UNIT-4

IP Datagrams and Forwarding Connectionless

service Virtual Packets IP Datagram Forwarding
IP Address and Routing Table Entry Mask Field and
Datagram Forwarding IP Datagram Header Format.
IP Encapsulation and Fragmentation Datagram
Transmission and Frames, Encapsulation, Transmission
across an Internet, MTU, Datagram size and
Encapsulation Reassembly Fragment Loss.
TCP Reliable Transport service Need for Reliable
Transport TCP- Services provided End to end
service, Packet loss and Retransmission,
Retransmission Times, Buffers, Flow control and
Windows, Three-Way Handshake, Congestion control,
TCP Segment Format.

IP Datagrams and Forwarding

Connectionless services :
Initially the designers must decide the type of the service.
Two types of connection service provided are: (1)
Connection-Oriented service. (2) Connectionless service.
Connectionless services do not require a session
connection between the Sender and the Receiver. In this type
of service, the Application programs are unaware of the
underlying physical hardware. The sender simply starts

simply sending packets(called Datagrams) to the destination.

There is no acknowledgement confirmation from the receiver.
So the Connectionless service always provides the Unreliable
communication. A Connectionless service provides minimal
services.
In contrast, connectionless service is modeled after the
postal system. Each message (letter) carries the full
destination address, and each one is routed through the system
independent of all the others. Normally, when two messages
are sent to the same destination, the first one sent will be the
first one to arrive. However, it is possible that the first one
sent can be delayed so that the second one arrives first.

Virtual Packets :
To overcome the heterogeneity, internet protocol software
defines an internet packet format that is independent of the
underlying hardware. The result is universal, Virtual packet
that can be transferred across the underlying hardware. As the
term virtual implies, protocol software creates and handles
internet packets-the underlying hardware does not understand
or recognize the internet packet format. As the term universal
implies, each host or router in an internet contains protocol
software that understands internet packets.
TCP/IP protocols use the name IP datagram to refer to an
internet packet. An IP datagram has the same general format
as a hardware frame; the datagram begins with a header
following by a data area.

The Header contains information that controls where and

how the datagram is to be sent. The amount of data carried in
a datagram is fixed. The sender chooses an amount of data
that is appropriate to a particular packet. The size of a
datagram is determined by the application that sends data.
Example: An application that transmits keystrokes across a
network can place each keystroke in a separate datagram,
while an application that transfers large files can send large
datagrams.

IP Datagram forwarding :
Datagrams traverse an internet by following a path from
their initial source through routers to the final destination.
Each router along the path receives the datagram, extracts the
destination address from the header, and uses the destination
address to determine a next hop to which datagram should be
sent. The router then forwards the datagram to the next hop,
either the final destination or another router.
To make the selection of a next hop efficient, each IP
router keeps the information in a routing table. A routing table
must be initialized, when the router boots, and must be
updated if the topology changes/hardware fails.
Conceptually, a routing table contains a set of entries that
each specify a destination and the next hop used to reach that
destination.
Example:

Destination
Net 1
Net 2
Net 3
Net 4

Next hop
R1
deliver direct
deliver direct
R3

In the above diagram, router R2 directly connects the

networks: Net 1 and Net 2. Therefore, R2 can directly deliver
a datagram to any destination attached to those networks.
When a datagram is destined for Net 4, R2 sends the datagram
to router R3.

IP address and Routing Table Entry :

In practice, IP routing table consists of 3 fields. First, the
destination field in each entry contains the network prefix of
the destination network. Second, an additional field in each
entry contains an address mask that specifies which bits of the
destination correspond to the network prefix. Third, an IP
address is used when the next hop fields denotes a router.

Destination

Mask

Next hop

30.0.0.0
40.0.0.0
50.8.0.0
60.6.10.0

255.0.0.0
255.0.0.0
255.255.0.0
255.255.255.0

40.0.0.7
deliver dierct
deliver dierct
50.8.0.9

Mask Field and Datagram forwarding :

The process of using a routing table to select a next hop
for a given datagram is called routing or forwarding. The
mask field in a routing table entry is used to extract the
network part of an addressing during lookup.
Imagine that a routing software is given a datagram to
forward. Also assume that the datagram contains a destination
IP address D. The routing software must find an entry in the
routing table that specifies a next hop for D.
To do so, the software examines each entry in the table by
using the mask in the entry to extract a prefix of address D
and comparing the result to the destination field of the entry.
If the two are equal, the datagram will be forwarded to the
Next hop in the entry. A bit mask makes prefix extraction
efficient. Software computes the Boolean AND of the mask
and the datagram destination address D.
If((Mask[i] & D)==Destination[i])
forward to nextHop[i];

IP Datagram Header Format

VERS : Version - 4-bit protocol version(ipv4 or ipv6).

