Chap 08 Modified PP2003

7/25/2019 Chap 08 Modified PP2003

1/68

TCP/IP Protocol Suite 1

Chapter 8Chapter 8

Upon completion you will be able to:

Internet ProtocolInternet Protocol

Understand the format and fields of a datagram

Understand the need for fragmentation and the fields involved

Understand the options available in an IP datagram

Be able to perform a checksum calculation

Understand the components and interactions of an IP package

Objectives


2/68


Figure 8.1 Position of IP in TCP/IP protocol suite

IP: connectionless, unreliable, packet switching w/ datagrams best-efort delier! serice


3/68


Internet Protocol (IP) is the transmissionmechanism used by the TCP/IP protocols at thenetwork layer.

The term best-efort means that IP packets canbe corrupted, lost, arrive out o order or delayedand may create con!estion "or the network.


4/68

TCP/IP Protocol Suite #

8.1 DATAGRAM

packet in the IP layer is called a datagram! packet in the IP layer is called a datagram!a variable"length packet consisting of two parts: header and data#a variable"length packet consisting of two parts: header and data#

The header is $% to &% bytes in length and contains information essentialThe header is $% to &% bytes in length and contains information essential

to routing and delivery#to routing and delivery#

'ome of the fields:'ome of the fields:

()* " version numbers! + and &()* " version numbers! + and &

,-). " header length in +"byte words# (alue of means $% byte header,-). " header length in +"byte words# (alue of means $% byte header


5/68

TCP/IP Protocol Suite $

Figure 8.2 IP datagram


6/68

TCP/IP Protocol Suite %

Figure 8.3 'ervice type or 0ifferentiated 'ervices " 0' field

This "eld was preiousl! called Serice t!pe# $ow called%iferentiated Serices#The precedence sub"eld was designed,but neer used in ersion T'S (T!pe o) Serice* bits - see ne+t slide#


7/68

TCP/IP Protocol Suite &

Table 1#2Table 1#2 Types of serviceTypes of service

I) !ou want to send a packet with a special t!pe o) serice,use one o) the aboe -bit sets#


8/68

TCP/IP Protocol Suite '

Table 1#$Table 1#$ 0efault types of service0efault types of service

Some apps

hae de)aultserice t!pes#


9/68

TCP/IP Protocol Suite

Table 1#3Table 1#3 (alues for codepoints(alues for codepoints

I) we call these 8 bits %iferentiated Serices (and not theolder Serice T!pe*, then the "rst si+ bits are called code-points#

hen the . right-most bits are , the . le)t-most bits are the saas the precedence bits )rom the preious slides#hen the . right-most bits are not all s, the 0 bits de"ne 0&Serices based on the priorit! assignment b! the Internet orlocal authorities# 1ssignments hae not !et been "nali2ed#


10/68


Figure 8.4 )ncapsulation of a small datagram in an )thernet frame

The total length "eld de"nes the total length o) the datagram

including the header#

Total length "eld is 30 bits, or 0,. b!tes# ') which 4 to0 b!tes are the header#

I) an IP datagram is short, and is packaged into an 5thernet

)rame, remember that the minimum pa!load si2e o) an5thernet )rame is &0 b!tes#


11/68


Identi"cation, 6lags, and 6ragmentation ofset are all usedto per)orm )ragmentation, which we will coer shortl!#

Time to 7ie - 8-bit "eld, so Time to 7ie can be set to 4#1s it passes thru a router, the router decrements thecounter# hen counter hits , the datagram is deleted(and ICP sends an error message back to the source*#

Time to li*e. + data!ram has a limited li"etime in its tra*el throu!h an intern,eld was ori!inally desi!ned to hold a timestamp which was decremented b*isited router. The data!ram was discarded when the *alue became -ero.

h! might a host set the Time to 7ie "eld to 39


12/68


Figure 8.5 4ultiple5ing

The Protocol "eld (8 bits* identi"es the upper la!er protocol

that is using IP )or transmission o) its data#


13/68


Table 1#+Table 1#+ ProtocolsProtocols


14/68

TCP/IP Protocol Suite 1#

n IP packet has arrived with the first 1 bits as shown:

Example1

%2%%%%2%

The receiver discards the packet# 6hy7


15/68

TCP/IP Protocol Suite 1$

n IP packet has arrived with the first 1 bits as shown:

Example1

%2%%%%2%

The receiver discards the packet# 6hy7

'olution

There is an error in this packet# The + left"most bits 8%2%%9

show the version! which is correct# The ne5t + bits 8%%2%9 showthe header length which means 8$ ; + < 19! which is wrong#

The minimum number of bytes in the header must be $%# The

packet has been corrupted in transmission#


16/68

TCP/IP Protocol Suite 1%

In an IP packet! the value of ,-). is 2%%% in binary# ,owmany bytes of options are being carried by this packet7

Example2


17/68

TCP/IP Protocol Suite 1&

In an IP packet! the value of ,-). is 2%%% in binary# ,owmany bytes of options are being carried by this packet7

