co
not wait for the window to be filled. It must create a segment and send it
immediately. The sending TCP must also set the push bit (PSH) to let the
ya
receiving TCP know that the segment includes data that must be
delivered to the receiving application program as soon as possible and
not to wait for more data to come.
i
un
Urgent Data : TCP is a stream-oriented protocol. This means that the
data are presented from the application program to TCP as a stream of
lsD
bytes. Each byte of data has a position in the stream. However, sending
application program wants a piece of data to be read out of order by the
ria
receiving application program.
to
Tu
m
1. In a normal situation, the client TCP, after receiving a close command
.co
from the client process, sends the first segment, a FIN segment in which the
FIN flag is set.
ya
Note that a FIN segment can include the last chunk of data sent by the
client, or it can be just a control segment as shown in Figure. If it is only a
i
control segment, it consumes only one sequence number.
un
NOTE: The FIN segment consumes one sequence number ifit does
not carry data.
lsD
2 The server TCP, after receiving the FIN segment, informs its process of
the situation and sends the second segment, a FIN +ACK segment, to
ria
confirm the receipt of the FIN segment from the client and at the same time
to announce the closing of the connection in the other direction. This
segment can also contain the last chunk of data from the server. If it does
to
3. The client TCP sends the last segment, an ACK segment, to confirm the
receipt of the FIN segment from the TCP server. This segment contains the
m
acknowledgment number, which is 1 plus the sequence number received in
the FIN segment from the server. This segment cannot carry data and
.co
consumes no sequence numbers.
ya
Half-Close In TCP, one end can stop sending data while still receiving
data. This is called a half-close. Although either end can issue a half-close,
it is normally initiated by the client. It can occur when the server needs all
i
un
the data before processing can begin.
A good example is sorting. When the client sends data to the server to be
sorted, the server needs to receive all the data before sorting can start.
lsD
This means the client, after sending all the data, can close the connection
in the outbound direction. However, the inbound direction must remain
ria
open to receive the sorted data. The server, after receiving the data, still
needs time for sorting; its outbound direction must remain open
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
protocol used by TCP, however, is something between the Go-Back-N and
Selective Repeat sliding window.
.co
The sliding window protocol in TCP looks like the Go-Back-N protocol
ya
because it does not use NAKs;
it looks like Selective Repeat because the receiver holds the out-of-order
i
segments until the missing ones arrive.
un
There are two big differences between this sliding window and the one
lsD
we used at the data link layer.
1 the sliding window of TCP is byte-oriented; the one we discussed in the
data link layer is frame-oriented.
ria
2 the TCP's sliding window is of variable size; the one we discussed in
the data link layer was of fixed size
to
Tu
Sliding window
m
.co
i ya
un
lsD
ria
to
Tu
m
see, are in the control of the receiver (and depend on congestion in the
network), not the sender.
.co
The sender must obey the commands of the receiver in this matter.
Opening a window means moving the right wall to the right. This allows
more new bytes in the buffer that are eligible for sending.
ya
Closing the window means moving the left wall to the right. This means
that some bytes have been acknowledged and the sender need not worry
i
un
about them anymore.
Shrinking the window means moving the right wall to the left.
lsD
The size of the window at one end is determined by the lesser of two
values: receiver window (rwnd) or congestion window (cwnd).
The receiver window is the value advertised by the opposite end in a
ria
segment containing acknowledgment. It is the number of bytes the other
end can accept before its buffer overflows and data are discarded.
to
m
.co
i ya
un
lsD
ria
to
Tu
m
running on the remote machine.
.co
2) the sender may send a 1-byte segment to force the receiver to
reannounce the next byte expected and the window size. This packet is
called a window probe.
ya
The TCP standard explicitly provides this option to prevent deadlock if a
window update ever gets lost.
i
un
Senders are not required to transmit data as soon as they come in from the
application. Neither are receivers required to send acknowledgements as
soon as possible. lsD
ria
For example, in Fig. when the first 2 KB of data came in, TCP, knowing that it
had a 4-KB window, would have been completely correct in just buffering the
data until another 2 KB came in, to be able to transmit a segment with a 4-KB
to
m
3.
.co
send echo of character
and/or output 2.
interpret
character
1.
ya
send character
Host with Host with
Telnet client Telnet server
i
un
Remote terminal applications (e.g., Telnet) send characters to a server.
lsD
The server interprets the character and sends the output at the server
to the client.
ria
For each character typed, you see three packets:
Client Server: Send typed character
to
Delayed Acknowledgement
m
• TCP delays transmission of ACKs for up to 500ms
.co
• Avoid to send ACK packets that do not carry data.
– The hope is that, within the delay, the receiver will have data ready to
ya
be sent to the receiver. Then, the ACK can be piggybacked with a data
segment
i
un
Exceptions:
• ACK should be sent for every full sized segment
lsD
• Delayed ACK is not used when packets arrive out of order
59
Nagel’s Rule
m
Send one byte and buffer all subsequent bytes until acknowledgement is
received. Then send all buffered bytes in a single TCP segment and start
.co
buffering again until the sent segment is acknowledged.
Nagle’s algorithm will put the many pieces in one segment, greatly reducing
the bandwidth used
ya
Nagle’s algorithm is widely used by TCP implementations, but there are
i
un
times when it is better to disable it. In particular, in interactive games that are
run over the Internet.
A more subtle problem is that Nagle’s algorithm can sometimes interact with
lsD
delayed acknowledgements to cause a temporary deadlock: the receiver
waits for data on which to piggyback an acknowledgement, and the sender
waits on the acknowledgement to send more data.
ria
Another problem that can degrade TCP performance is the silly window
syndrome (Clark, 1982).
m
.co
i ya
un
lsD
ria
to
Tu
Clark’s solution is to prevent the receiver from sending a window update for
1 byte. Instead, it is forced to wait until it has a decent amount of space
m
available and advertise that instead. Specifically, the receiver should not
.co
send a window update until it can handle the maximum segment size it
advertised when the connection was established or until its buffer is half
empty, whichever is smaller.
ya
Furthermore, the sender can also help by not sending tiny segments.
i
Instead, it should wait until it can send a full segment, or at least one
un
containing half of the receiver’s buffer size.
lsD
The goal is for the sender not to send small segments and the receiver not
to ask for them. (Nagel + Clark). Both are used to improve TCP
performance
ria
The receiver will buffer the data until it can be passed up to the
application in order (handling out of order segments)
to
Tu
Cumulative acknowledgements
Error Control
TCP is a reliable transport layer protocol. This means that an application
m
program that delivers a stream of data to TCP relies on TCP to deliver the
entire stream to the application program on the other end in order, without
.co
error, and without any part lost or duplicated.
ya
TCP provides reliability using error control. Error control includes
mechanisms for detecting corrupted segments, lost segments, out-of-order
segments, and duplicated segments. Error control also includes a
i
un
mechanism for correcting errors after they are detected. Error detection and
correction in TCP is achieved through the use of three simple tools:
checksum, acknowledgment, and time-out.
Checksum
lsD
ria
Each segment includes a checksum field which is used to check for a
corrupted segment. If the segment is corrupted, it is discarded by the
destination TCP and is considered as lost. TCP uses a 16-bit checksum that
to
m
.co
i ya
un
lsD
ria
to
Tu
23.64
Acknowledgment
TCP uses acknowledgments to confirm the receipt of data segments.
m
Control segments that carry no data but consume a sequence number are
also acknowledged. ACK segments are never acknowledged.