H.LEN : Header Length - Number of 32 bit entities in
the header.
SERVICE TYPE : Minimum delay or maximum
throughput.

 TOTAL LENGTH : Total number of octets in the

header and data.
IDENTIFICATION : : After fragmentation each
fragment is given a number).
FLAGS : Indicates datagram is a fragment.
FRAGMENT OFFSET : How to order the fragments.
TIME TO LIVE : Positive integer from 1 to 255.
Counter decrements from the value when the datagram
begins to send. When counter reaches 0, datagram is
discarded and error message will be sent to the source.
HEADER CHECKSUM : bits of the header are not
changed in transmit. Sender computes 1s complement
sum of all 16 bit entities in the header excluding header
checksum field. Receiver computes the same 16 bit
including header checksum. Checksum is correct if result
is zero.
SOURCE IP ADDRESS : Logical address of the
Source.
DESTINATION IP ADDRESS : Logical address of the
Destination
PADDING : That contains zero bits to make header
multiple of 32 bits.

Data Transmission and Frames,

Encapsulation :

When a host or router handles a datagram, IP software

first select the next hop to which the datagram should be sent,
N, and then transmits the datagram across a physical network
to N. Network hardware does not understand the datagram
format or the Internet addressing. Instead, each hardware
technology defines a frame format and a physical addressing
scheme; the hardware only accepts and delivers packets that
adhere to the specified frame format and use the specified
hardware address.
Encapsulation is a technique used to transmit a datagram
across a physical network that does not understand the
datagram format. When an IP datagram is encapsulated in a
frame, the entire datagram is placed in the data area of a
frame. The network treats a frame that contains a datagram
exactly like any other frame. In fact, the hardware does not
examine/change the contents of the frame data area.
When datagram is placed in a frame, the sender assigns
the frame type field, the value that is reserved for IP. When
the frame arrives with the special value in its type field, the
receiver knows that the data area contains an IP datagram.
A frame that carries an IP datagram must have a
destination address as usual. Therefore, in addition to placing
a datagram in the data area of a frame, encapsulation requires
the sender to supply the physical address of the next computer
to which the datagram should be sent.

An IP datagram is encapsulated in a Frame

Transmission across an Internet

Encapsulation applies to one transmission at a time. After
the sender selects a next hop, the sender encapsulates the
datagram in a frame and transmits the result across the
physical network to the next hop. When the frame reaches the
next hop, the receiving software removes the IP datagram and
discards the frame. If the datagram has to be forwarded across
another network, a new frame is created. Each network can
use a different hardware technology than the others, meaning
that the frame formats can differ.
In the diagram given below, hosts and routers store a
datagram in memory with no additional header. When the
datagram passes across a physical network, the datagram is
encapsulated in a frame suitable for the network. The size of
the frame header that appears before the datagram depends on
the network technology. When a datagram arrives in a

network frame, the receiver extracts the datagram from the

frame data area and discards the frame header.

An IP datagram as it appears at each step during a trip across an internet.

MTU, Datagram size and encapsulation :

Each hardware technology specifies the maximum amount
of data that a frame can carry. The limit is known as
maximum transmission unit (MTU). There is no exception to
the MTU limit. The network hardware is not designed to
accept or transfer frames that carry more data than the MTU
allows. The datagram must be smaller or equal to the network
MTU or it cannot be encapsulated for transmission. An IP
router uses a technique known as fragmentation to solve the
problem of heterogeneous MTUs. When a datagram is larger

than the MTU of the network over which it must be sent, the
router divides the datagram into smaller pieces called
fragments, and sends each fragment independently. Fragment
has same format as other datagrams.

MTU

A bit in the FLAGS field in the header indicates whether a

datagram is a fragment or a complete datagram. Fragment
offset field in the header of a fragment specifies where in the
original datagram the fragment belongs.
To fragment a datagram, a router uses the network MTU
and the datagram header size to calculate the maximum
amount of data that can be sent in each fragment and the
number of fragments that will be needed. The router then
creates the fragments. It begins by starting each fragment with
a copy of the original header, and then modifies individual
header fields.