Example2

'olution

The ,-). value is 1! which means the total number of bytes

in the header is 1 ; + or 3$ bytes# The first $% bytes are the

base header! the ne5t 2$bytes are the options#


18/68

TCP/IP Protocol Suite 1'

In an IP packet! the value of ,-). is 2&and the value of the total length field is %%$12& # ,ow

many bytes of data are being carried by this packet7

Example3


19/68


In an IP packet! the value of ,-). is 2&and the value of the total length field is %%$12& # ,ow

many bytes of data are being carried by this packet7

Example3

'olution

The ,-). value is ! which means the total number of bytes

in the header is ; + or $% bytes 8no options9# The total length

is +% bytes! which means the packet is carrying $%bytes of data

8+% = $%9#


20/68


n IP packet has arrived with the first few he5adecimal digitsas shown below:

Example4

+%%%%$1%%%2%%%%%2%$ # # #

,ow many hops can this packet travel before being dropped7The data belong to what upper layer protocol7


21/68


n IP packet has arrived with the first few he5adecimal digitsas shown below:

Example4

+%%%%$1%%%2%%%%%2%$ # # #

,ow many hops can this packet travel before being dropped7The data belong to what upper layer protocol7

'olution

To find the time"to"live field! we skip 1 bytes 82& he5adecimaldigits9# The time"to"live field is the ninth byte! which is %2# This

means the packet can travel only one hop# The protocol field is

the ne5t byte 8%$9! which means that the upper layer protocol is

I>4P 8see Table 1#+9#


22/68


8.2 FRAGMENTATION

The format and si?e of a frame depend on the protocol used by theThe format and si?e of a frame depend on the protocol used by thephysical network# datagram may have to be fragmented to fit thephysical network# datagram may have to be fragmented to fit the

protocol regulations#protocol regulations#

The topics discussed in this section include:The topics discussed in this section include:

4a5imum Transfer Unit 84TU94a5imum Transfer Unit 84TU9

@ields *elated to @ragmentation@ields *elated to @ragmentation


23/68


Figure 8.6 4TU

T - a+imum Trans)er nit


24/68


Table 1#Table 1# 4TUs for some networks4TUs for some networks

a+ datagram si2e )or IP is 0. b!tes# So i) we hae a ma+si2ed datagram to send oer 5thernet, what do we do9


25/68


Figure 8.7 @lags field

1 datagram can be )ragmented b! the source host or an!router in the path# ;eassembl! is done onlyb! the destination

host#

ost "elds are copied )rom one )ragment to the ne+t# The ."elds that are not copied are the


26/68


Figure 8.8 @ragmentation e5ample

The 6ragmentation 'fset tells what position this )ragmentis in the whole stream# The ofset counts b! 8# So i) a)ragment is supposed to start at b!te &, the ofset e>uals

#

= It is the ofset o) the data in the original datagram measurein units o) 8 b!tes


27/68


Figure 8.9 0etailed fragmentation e5ample


28/68


packet has arrived with an 4 bit value of %# Is this the firstfragment! the last fragment! or a middle fragment7 0o we

know if the packet was fragmented7

Example5


29/68


packet has arrived with an 4 bit value of %# Is this the firstfragment! the last fragment! or a middle fragment7 0o we

know if the packet was fragmented7

Example5

'olution

If the 4 bit is %! it means that there are no more fragments the

fragment is the last one# ,owever! we cannot say if the original

packet was fragmented or not# non"fragmented packet isconsidered the last fragment#


30/68


packet has arrived with an 4 bit value of 2 and afragmentation offset value of ?ero# Is this the first fragment!

the last fragment! or a middle fragment7#

Example7


31/68


packet has arrived with an 4 bit value of 2 and afragmentation offset value of ?ero# Is this the first fragment!

the last fragment! or a middle fragment7#

Example7

'olution

Because the 4 bit is 2! it is either the first fragment or a middle

one# Because the offset value is %! it is thefirstfragment#


32/68


packet has arrived in which the offset value is 2%%# 6hat isthe number of the first byte7 0o we know the number of the

last byte7

Example8


33/68


packet has arrived in which the offset value is 2%%# 6hat isthe number of the first byte7 0o we know the number of the

last byte7

Example8

'olution

To find the number of the first byte! we multiply the offset

value by 1# This means that the first byte number is 1%%# 6e

cannot determine the number of the last byte unless we know

the length of the data#


34/68


packet has arrived in which the offset value is 2%%! the valueof ,-). is and the value of the total length field is 2%%#

6hat is the number of the first byte and the last byte7

Example9


35/68


packet has arrived in which the offset value is 2%%! the valueof ,-). is and the value of the total length field is 2%%#