.co
ACK segments do not consume sequence numbers and are not
acknowledged.
ya
Retransmission
i
The heart of the error control mechanism is the retransmission of
un
segments. When a segment is corrupted, lost, or delayed, it is
retransmitted.
lsD
In modern implementations, a retransmission occurs if the retransmission
timer expires or three duplicate ACK segments have arrived.
ria
Retransmission After RTO (retransmission time out)
Retransmission After Three Duplicate ACK Segments (also called fast
retransmission)
to
Out-of-Order Segments
Tu
Data may arrive out of order and be temporarily stored by the receiving
TCP, but yet guarantees that no out-of-order segment is delivered to the
process
Download FREE Computer Science Notes at TutorialsDuniya.com
Download FREE Computer Science Notes at TutorialsDuniya.com
m
When the load offered to any network is more than it can handle,
congestion builds up.
.co
The network layer detects congestion when queues grow large at routers
and tries to manage it, if only by dropping packets. It is up to the transport
ya
layer to receive congestion feedback from the network layer and slow down
the rate of traffic that it is sending into the network.
i
un
For Congestion control, transport protocol uses an AIMD (Additive Increase
Multiplicative Decrease) control law.
lsD
TCP congestion control is based on implementing this approach using a
window called congestion window. TCP adjusts the size of the window
ria
according to the AIMD rule.
where
flow control window is advertised by the receiver (rwnd)
congestion window is adjusted based on feedback from the
Download FREE Computer Science Notes at TutorialsDuniya.com
Download FREE Computer Science Notes at TutorialsDuniya.com
Modern congestion control was added to TCP largely through the efforts of
Van Jacobson (1988). It is a fascinating story. Starting in 1986, the growing
m
popularity of the early Internet led to the first occurrence of what became
known as a congestion collapse, a prolonged period during which good
.co
put dropped suddenly (i.e., by more than a factor of 100) due to congestion
in the network. Jacobson (and many others) set out to understand what
ya
was happening and remedy the situation.
i
To start, he observed that packet loss is a suitable signal of congestion.
un
This signal comes a little late (as the network is already congested) but it is
quite dependable
lsD
At the beginning how sender knows at what speed receiver can receive the
packets?
ria
to
Tu
m
.co
i ya
un
lsD
The key observation is this: the acknowledgements return to the sender
at about the rate that packets can be sent over the slowest link in the
path. This is precisely the rate that the sender wants to use. If it injects
ria
new packets into the network at this rate, they will be sent as fast as the
slow link permits, but they will not queue up and congest any router
along the path. This timing is known as an ack clock. It is an essential
to
part of TCP. By using an ack clock, TCP smoothes out traffic and avoids
unnecessary queues at routers. This is first consideration
Tu
A second consideration is that the AIMD rule will take a very long time to
reach a good operating point on fast networks if the congestion window is
m
started from a small size
.co
Instead, the solution Jacobson chose to handle both of these
considerations is a mix of linear and multiplicative increase.
ya
SLOW-START
i
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
process is restarted.
.co
Congestion avoidance phase is started if cwnd has reached the slow
start threshold value
ya
Whenever the slow start threshold is crossed, TCP switches from slow
start to additive increase. In this mode, the congestion window is
i
increased by one segment every round-trip time.
un
lsD
ria
to
Tu
71
m
.co
i ya
un
lsD
ria
to
Tu
Responses to Congestion
m
.co
• So, TCP assumes there is congestion if it
detects a packet loss
ya
• A TCP sender can detect lost packets via:
• Timeout of a retransmission timer
i
un
• Receipt of a duplicate ACK
lsD
• TCP interprets a Timeout as a binary congestion signal. When a
timeout occurs, the sender performs:
ria
– cwnd is reset to one:
cwnd = 1
– ssthresh is set to half the current size of the congestion window:
to
ssthressh = cwnd / 2
– and slow-start is entered
Tu
73
Fast Retransmit
m
.co
• If three or more duplicate
ACKs are received in a
ya
row, the TCP sender
believes that a segment
has been lost.
i
un
• Then TCP performs a
lsD
retransmission of what
seems to be the missing
segment, without waiting
ria
for a timeout to happen.
to
cwnd = 1
74
m
.co
• TCP Tahoe (1988)
ya
– Slow Start
– Congestion Avoidance
i
– Fast Retransmit
un
• TCP Reno (1990) (TCP Tahoe+FR)
– Fast Recovery
• New Reno (1996) lsD
• SACK (1996) (SACK (Selective
ria
ACKnowledgements))
to
75
m
.co
i ya
un
lsD
ria
to
Tu
m
as a congestion signal. ECN is an IP layer mechanism to notify hosts of
congestion.
.co
The sender tells the receiver that it has heard the signal by using the CWR
(Congestion Window Reduced) flag.
i ya
un
lsD
ria
to
Tu
USER DATAGRAM PROTOCOL (UDP)
m
.co
The User Datagram Protocol (UDP) is called a
ya
connectionless, unreliable transport protocol. It does
not add anything to the services of IP except to provide
i
un
process-to-process communication instead of host-to-
host communication.
lsD
Topics discussed in this section:
ria
Well-Known Ports for UDP
User Datagram
Checksum
to
UDP Operation
Tu
Use of UDP
m
.co
i ya
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
23.80
m
The UDP checksum calculation is different from the one for IP and ICMP. Here
the checksum includes three sections: a pseudo header, the UDP header,
.co
and the data coming from the application layer.
The pseudo header is the part of the header of the IP packet in which the user
ya
datagram is to be encapsulated with some fields filled with Os
If the checksum does not include the pseudo header, a user datagram may
i
un
arrive safe and sound. However, if the IP header is corrupted, it may be
delivered to the wrong host.
The protocol field is added to ensure that the packet belongs to UDP, and not to
lsD
other transport-layer protocols.
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
23.82
UDP Operation
Connectionless Services
m
UDP provides a connectionless service. This means that each user
datagram sent by UDP is an independent datagram. There is no
.co
relationship between the different user datagrams even if they are coming
from the same source process and going to the same destination program.
The user datagrams are not numbered. Also, there is no connection
ya
establishment and no connection termination, as is the case for TCP. This
means that each user datagram can travel on a different path.
i
Flow and Error Control
un
UDP is a very simple, unreliable transport protocol. There is no flow control
and hence no window mechanism. The receiver may overflow with
lsD
incoming messages. There is no error control mechanism in UDP except for
the checksum. This means that the sender does not know if a message has
been lost or duplicated. When the receiver detects an error through the
checksum, the user datagram is silently discarded. The lack of flow control
ria
and error control
Encapsulation and Decapsulation
To send a message from one process to another, the UDP protocol
to
m
Figure 23.11 shows the checksum calculation for a very
.co
small user datagram with only 7 bytes of data. Because the
ya
number of bytes of data is odd, padding is added for
checksum calculation. The pseudoheader as well as the
i
un
padding will be dropped when the user datagram is
delivered to IP.
lsD
ria
to
Tu
23.84
m
.co
i ya
un
lsD
ria
to
Tu
23.85
m
.co
iya
un
lsD
ria
to
Tu
23.86
m
Birrell and Nelson suggested was allowing programs to call procedures
located on remote hosts. When a process on machine 1 calls a procedure
.co
on machine 2, the calling process on 1 is suspended and execution of the
called procedure takes place on 2. Information can be transported from the
caller to the callee in the parameters and can come back in the procedure
ya
result. No message passing is visible to the application programmer. This
technique is known as RPC (Remote Procedure Call). Traditionally, the
i
un
calling procedure is known as the client and the called procedure is known
as the server, and we will use those names here too.
lsD
to call a remote procedure, the client program must be bound with a small
library procedure, called the client stub, that represents the server
procedure in the client’s address space. Similarly, the server is bound with a
ria
procedure called the server stub. These procedures hide the fact that the
procedure call from the client to the server is not local
to
Tu
Step 1 is the client calling the client stub. This call is a local procedure call,
with the parameters pushed onto the stack in the normal way.
m
Step 2 is the client stub packing the parameters into a message and
making a system call to send the message. Packing the parameters is
.co
called marshaling.