Reassembly and Fragment loss :

The process of creating the copy of the original datagram
from fragments is called Reassembly. Because each fragments
begins with a copy of the original datagram header, all
fragments have the same destination address as the original
datagram from which they are derived. Furthermore, the

fragment that carries the final piece of data has an additional

bit sent in the header. Thus, a receiver performing reassembly
can tell whether all fragments have arrives successfully. IP
does not guarantee datagram delivery. If an underlying
network drops packets, an encapsulated datagram or fragment
can be lost.
When all fragments from a datagram arrive, the datagram
can be reassembled. However, a problem arises when one or
more fragments from a datagram arrive, and some fragments
are delayed or lost. A receiver can not hold fragments
arbitrarily long time because fragments occupy space in the
receivers memory. When the first fragment arrives from a
given datagram, the receiver starts a timer. If all fragments of
a datagram arrive before the timer expires, the receiver
cancels the timer and reassembles the datagram. If the timer
expires before all fragments arrive, the receiver discards those
fragments that have arrived. Although the datagram cannot be
reassembled, the receiver must save the fragments in case
missing fragments are only delayed.

TCP
TCP-Transmission Control Protocol :
TCP is a connection-oriented protocol. It does not mean it
has a physical connection between sender and receiver. TCP
provides the function to allow a connection virtually exists
also called virtual circuit or connection. TCP offers a reliable
byte-stream delivery service .

Functions provides by TCP are :

Dividing a chunk of data into segments.
Reassembly segments into the original chunk.
Provide further the functions such as reordering and data
resend.

TCP-Services :
Services provided by TCP applications are :

a. Connection-orientation : TCP provides connectionoriented service in which an application must first

request a connection to a destination, and then use the
connection to transfer data.
b. Point-to-Point communication : Each TCP connection
has exactly two endpoints. TCP always provides the endto-end communication.
c. Complete Reliability : TCP guarantees that the data sent
across a connection will be delivered exactly as sent,
with no data missing or out of order.
d. Full Duplex communication : A TCP connection allows
data to flow in either direction, and allows either
application program to send data at any time. TCP can
buffer outgoing and incoming data in both directions,
making it possible for an application to send data and

then to continue computation while the data is being

transferred.
e. Stream Interface : TCP provides a stream interface in
which an application sends a continuous sequence of
octets across a connection. That is, TCP does not
guarantee that data will be delivered to the receiving
application in the same size pieces that it was transferred
by the sending applications.
f. Reliable Connection Start-Up : TCP requires that when
two applications create a connection, both must agree to
the new connection; duplicate packets used in previous
connections will not appear to be valid responses or
otherwise interface with the new connection.
g. Graceful Connection Shutdown : An application
program can open a connection, send arbitrary amounts
of data, and then request that the connection be
shutdown. TCP guarantees to deliver all the data reliably
before closing the application.

End to End services and datagrams :

TCP is called end-to-end protocol because it provides a
connection directly from an application on one computer to
another application on a remote computer. The applications
can request that TCP form a connection, send and receive data
and close the connection. The connections provided by TCP
are called Virtual connections because they are achieved in
software. Indeed, the underlying internet system does not
provide hardware/software support for connections. Instead,

the TCP software modules on two machines message

exchanges to achieve the illusion of a connection.
TCP uses IP to carry messages. Each TCP message is
encapsulated in an IP datagram and sent across the internet.
When the datagram arrives on the destination host, IP passes
the contents to TCP.

Packet Loss and Retransmission :

Retransmission is a technique used to handle parts of the
problem. When TCP sends data, the sender compensates for
packet loss by implementing a retransmission scheme. Both
sides of a communication participates. When TCP receives
data, it sends an acknowledgment back to the sender.
Whenever it sends data, TCP starts a timer. If the timer
expires before an acknowledgment arrives, the sender
retransmits the data. This scheme handles communication
across an arbitary internet and allows multiple application
programs to communicate concurrently.
TCP faces a more difficult challenge than distinguishing
between local and remote destinations : bursts of datagrams

can cause congestion, which causes transmission delays along

a given path to change rapidly.
Adaptive retransmission :
Earlier a fixed value for retransmission delay was chosen
by designers. Then they chose delay to be adaptive ,
considering current delay on each connection. TCP estimates
round-trip delay for each connection which measures the
time to get response. The average calculated by applying
statistical function to sequences of round trip estimates.