Step 3 is the operating system sending the message from the client
ya
machine to the server machine.
Step 4 is the operating system passing the incoming packet to the server
i
stub.
un
Finally, step 5 is the server stub calling the server procedure with the
unmarshaled parameters.
lsD
The reply traces the same path in the other direction.
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
1 With RPC, passing pointers is impossible because the client and server are in
different address spaces.
.co
2 It is essentially impossible for the client stub to marshal the parameters: it
has no way of determining how large they are.
ya
3 A third problem is that it is not always possible to deduce the types of the
parameters, not even from a formal specification or the code itself.(exa:
i
printf)
un
4 A fourth problem relates to the use of global variables. Normally, the calling
and called procedure can communicate by using global variables, in addition to
lsD
communicating via parameters. But if the called procedure is moved to a
remote machine, the code will fail because the global variables are no longer
shared
ria
to
Tu
TCP
TCP is a connection oriented protocol; it creates a virtual connection
m
between two TCPs to send data. In addition, TCP uses flow and error
.co
control mechanisms at the transport level. In brief, TCP is called a
connection-oriented, reliable transport protocol. It adds connection-oriented
and reliability features to the services of IP.
ya
Topics discussed in this section:
i
un
TCP Services
TCP Features
Segment lsD
A TCP Connection
ria
Flow Control
Error Control
to
Tu
TCP Services
m
1 Process-to-Process Communication
TCP provides process-to-process communication using port
.co
numbers. Below Table lists some well-known port numbers used by
TCP.
i ya
un
lsD
ria
to
Tu
m
stream of bytes and allows the receiving process to obtain data as a
stream of bytes. TCP creates an environment in which the two processes
.co
seem to be connected by an imaginary "tube“ that carries their data
across the Internet. This imaginary environment is showed in below
ya
Figure. The sending process produces (writes to) the stream of bytes, and
the receiving process consumes (reads from) them
i
un
lsD
ria
to
Tu
3 Sending and Receiving Buffers Because the sending and the receiving
processes may not write or read data at the same speed, TCP needs
m
buffers for storage. There are two buffers, the sending buffer and the
receiving buffer, one for each direction. One way to implement a buffer is
.co
to use a circular array of I-byte locations as shown in Figure. For simplicity,
we have shown two buffers of 20 bytes each. Normally the buffers are
ya
hundreds or thousands of bytes, depending on the implementation. We
also show the buffers as the same size, which is not always the case.
i
un
lsD
ria
to
Tu
Figure shows the movement of the data in one direction. At the sending
m
site, the buffer has three types of chambers. The white section contains
empty chambers that can be filled by the sending process (producer). The
.co
gray area holds bytes that have been sent but not yet acknowledged. TCP
keeps these bytes in the buffer until it receives an acknowledgment. The
colored area contains bytes to be sent by the sending TCP.
ya
However, as we will see later in this chapter, TCP may be able to send
only part of this colored section. This could be due to the slowness of the
i
un
receiving process or perhaps to congestion in the network. Also note that
after the bytes in the gray chambers are acknowledged, the chambers are
recycled and available for use by the sending process.
lsD
This is why we show a circular buffer.
The operation of the buffer at the receiver site is simpler. The circular
ria
buffer is divided into two areas (shown as white and colored). The white
area contains empty chambers to be filled by bytes received from the
to
network. The colored sections contain received bytes that can be read by
the receiving process. When a byte is read by the receiving process, the
Tu
4 TCP segments
m
.co
i ya
un
At the transport layer, TCP groups a number of bytes together into a packet
lsD
called a segment. TCP adds a header to each segment (for control
purposes) and delivers the segment to the IP layer for transmission. The
ria
segments are encapsulated in IP datagrams and transmitted.
This entire operation is transparent to the receiving process. Later we will
see that segments may be received out of order, lost, or corrupted and
to
resent. All these are handled by TCP with the receiving process unaware of
any activities. Above fig shows how segments are created from the bytes in
Tu
the buffers
23.6
5 Full-Duplex Communication
TCP offers full-duplex service, in which data can flow in both directions at
m
the same time. Each TCP then has a sending and receiving buffer, and
segments move in both directions
.co
6 Connection-Oriented Service
ya
TCP is a connection-oriented protocol. When a process at site A wants to
send and receive data from another process at site B, the following occurs:
i
1. The two TCPs establish a connection between them.
un
2. Data are exchanged in both directions.
3. The connection is terminated.
TCP Features
1 Numbering System
m
There are two fields called the sequence number and the
.co
acknowledgment number. These two fields refer to the byte number and
not the segment number.
Byte Number The bytes of data being transferred in each connection
ya
are numbered by TCP. The numbering starts with a randomly generated
number. For example, if the random number happens to be 1057 and the
i
total data to be sent are 6000 bytes, the bytes are numbered from 1057
un
to 7056. We will see that byte numbering is used for flow and error
control.
lsD
Sequence Number After the bytes have been numbered, TCP assigns a
sequence number to each segment that is being sent. The sequence
number for each segment is the number of the first byte carried in that
ria
segment.
Acknowledgment Number The value of the acknowledgment field in a
segment defines the number of the next byte a party expects to receive.
to
2 Flow Control
TCP, provides flow control. The receiver of the data controls the amount
m
of data that are to be sent by the sender. This is done to prevent the
.co
receiver from being overwhelmed with data. The numbering system
allows TCP to use a byte-oriented flow control.
ya
3 Error Control
To provide reliable service, TCP implements an error control mechanism.
i
Although error control considers a segment as the unit of data for error
un
detection (loss or corrupted segments), error control is byte-oriented, as
we will see later.
m
.co
i ya
un
lsD
ria
to
Tu
23.10
m
application program in the host that is sending the segment.
Destination port address. This is a 16-bit field that defines the port number of the
.co
application program in the host that is receiving the segment.
Sequence number. This 32-bit field defines the number assigned to the first byte
of data contained in this segment. As we said before, TCP is a stream transport
ya
protocol. To ensure connectivity, each byte to be transmitted is numbered. The
sequence number tells the destination which byte in this sequence comprises the
i
first byte in the segment. During connection establishment, each party uses a
un
random number generator to create an initial sequence number (ISN), which is
usually different in each direction.
lsD
Acknowledgment number. This 32-bit field defines the byte number that the
receiver of the segment is expecting to receive from the other party. If the
receiver of the segment has successfully received byte number x from the other
ria
party, it defines x + I as the acknowledgment number. Acknowledgment and data
can be piggybacked together.
Header length. This 4-bit field indicates the number of 4-byte words in the TCP
to
header. The length of the header can be between 20 and 60 bytes. Therefore,
the value of this field can be between 5 (5 x 4 =20) and 15 (15 x 4 =60).
Tu
m
.co
ya
These bits enable flow control, connection establishment and termination,
i
connection abortion, and the mode of data transfer in TCP.
un
Window size. This field defines the size of the window, in bytes, that the other
party must maintain. Note that the length of this field is 16 bits, which means
lsD
that the maximum size of the window is 65,535 bytes. This value is normally
referred to as the receiving window (rwnd) and is determined by the receiver.
The sender must obey the dictation of the receiver in this case.
ria
Checksum. This 16-bit field contains the checksum. The calculation of the
checksum for TCP follows the same procedure as the one described for UDP.
to
pseudoheader, serving the same purpose, is added to the segment. For the
TCP pseudoheader, the value for the protocol field is 6.
Urgent pointer. This l6-bit field, which is valid only if the urgent flag is set, is
used when the segment contains urgent data. It defines the number that
m
must be added to the sequence number to obtain the number of the last
urgent byte in the data section of the segment. This will be discussed later
.co
in this chapter.
Options. There can be up to 40 bytes of optional information in the TCP
ya
header. We will not discuss these options here; please refer to the
reference list for more information.
i
un
lsD
ria
to
Tu
m
A TCP Connection
.co
TCP is connection-oriented. A connection-oriented transport protocol
establishes a virtual path between the source and destination. All the
segments belonging to a message are then sent over this virtual path.
ya
Using a single virtual pathway for the entire message facilitates the
acknowledgment process as well as retransmission of damaged or lost
i
frames.
un
In TCP, connection-oriented transmission requires three phases:
3. connection termination.
ria
to
Tu
m
1 The client sends the first segment, a SYN segment, in which only the SYN
flag is set.
.co
NOTE:A SYN segment cannot carry data, but it consumes one
sequence number.
ya
2. The server sends the second segment, a SYN +ACK segment, with 2 flag
bits set: SYN and ACK. This segment has a dual purpose. It is a SYN
i
un
segment for communication in the other direction and serves as the
acknowledgment for the SYN segment. It consumes one sequence number.
NOTE:A SYN+ACK segment cannot carry data, but does consume
one sequence number lsD
3. The client sends the third segment. This is just an ACK segment. It
ria
acknowledges the receipt of the second segment with the ACK flag and
acknowledgment number field. Note that the sequence number in this
to
segment is the same as the one in the SYN segment; the ACK segment does
not consume any sequence numbers.
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
segments to a server, pretending that each of them is corning from a
different client by faking the source IP addresses in the datagram's.
.co
The server, assuming that the clients are issuing an active open, allocates
the necessary resources, such as creating communication tables and
ya
setting timers. The TCP server then sends the SYN +ACK segments to the
fake clients, which are lost. During this time, however, a lot of resources
i
are occupied without being used. If, during this short time, the number of
un
SYN segments is large, the server eventually runs out of resources and
may crash. This SYN flooding attack belongs to a type of security attack
lsD
known as a denial-of-service attack, in which an attacker monopolizes a
system with so many service requests that the system collapses and
denies service to every request.
ria
SOLUTIONS:
1 Some have imposed a limit on connection requests during a
specified period of
to
time.
2 Others filter out datagrams coming from unwanted source
Tu
addresses.
3 One recent strategy is to postpone resource allocation until the
entire connection is set up, using what is called a cookie.
Download FREE Computer Science Notes at TutorialsDuniya.com
Download FREE Computer Science Notes at TutorialsDuniya.com
Data Transfer
After connection is established, bidirectional data transfer can take place.
m
The client and server can both send data and acknowledgments. Data
traveling in the same direction as an acknowledgment are carried on the
.co
same segment. The acknowledgment is piggybacked with the data
ya
In this example, after connection is established (not shown in the figure),
the client sends 2000 bytes of data in two segments. The server then
i
sends 2000 bytes in one segment. The client sends one more segment.
un
The first three segments carry both data and acknowledgment, but the last
segment carries only an acknowledgment because there are no more data
to be sent.
lsD
Note the values of the sequence and acknowledgment numbers. The data
segments sent by the client have the PSH (push) flag set so that the
ria
server TCP knows to deliver data to the server process as soon as they
are received.
to
Tu
Data transfer
m
.co
i ya
un
lsD
ria
to
Tu
m
TCP can handle such a situation. The application program at the sending
site can request a push operation. This means that the sending TCP must
.co
not wait for the window to be filled. It must create a segment and send it
immediately. The sending TCP must also set the push bit (PSH) to let the
ya
receiving TCP know that the segment includes data that must be
delivered to the receiving application program as soon as possible and
not to wait for more data to come.
i
un
Urgent Data : TCP is a stream-oriented protocol. This means that the
data are presented from the application program to TCP as a stream of
lsD
bytes. Each byte of data has a position in the stream. However, sending
application program wants a piece of data to be read out of order by the
ria
receiving application program.
to
Tu
m
1. In a normal situation, the client TCP, after receiving a close command
.co
from the client process, sends the first segment, a FIN segment in which the
FIN flag is set.
ya
Note that a FIN segment can include the last chunk of data sent by the
client, or it can be just a control segment as shown in Figure. If it is only a
i
control segment, it consumes only one sequence number.
un
NOTE: The FIN segment consumes one sequence number ifit does
not carry data.
lsD
2 The server TCP, after receiving the FIN segment, informs its process of
the situation and sends the second segment, a FIN +ACK segment, to
ria
confirm the receipt of the FIN segment from the client and at the same time
to announce the closing of the connection in the other direction. This
segment can also contain the last chunk of data from the server. If it does
to
3. The client TCP sends the last segment, an ACK segment, to confirm the
receipt of the FIN segment from the TCP server. This segment contains the
m
acknowledgment number, which is 1 plus the sequence number received in
the FIN segment from the server. This segment cannot carry data and
.co
consumes no sequence numbers.
ya
Half-Close In TCP, one end can stop sending data while still receiving
data. This is called a half-close. Although either end can issue a half-close,
it is normally initiated by the client. It can occur when the server needs all
i
un
the data before processing can begin.
A good example is sorting. When the client sends data to the server to be
sorted, the server needs to receive all the data before sorting can start.
lsD
This means the client, after sending all the data, can close the connection
in the outbound direction. However, the inbound direction must remain
ria
open to receive the sorted data. The server, after receiving the data, still
needs time for sorting; its outbound direction must remain open
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
protocol used by TCP, however, is something between the Go-Back-N and
Selective Repeat sliding window.
.co
The sliding window protocol in TCP looks like the Go-Back-N protocol
ya
because it does not use NAKs;
it looks like Selective Repeat because the receiver holds the out-of-order
i
segments until the missing ones arrive.
un
There are two big differences between this sliding window and the one
lsD
we used at the data link layer.
1 the sliding window of TCP is byte-oriented; the one we discussed in the
data link layer is frame-oriented.
ria
2 the TCP's sliding window is of variable size; the one we discussed in
the data link layer was of fixed size
to
Tu
Sliding window
m
.co
i ya
un
lsD
ria
to
Tu
m
see, are in the control of the receiver (and depend on congestion in the
network), not the sender.
.co
The sender must obey the commands of the receiver in this matter.
Opening a window means moving the right wall to the right. This allows
more new bytes in the buffer that are eligible for sending.
ya
Closing the window means moving the left wall to the right. This means
that some bytes have been acknowledged and the sender need not worry
i
un
about them anymore.
Shrinking the window means moving the right wall to the left.
lsD
The size of the window at one end is determined by the lesser of two
values: receiver window (rwnd) or congestion window (cwnd).
The receiver window is the value advertised by the opposite end in a
ria
segment containing acknowledgment. It is the number of bytes the other
end can accept before its buffer overflows and data are discarded.
to
m
.co
i ya
un
lsD
ria
to
Tu
m
running on the remote machine.
.co
2) the sender may send a 1-byte segment to force the receiver to
reannounce the next byte expected and the window size. This packet is
called a window probe.
ya
The TCP standard explicitly provides this option to prevent deadlock if a
window update ever gets lost.
i
un
Senders are not required to transmit data as soon as they come in from the
application. Neither are receivers required to send acknowledgements as
soon as possible. lsD
ria
For example, in Fig. when the first 2 KB of data came in, TCP, knowing that it
had a 4-KB window, would have been completely correct in just buffering the
data until another 2 KB came in, to be able to transmit a segment with a 4-KB
to
m
3.