Retransmission Times :
Consider a case of packet loss on two connections that
have different round-trip delays.
As the diagram shows, TCP sets the transmission to be
slightly longer than the mean round-trip delay. If the delay is
large, TCP uses a large transmission timeout; if the delay is
small, TCP uses a small timeout. The goal is to wait long
enough to determine that a packet was lost, without waiting
longer than necessary.

Timeout and retransmission on two connections that that have different round-trip
delays. TCP optimizes throughput by using a round-trip estimate to compute a
retransmission timer.

Buffers, Flow control and Windows :

TCP uses a window mechanism tool to control the flow of
data. When a connection is established, each end of the
connection allocates a buffer to hold incoming data, and sends
the size of the buffer to the other end. As data arrives, the
receiver sends acknowledgements, which also specify the
remaining buffer size. The amount of buffer size available at
any time is called the Window, and a notification that specifies
the size is called a Window Advertisement.
If the receiving application can read data as quickly as it
arrives, a receiver will send a positive window advertisement
along with each acknowledgement. However, if the sending
side operates faster than a receiving side, incoming data will
eventually fill the receivers buffer, causing the receiver to
advertise a zero window. A sender that receives a zero window

advertisement must stop sending until the receiver again

advertises a positive window.

Three-Way handshake :
To guarantee that connections are established or
terminated reliably, TCP uses a Three-Way -Handshake in
which three messages are exchanged. Scientists have proved
that a three way exchange is necessary and sufficient to ensure
unambiguous agreement despite packet loss, duplication and
delay.
TCP uses the term Synchronization segment(SYN
segment) to describe messages in a three-way-handshake used
to create a connection, and the term Finish segment(FIN

segment) to describe messages in a three-way-handshake used

to close a connection.
Like other messages, TCP transmits lost SYN or FIN
segments. Furthermore, the handshake guarantees that TCP
will not open or close a connection until both ends have
interacted. Part of the three-way-handshake used to create a
connection requires each end to generate a random 32-bit
sequence number. If an application attempts to establish a new
TCP connection after a computer reboots, TCP chooses a new
random sequence, and then establish a new TCP connection
without interference from duplicate or delayed packets.

Congestion Control :
One of the most interesting aspects of TCP is a mechanism
for Congestion control. In most modern internets, packet loss
is more likely to be caused by congestion than a hardware
failure. Interestingly, transport protocols that retransmit can
exacerbate the problem of congestion by injecting additional

copies of a message. If congestion triggers excessive

retransmission, the entire system can reach a state of
congestion collapse, analogous to a traffic jam on a highway.
To avoid the problem, TCP always uses packet loss as a
measure of congestion, and responds by reducing the rate at
which it retransmits data.
Whenever a message is lost, TCP begins congestion
control. Instead of retransmitting enough data to fill the
receivers buffer, TCP begins by sending a single message
containing data. If the acknowledgement arrives without
additional loss, TCP doubles the amount of data being sent,
and sends two additional messages. If acknowledgement
arrive for those two, TCP sends four more and so on. The
exponential increase continues until TCP is sending half of the
receivers advertisement window, at which time TCP slows
down the rate of increase.
TCPs congestion control scheme responds well to
increased traffic in an internet. By backing of quickly, TCP is
able to alleviate congestion.

TCP Segment Format :

TCP uses a single format for all messages, including
messages that carry data, acknowledgements, and messages
that are part of three-way handshake used to create or
terminate a connection. TCP uses the term segment to refer to
a message.

A TCP connection contains two streams of data, one

flowing in each direction. If the application at each end are
sending data simultaneously, TCP can send a single segment
that carries the acknowledgement for incoming data, a
window advertisement that specifies the amount of additional
buffer space available for incoming data and incoming data.
Thus, some of the fields in the segment refer to the data
stream travelling in the forward direction, while other fields
refer to the data stream travelling in the reverse direction.
Source port identifies the application sending data.
Destination port - identifies the application recieving
data.
Sequence number- sequence number of outgoing
message number of first octet.
ACKN Number- sequence number expected next.
Window buffer available.

 Checksum- for header and data.