.co
send echo of character
and/or output 2.
interpret
character
1.
ya
send character
Host with Host with
Telnet client Telnet server
i
un
Remote terminal applications (e.g., Telnet) send characters to a server.
lsD
The server interprets the character and sends the output at the server
to the client.
ria
For each character typed, you see three packets:
Client Server: Send typed character
to
Delayed Acknowledgement
m
• TCP delays transmission of ACKs for up to 500ms
.co
• Avoid to send ACK packets that do not carry data.
– The hope is that, within the delay, the receiver will have data ready to
ya
be sent to the receiver. Then, the ACK can be piggybacked with a data
segment
i
un
Exceptions:
• ACK should be sent for every full sized segment
lsD
• Delayed ACK is not used when packets arrive out of order
31
Nagel’s Rule
m
Send one byte and buffer all subsequent bytes until acknowledgement is
received. Then send all buffered bytes in a single TCP segment and start
.co
buffering again until the sent segment is acknowledged.
Nagle’s algorithm will put the many pieces in one segment, greatly reducing
the bandwidth used
ya
Nagle’s algorithm is widely used by TCP implementations, but there are
i
un
times when it is better to disable it. In particular, in interactive games that are
run over the Internet.
A more subtle problem is that Nagle’s algorithm can sometimes interact with
lsD
delayed acknowledgements to cause a temporary deadlock: the receiver
waits for data on which to piggyback an acknowledgement, and the sender
waits on the acknowledgement to send more data.
ria
Another problem that can degrade TCP performance is the silly window
syndrome (Clark, 1982).
m
.co
i ya
un
lsD
ria
to
Tu
Clark’s solution is to prevent the receiver from sending a window update for
1 byte. Instead, it is forced to wait until it has a decent amount of space
m
available and advertise that instead. Specifically, the receiver should not
.co
send a window update until it can handle the maximum segment size it
advertised when the connection was established or until its buffer is half
empty, whichever is smaller.
ya
Furthermore, the sender can also help by not sending tiny segments.
i
Instead, it should wait until it can send a full segment, or at least one
un
containing half of the receiver’s buffer size.
lsD
The goal is for the sender not to send small segments and the receiver not
to ask for them. (Nagel + Clark). Both are used to improve TCP
performance
ria
The receiver will buffer the data until it can be passed up to the
application in order (handling out of order segments)
to
Tu
Cumulative acknowledgements
Error Control
TCP is a reliable transport layer protocol. This means that an application
m
program that delivers a stream of data to TCP relies on TCP to deliver the
entire stream to the application program on the other end in order, without
.co
error, and without any part lost or duplicated.
ya
TCP provides reliability using error control. Error control includes
mechanisms for detecting corrupted segments, lost segments, out-of-order
segments, and duplicated segments. Error control also includes a
i
un
mechanism for correcting errors after they are detected. Error detection and
correction in TCP is achieved through the use of three simple tools:
checksum, acknowledgment, and time-out.
Checksum
lsD
ria
Each segment includes a checksum field which is used to check for a
corrupted segment. If the segment is corrupted, it is discarded by the
destination TCP and is considered as lost. TCP uses a 16-bit checksum that
to
m
.co
i ya
un
lsD
ria
to
Tu
23.36
Acknowledgment
TCP uses acknowledgments to confirm the receipt of data segments.
m
Control segments that carry no data but consume a sequence number are
also acknowledged. ACK segments are never acknowledged.
.co
ACK segments do not consume sequence numbers and are not
acknowledged.
ya
Retransmission
i
The heart of the error control mechanism is the retransmission of
un
segments. When a segment is corrupted, lost, or delayed, it is
retransmitted.
lsD
In modern implementations, a retransmission occurs if the retransmission
timer expires or three duplicate ACK segments have arrived.
ria
Retransmission After RTO (retransmission time out)
Retransmission After Three Duplicate ACK Segments (also called fast
retransmission)
to
Out-of-Order Segments
Tu
Data may arrive out of order and be temporarily stored by the receiving
TCP, but yet guarantees that no out-of-order segment is delivered to the
process
Download FREE Computer Science Notes at TutorialsDuniya.com
Download FREE Computer Science Notes at TutorialsDuniya.com
m
When the load offered to any network is more than it can handle,
congestion builds up.
.co
The network layer detects congestion when queues grow large at routers
and tries to manage it, if only by dropping packets. It is up to the transport
ya
layer to receive congestion feedback from the network layer and slow down
the rate of traffic that it is sending into the network.
i
un
For Congestion control, transport protocol uses an AIMD (Additive Increase
Multiplicative Decrease) control law.
lsD
TCP congestion control is based on implementing this approach using a
window called congestion window. TCP adjusts the size of the window
ria
according to the AIMD rule.
where
flow control window is advertised by the receiver (rwnd)
congestion window is adjusted based on feedback from the
Download FREE Computer Science Notes at TutorialsDuniya.com
Download FREE Computer Science Notes at TutorialsDuniya.com
Modern congestion control was added to TCP largely through the efforts of
Van Jacobson (1988). It is a fascinating story. Starting in 1986, the growing
m
popularity of the early Internet led to the first occurrence of what became
known as a congestion collapse, a prolonged period during which good
.co
put dropped suddenly (i.e., by more than a factor of 100) due to congestion
in the network. Jacobson (and many others) set out to understand what
ya
was happening and remedy the situation.
i
To start, he observed that packet loss is a suitable signal of congestion.
un
This signal comes a little late (as the network is already congested) but it is
quite dependable
lsD
At the beginning how sender knows at what speed receiver can receive the
packets?
ria
to
Tu
m
.co
i ya
un
lsD
The key observation is this: the acknowledgements return to the sender
at about the rate that packets can be sent over the slowest link in the
path. This is precisely the rate that the sender wants to use. If it injects
ria
new packets into the network at this rate, they will be sent as fast as the
slow link permits, but they will not queue up and congest any router
along the path. This timing is known as an ack clock. It is an essential
to
part of TCP. By using an ack clock, TCP smoothes out traffic and avoids
unnecessary queues at routers. This is first consideration
Tu
A second consideration is that the AIMD rule will take a very long time to
reach a good operating point on fast networks if the congestion window is
m
started from a small size
.co
Instead, the solution Jacobson chose to handle both of these
considerations is a mix of linear and multiplicative increase.
ya
SLOW-START
i
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
process is restarted.
.co
Congestion avoidance phase is started if cwnd has reached the slow
start threshold value
ya
Whenever the slow start threshold is crossed, TCP switches from slow
start to additive increase. In this mode, the congestion window is
i
increased by one segment every round-trip time.
un
lsD
ria
to
Tu
43
m
.co
i ya
un
lsD
ria
to
Tu
Responses to Congestion
m
.co
• So, TCP assumes there is congestion if it
detects a packet loss
ya
• A TCP sender can detect lost packets via:
• Timeout of a retransmission timer
i
un
• Receipt of a duplicate ACK
lsD
• TCP interprets a Timeout as a binary congestion signal. When a
timeout occurs, the sender performs:
ria
– cwnd is reset to one:
cwnd = 1
– ssthresh is set to half the current size of the congestion window:
to
ssthressh = cwnd / 2
– and slow-start is entered
Tu
45
Fast Retransmit
m
.co
• If three or more duplicate
ACKs are received in a
ya
row, the TCP sender
believes that a segment
has been lost.
i
un
• Then TCP performs a
lsD
retransmission of what
seems to be the missing
segment, without waiting
ria
for a timeout to happen.
to
cwnd = 1
46
m
.co
• TCP Tahoe (1988)
ya
– Slow Start
– Congestion Avoidance
i
– Fast Retransmit
un
• TCP Reno (1990) (TCP Tahoe+FR)
– Fast Recovery
• New Reno (1996) lsD
• SACK (1996) (SACK (Selective
ria
ACKnowledgements))
to
47
m
.co
i ya
un
lsD
ria
to
Tu
m
as a congestion signal. ECN is an IP layer mechanism to notify hosts of
congestion.
.co
The sender tells the receiver that it has heard the signal by using the CWR
(Congestion Window Reduced) flag.
i ya
un
lsD
ria
to
Tu
USER DATAGRAM PROTOCOL (UDP)
m
.co
The User Datagram Protocol (UDP) is called a
ya
connectionless, unreliable transport protocol. It does
not add anything to the services of IP except to provide
i
un
process-to-process communication instead of host-to-
host communication.
lsD
Topics discussed in this section:
ria
Well-Known Ports for UDP
User Datagram
Checksum
to
UDP Operation
Tu
Use of UDP
m
.co
i ya
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
23.52
m
The UDP checksum calculation is different from the one for IP and ICMP. Here
the checksum includes three sections: a pseudo header, the UDP header,
.co
and the data coming from the application layer.
The pseudo header is the part of the header of the IP packet in which the user
ya
datagram is to be encapsulated with some fields filled with Os
If the checksum does not include the pseudo header, a user datagram may
i
un
arrive safe and sound. However, if the IP header is corrupted, it may be
delivered to the wrong host.
The protocol field is added to ensure that the packet belongs to UDP, and not to
lsD
other transport-layer protocols.
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
23.54
UDP Operation
Connectionless Services
m
UDP provides a connectionless service. This means that each user
datagram sent by UDP is an independent datagram. There is no
.co
relationship between the different user datagrams even if they are coming
from the same source process and going to the same destination program.
The user datagrams are not numbered. Also, there is no connection
ya
establishment and no connection termination, as is the case for TCP. This
means that each user datagram can travel on a different path.
i
Flow and Error Control
un
UDP is a very simple, unreliable transport protocol. There is no flow control
and hence no window mechanism. The receiver may overflow with
lsD
incoming messages. There is no error control mechanism in UDP except for
the checksum. This means that the sender does not know if a message has
been lost or duplicated. When the receiver detects an error through the
checksum, the user datagram is silently discarded. The lack of flow control
ria
and error control
Encapsulation and Decapsulation
To send a message from one process to another, the UDP protocol
to
m
Figure 23.11 shows the checksum calculation for a very
.co
small user datagram with only 7 bytes of data. Because the
ya
number of bytes of data is odd, padding is added for
checksum calculation. The pseudoheader as well as the
i
un
padding will be dropped when the user datagram is
delivered to IP.
lsD
ria
to
Tu
23.56
m
.co
i ya
un
lsD
ria
to
Tu
23.57
m
.co
iya
un
lsD
ria
to
Tu
23.58
m
Birrell and Nelson suggested was allowing programs to call procedures
located on remote hosts. When a process on machine 1 calls a procedure
.co
on machine 2, the calling process on 1 is suspended and execution of the
called procedure takes place on 2. Information can be transported from the
caller to the callee in the parameters and can come back in the procedure
ya
result. No message passing is visible to the application programmer. This
technique is known as RPC (Remote Procedure Call). Traditionally, the
i
un
calling procedure is known as the client and the called procedure is known
as the server, and we will use those names here too.
lsD
to call a remote procedure, the client program must be bound with a small
library procedure, called the client stub, that represents the server
procedure in the client’s address space. Similarly, the server is bound with a
ria
procedure called the server stub. These procedures hide the fact that the
procedure call from the client to the server is not local
to
Tu
Step 1 is the client calling the client stub. This call is a local procedure call,
with the parameters pushed onto the stack in the normal way.
m
Step 2 is the client stub packing the parameters into a message and
making a system call to send the message. Packing the parameters is
.co
called marshaling.
Step 3 is the operating system sending the message from the client
ya
machine to the server machine.
Step 4 is the operating system passing the incoming packet to the server
i
stub.
un
Finally, step 5 is the server stub calling the server procedure with the
unmarshaled parameters.
lsD
The reply traces the same path in the other direction.
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
1 With RPC, passing pointers is impossible because the client and server are in
different address spaces.
.co
2 It is essentially impossible for the client stub to marshal the parameters: it
has no way of determining how large they are.
ya
3 A third problem is that it is not always possible to deduce the types of the
parameters, not even from a formal specification or the code itself.(exa:
i
printf)
un
4 A fourth problem relates to the use of global variables. Normally, the calling
and called procedure can communicate by using global variables, in addition to
lsD
communicating via parameters. But if the called procedure is moved to a
remote machine, the code will fail because the global variables are no longer
shared
ria
to
Tu
m
Another one is for real-time multimedia applications.
Internet radio,
.co
Internet telephony,
music-on-demand,
ya
videoconferencing,
video-on-demand,
i
and other multimedia applications became more commonplace, people
un
have discovered that each application was reinventing more or less the
same real-time transport protocol.
lsD
It gradually became clear that having a generic real-time transport protocol
for multiple applications would be a good idea.
ria
Thus was RTP (Real-time Transport Protocol) born. It is described in
RFC 3550 and is now in widespread use for multimedia applications. We
will describe two aspects of real-time transport.
to
The first is the RTP protocol for transporting audio and video data in
packets.
Tu
The second is the processing that takes place, mostly at the receiver, to
play out the audio and video at the right time..
Download FREE Computer Science Notes at TutorialsDuniya.com
Download FREE Computer Science Notes at TutorialsDuniya.com
m
.co
i ya
un
lsD
ria
to
Tu
RTP normally runs in user space over UDP (in the operating system).
It operates as follows. The multimedia application consists of multiple
m
audio, video, text, and possibly other streams. These are fed into the RTP
library, which is in user space along with the application. This library
.co
multiplexes the streams and encodes them in RTP packets, which it stuffs
into a socket.
ya
On the operating system side of the socket, UDP packets are generated to
wrap the RTP packets and handed to IP for transmission over a link such
i
as Ethernet.
un
The reverse process happens at the receiver. The multimedia application
eventually receives multimedia data from the RTP library. It is responsible
lsD
for playing out the media. The protocol stack for this situation is shown in
Fig. 6-30(a). The packet nesting is shown in Fig. 6-30(b).
ria
to
Tu
m
onto
a single stream of UDP packets. The UDP stream can be sent to a single
.co
destination (unicasting) or to multiple destinations (multicasting).
Because RTP just uses normal UDP, its packets are not treated specially by
the routers unless some normal IP quality-of-service features are enabled.
ya
In particular, there are no special guarantees about delivery, and packets
may be lost, delayed, corrupted, etc.
i
un
The RTP format contains several features.
Each packet sent in an RTP stream is given a number one higher than its
lsD
predecessor. This numbering allows the destination to determine if any
packets are missing.
RTP has no acknowledgements, and no mechanism to request
ria
retransmissions.
Each RTP payload may contain multiple samples, and they may be coded
to
any way that the application wants. To allow for interworking, RTP defines
several profiles (e.g., a single audio stream), and for each profile, multiple
Tu
m
.co
i ya
un
lsD
ria
to
Tu
m
bytes. The last padding byte tells how many bytes were added.
.co
The X bit indicates that an extension header is present.
The CC field tells how many contributing sources are present, from 0
to 15
ya
The M bit is an application-specific marker bit. It can be used to mark
the start of a video frame, the start of a word in an audio channel, or
i
something else that the application understands.
un
The Payload type field tells which encoding algorithm has been used
(e.g., uncompressed 8-bit audio, MP3, etc.). Since every packet
lsD
carries this field, the encoding can change during transmission.
The Sequence number is just a counter that is incremented on each
RTP packet sent. It is used to detect lost packets.
ria
The Timestamp, this value can help reduce timing variability called
jitter at the receiver by decoupling the playback from the packet
arrival time.
to
m
Transport Control Protocol). It is defined along with RTP in RFC 3550 and
handles feedback, synchronization, and the user interface. It does not
.co
transport any media samples.
The first function can be used to provide feedback on delay, variation in delay
ya
or jitter, bandwidth, congestion, and other network properties to the sources.
This information can be used by the encoding process to increase the data
rate (and give better quality) when the network is functioning well and to cut
i
un
back the data rate when there is trouble in the network. By providing
continuous feedback, It provides the best quality
The Payload type field is used to tell the destination what encoding algorithm
lsD
is used for the current packet, making it possible to vary it on demand.
RTCP also handles inter stream synchronization. The problem is that
ria
different streams may use different clocks, with different
granularities and different drift rates. RTCP can be used to keep
them in sync.
to
Finally, RTCP provides a way for naming the various sources (e.g., in
ASCII text). This information can be displayed on the receiver’s
Tu
m
between them at the sender, they will reach the receiver with different
.co
relative times. This variation in delay is called jitter. Even a small amount of
packet jitter can cause distracting media artifacts, such as jerky video
frames and unintelligible audio, if the media is simply played out as it
ya
arrives.
The solution to this problem is to buffer packets at the receiver before they
i
are played out to reduce the jitter.
un
lsD
ria
to
Tu
m
Deciding how long to wait depends on the jitter. The difference
between a low-jitter and high-jitter connection is shown in Fig. The
.co
average delay may not differ greatly between the two, but if there is
high jitter the playback point may need to be much further out to
ya
capture 99% of the packets than if there is low jitter.
i
un
lsD
ria
to
One way to avoid this problem for audio is to adapt the playback
Tu
TELNET
It is client/server application program. TELNET is an abbreviation for
m
TErminaL NETwork. TELNET enables the establishment of a connection to a
remote system in such a way that the local terminal appears to be a terminal
.co
at the remote system.
ya
Timesharing Environment
A large computer supports multiple users. The interaction between a
user and the computer occurs through a terminal, which is usually a
i
un
combination of keyboard, monitor, and mouse.
Logging
lsD
To access the system, the user logs into the system with a user id or
log-in name. The system also includes password checking to prevent an
unauthorized user from accessing the resources.
ria
Local login
Remote login
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
When a user logs into a local timesharing system, it is called local log-in.
As a user types at a terminal or at a workstation running a terminal
m
emulator, the keystrokes are accepted by the terminal driver. The terminal
driver passes the characters to the operating system. The operating
.co
system, in turn, interprets the combination of characters and invokes the
desired application program or utility.
ya
When a user wants to access an application program or utility located on a
remote Machine, it is called remote log-in. Here the TELNET client and
i
server programs come into use. The user sends the keystrokes to the
un
terminal driver, where the local operating system accepts the characters
but does not interpret them. The characters are sent to the TELNET client,
which transforms the characters to a universal character set called
lsD
network virtual terminal (NVT) characters and delivers them to the local
TCP/IP protocol stack.
ria
The commands or text, in NVT form, travel through the Internet and arrive
at the TCP/IP stack at the remote machine. Here the characters are
delivered to the operating system and passed to the TELNET server, which
to
m
Concept of NVT(network virtual terminal)
.co
i ya
un
lsD
ria
to
Tu
26.75
m
.co
i ya
un
WWW and HTTP
lsD
ria
to
Tu
27.76
Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display.
271 ARCHITECTURE
m
.co
The WWW today is a distributed client/server service, in
ya
which a client using a browser can access a service
using a server. However, the service provided is
i
un
distributed over many locations called sites as shown in
fig.
lsD
Topics discussed in this section:
ria
Client (Browser)
Server
Uniform Resource Locator
to
Cookies
Tu
27.77
m
Figure 27.1 Architecture of WWW
.co
i ya
un
lsD
ria
to
Tu
27.78
Client (Browser)
A variety of vendors offer commercial browsers that interpret and
m
display a Web document,and all use nearly the same
.co
architecture.
Each browser usually consists of three parts: a controller, client
protocol, and interpreters.
ya
The controller receives input from the keyboard or the mouse and
uses the client programs to access the document.
i
After the document has been accessed, the controller uses one of
un
the interpreters to display the document on the screen. The
interpreter can be HTML, Java, or JavaScript, depending on the
type of document
lsD
The client protocol can be one of the protocols described
previously such as FTP or HTTP.
ria
Server
The Web page is stored at the server. Each time a client request
arrives, the corresponding document is sent to the client. To
to
m
Figure 27.2 Browser
.co
i ya
un
lsD
ria
to
Tu
27.80
m
facilitate the access of documents distributed throughout the world,
.co
HTTP uses locators. The uniform resource locator (URL) is a standard
for specifying any kind of information on the Internet. The URL
defines four things: protocol, host computer, port, and path.
ya
The protocol is the client/server program used to retrieve the
document. Many different protocols can retrieve a document; among
i
them are FTP or HTTP. The most common today is HTTP.
un
The host is the computer on which the information is located,
although the name of the computer can be an alias. Web pages are
lsD
usually stored in computers, and computers are given alias names
that usually begin with the characters "www".
The URL can optionally contain the port number of the server. If the
ria
port is included, it is inserted between the host and the path, and it
is separated from the host by a colon.
Path is the pathname of the file where the information is located.
to
Note that the path can itself contain slashes that, in the UNIX
Tu
m
Figure 27.3 URL
.co
i ya
un
An HTTP cookie (also called web cookie, Internetcookie,
lsD
browser cookie or simply cookie, the latter which is not to be
confused with the literal definition), is a small piece of data sent
from a website and stored in a user's web browser while the user is
ria
browsing that website
to
Tu
27.82
272 WEB DOCUMENTS
m
.co
The documents in the WWW can be grouped into three
broad categories: static, dynamic, and active. The
ya
category is based on the time at which the contents of
i
un
the document are determined.
lsD
Topics discussed in this section:
ria
Static Documents
Dynamic Documents
to
Active Documents
Tu
27.83
Static Documents
Static documents are fixed-content documents that are created
m
and stored in a server. The client can get only a copy of the
document. When a client accesses the document, a copy of the
.co
document is sent. The user can then use a browsing program to
display the document
i ya
un
lsD
ria
to
Tu
27.84
m
Figure 27.5 Boldface tags
.co
HTML
Hypertext Markup Language (HTML) is a language for creating
ya
Web pages.
i
un
lsD
ria
to
Tu
27.85
m
Figure 27.7 Beginning and ending tags
.co
i ya
un
lsD
ria
to
Tu
27.86
m
server runs an application program or a script that creates the
.co
dynamic document. The server returns the output of the program
or script as a response to the browser that requested the
document.
ya
A very simple example of a dynamic document is the retrieval of
the time and date from a server. Time and date are kinds of
i
information that are dynamic in that they change from moment to
un
moment. The client can ask the server to run a program such as
the date program in UNIX and send the result of the program to
the client.
lsD
Common Gateway Interface (CGI)
The Common Gateway Interface (CGI) is a technology that
ria
creates and handles dynamic documents.
Hypertext Preprocessor (pHP), which uses the Perl language; Java
Server Pages (JSP), which uses the Java language for scripting;
to
m
Figure 27.8 Dynamic document using CGI
.co
i ya
un
lsD
ria
to
Tu
27.88
m
Figure 27.9 Dynamic document using server-site script
.co
i ya
un
lsD
ria
to
Tu
27.89
m
.co
ya
Note
i
Dynamic documents are sometimes referred to as server-site
un
dynamic documents.
lsD
ria
to
Tu
27.90
m
Figure 27.10 Active document using Java applet
.co
Active Documents
For many applications, we need a program or a script to be run
ya
at the client site. These are called active documents
i
un
lsD
ria
to
Tu
27.91
m
Figure 27.11 Active document using client-site script
.co
i ya
un
lsD
ria
to
Tu
27.92
m
.co
ya
Note
i
Active documents are sometimes referred to as client-site dynamic
un
documents.
lsD
ria
to
Tu
27.93
m
.co
ya
Note
i
HTTP version 1.1 specifies a persistent connection by default.
un
lsD
ria
to
Tu
27.94
m
identifies the connection of a host to the Internet. However, people prefer to
use names instead of numeric addresses. Therefore, we need a system that
.co
can map a name to an address or an address to a name.
i ya
un
lsD
ria
to
Tu
NAME SPACE
A name space that maps each address to a unique name can be organized
m
in two ways: fiat or hierarchical.
Flat Name Space
.co
In a flat name space, a name is assigned to an address. A name in this
space is a sequence of characters without structure.
ya
Hierarchical Name Space
In a hierarchical name space, each name is made of several parts. The first
i
un
part can define the nature of the organization, the second part can define
the name of an organization, the third part can define departments in the
organization, and so on.
Exa:
lsD
challenger.jhda.edu, challenger.berkeley.edu, and
ria
challenger.smart.com
to
Tu
m
this design the names are defined in an inverted-tree structure with the root
at the top. The tree can have only 128 levels: level 0 (root) to level 127.
.co
ya
Domain name
space
i
un
lsD
ria
to
Tu
Label
Each node in the tree has a label, which is a string with a maximum of 63
m
characters. The root label is a null string (empty string). DNS requires that
children of a node (nodes that branch from the same node) have different
.co
labels, which guarantees the uniqueness of the domain names.
Domain Name
ya
Each node in the tree has a domain name. A full domain name is a
sequence of labels separated by dots (.). The domain names are always
read from the node up to the root. The last label is the label of the root (null).
i
un
This means that a full domain name always ends in a null label, which
means the last character is a dot because the null string is nothing. Below
Figure shows some domain names
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
Domain
A domain is a subtree of the domain name space. The name of the domain
m
is the domain name of the node at the top of the subtree.
.co
i ya
un
lsD
ria
to
Tu
m
However, it is very inefficient and also unreliable to have just one computer
store such a huge amount of information. In this section, we discuss the
.co
distribution of the domain name space
ya
1 Hierarchy of Name Servers
distribute the information among many computers called DNS
i
servers. we let the root stand alone and create as many domains
un
(subtrees) as there are first-level nodes
lsD
ria
to
Tu
2 Zone
Since the complete domain name hierarchy cannot be stored on a single
m
server, it is divided among many servers. What a server is responsible for or
has authority over is called a zone. We can define a zone as a contiguous
.co
part of the entire tree
i ya
un
lsD
ria
to
Tu
3 Root Server
A root server is a server whose zone consists of the whole tree. A root
m
server usually does not store any information about domains but delegates
its authority to other servers, keeping references to those servers. There are
.co
several root servers, each covering the whole domain name space. The
servers are distributed all around the world.
ya
4 Primary and Secondary Servers
i
A primary server is a server that stores a file about the zone for which it is
un
an authority. It is responsible for creating, maintaining, and updating the
zone file. It stores the zone file on a local disk
lsD
A secondary server is a server that transfers the complete information about
a zone from another server (primary or secondary) and stores the file on its
ria
local disk. The secondary server neither creates nor updates the zone files
to
Tu
m
domain name space (tree) is divided into three different sections: generic
domains, country domains, and the inverse domain
.co
i ya
un
lsD
ria
to
Tu
1 Generic Domains
The generic domains define registered hosts according to their generic
m
behavior. Each node in the tree defines a domain, which is an index to the
domain name space database
.co
i ya
un
lsD
ria
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
2 Country Domains
The country domains section uses two-character country abbreviations
m
(e.g., us for United States). Second labels can be organizational, or they
can be more specific, national designations. The United States, for example,
.co
uses state abbreviations as a subdivision of us (e.g., ca.us.).
i ya
un
lsD
ria
to
Tu
3 Inverse Domain
The inverse domain is used to map an address to a name.
m
.co
i ya
un
lsD
ria
to
Tu
RESOLUTION
Mapping a name to an address or an address to a name is called name-
m
address resolution
1 Resolver
.co
DNS is designed as a client/server application. A host that needs to
map an address to a name or a name to an address calls a DNS
ya
client called a resolver. The resolver accesses the closest DNS
server with a mapping request. If the server has the information, it
satisfies the resolver; otherwise, it either refers the resolver to
i
un
other servers or asks other servers to provide the information.
2 Mapping Names to Addresses
In this case, the server checks the generic domains or the country
lsD
domains to find the mapping.
3 Mapping Addresses to Names
ria
To answer queries of this kind, DNS uses the inverse domain
4 Recursive Resolution
The client (resolver) can ask for a recursive answer from a name
to
server. This means that the resolver expects the server to supply
the final answer.. When the query is finally resolved, the response
Tu
m
.co
Recursive resolution
i ya
un
lsD
ria
to
Tu
5 Iterative Resolution
If the client does not ask for a recursive answer, the mapping can
m
be done iteratively. If the server is an authority for the name, it
sends the answer. If it is not, it returns (to the client) the IP
.co
address of the server that it thinks can resolve the query
i ya
un
lsD
ria
to
Tu
6 Caching
Each time a server receives a query for a name that is not in its
m
domain, it needs to search its database for a server IP address.
Reduction of this search time would increase efficiency. DNS
.co
handles this with a mechanism called caching
ya
DNS MESSAGES
DNS has two types of messages: query and response. Both types have the
i
un
same format.
The query message consists of a header,
and question records;
lsD
the response message consists of a header,
question records,
answer records,
ria
authoritative records,
and additional records
to
Tu
m
.co
i ya
un
lsD
ria
to
Tu
Header
Both query and response messages have the same header format with some fields
m
set to zero for the query messages. The header is 12 bytes,
.co
i ya
un
lsD
ria
to
Tu
TYPES OF RECORDS
The question records are used in the question section of the query and
m
response messages. The resource records are used in the answer,
authoritative, and additional information sections of the response
.co
message.
ya
Question Record
A question record is used by the client to get information from a server..
Resource Record
i
un
Each domain name (each node on the tree) is associated with a record
called the resource record. The server database consists of resource
records. Resource records are also what is returned by the server to the
client. lsD
ria
REGISTRARS
How are new domains added to DNS? This is done through a registrar, a
commercial entity accredited by ICANN. A registrar first verifies that the
to
requested domain name is unique and then enters it into the DNS
database. A fee is charged. Today, there are many registrars; their names
Tu
m
host, or changing an IP address, the change must be made to the DNS
.co
master file. The size of today's Internet does not allow for this kind of
manual operation.
The DNS master file must be updated dynamically. The Dynamic Domain
ya
Name System (DDNS) therefore was devised to respond to this need.
i
un
ENCAPSULATION
DNS can use either UDP or TCP. In both cases the well-known port used by
the server is port 53.
lsD
ria
to
Tu