6hat is the number of the first byte and the last byte7

Example9

'olution

The first byte number is 2%% ; 1 < 1%%# The total length is 2%%

bytes and the header length is $% bytes 8 ; +9! which means

that there are 1% bytes in this datagram# If the first byte

number is 1%%! the last byte number must be 1A#


36/68


8.3 OPTIONS

The header of the IP datagram is made of two parts: a fi5ed part and aThe header of the IP datagram is made of two parts: a fi5ed part and avariable part# The variable part comprises the options that can be avariable part# The variable part comprises the options that can be a

ma5imum of +% bytes#ma5imum of +% bytes#


@ormat@ormat

ption Typesption Types

Figure 8 10 ption format


37/68


Figure 8.10 ption format

$ot all routers/hosts use these options, but the! must beread! to do so i) the! are present in the datagram#

Cop! - tells whether to cop! this option into a )ragmentClass - de"nes the general purpose o) the option

Figure 8 11 Categories of options


38/68


Figure 8.11 Categories of options

1s we ?ust saw, onl! 0 options in use currentl!# The single-options are onl! 3 b!te in length and do not re>uire lengthdata "elds#

Figure 8 12 .o operation option


39/68


Figure 8.12 .o operation option

= sed as a "ller between options# 6ore+ample,3#can be used to align the ne+t option #

4#Can be used to align beginning o) anoption #on a 30- or .4-bit boundar!#

Figure 8 13 )nd of option option


40/68


Figure 8.13 )nd of option option

enotes the end o" the options and that the data isne0t. used "or paddin! at the end o" the option ,eld can only be used as the last option. nly one endo"option option can be used

Figure 8 14 *ecord route option


41/68

TCP/IP Protocol Suite #1

Figure 8.14 *ecord route option

;ecords the route a datagram takes thru routers# Can onl! reco@ routers, since ma+ si2e o) the header is 0 b!tes, 4 b!tes )orbase header, leaing onl! & b!tes )or options#

used to record the Internet routers that handle thedata!ram.

en!th4 Total len!th o" the option(includin! code andlen!th ,elds)

Figure 8.15 *ecord route concept


42/68


Figure 8.15 *ecord route concept

The Pointer ,eld (# then ' then 12 then 1%) is the bytenumber o" the ,rst a*ailable space.

Figure 8.16 'trict source route option


43/68


Figure 8.16 'trict source route option

6or when a datagram has to )ollow a gien, "+ed route#

Figure 8.17 'trict source route concept


44/68

TCP/IP Protocol Suite ##

Figure 8.17 'trict source route concept

6irst hop address is here

Second hop address is here

$ote that as hops are made, ne+t hop is replaced withaddress o) router !ou ?ust went thru

Figure 8.18 -oose source route option


45/68

TCP/IP Protocol Suite #$

Figure 8.18 -oose source route option

Similar to "+ed route - each router in the list must be isited,but other routers can be isited too#

Figure 8.19 Timestamp option


46/68

TCP/IP Protocol Suite #%

g p p

Can be used i) !ou want to record the time the datagramisits each router# Time in milleseconds, niersal Time#

'-6low bits (oer


47/68

TCP/IP Protocol Suite #&

g f f g p

The 6lag bits tell the router whether to do one o) the)ollowing operations:

: add onl! the timestamp in the proided "eld

3 : add each routerAs outgoing IP address and the timestamp

. : each router must check the gien IP address with its ownincoming IP address# I) matched, the router oerwrites the IPaddress with its outgoing IP address and adds the timestamp

Figure 8.21 Timestamp concept


48/68

TCP/IP Protocol Suite #'

g p p


49/68


6hich of the si5 options must be copied to each fragment7

Example10


50/68


6hich of the si5 options must be copied to each fragment7

Example10

'olution

6e look at the first 8left"most9 bit of the code for each option#

a# .o operation: Code is %%%%%%%2 not copied#b# )nd of option: Code is %%%%%%%% not copied#

c# *ecord route: Code is %%%%%222 not copied#

d# 'trict source route: Code is 2%%%2%%2 copied#

e# -oose source route: Code is 2%%%%%22 copied#f# Timestamp: Code is %2%%%2%% not copied#


51/68

TCP/IP Protocol Suite $1

6hich of the si5 options are used for datagram control andwhich are used for debugging and management7

Example11


52/68


6hich of the si5 options are used for datagram control andwhich are used for debugging and management7

Example11

'olution

6e look at the second and third 8left"most9 bits of the code#

a# .o operation: Code is %%%%%%%2 datagram control#

b# )nd of option: Code is %%%%%%%% datagram control#

c# *ecord route: Code is %%%%%222 datagram control#

d# 'trict source route: Code is 2%%%2%%2 datagram control#e# -oose source route: Code is 2%%%%%22 datagram control#

f# Time stamp: Code is %2%%%2%% debugging and management

control#


53/68


8.4 CHECKSUM

The error detection method used by most TCP/IP protocols is called theThe error detection method used by most TCP/IP protocols is called thechecksum# The checksum protects against the corruption that may occurchecksum# The checksum protects against the corruption that may occur

during the transmission of a packet# It is redundant information addedduring the transmission of a packet# It is redundant information added

to the packet#to the packet#

The checksum in the IP packet covers only the header! not the data#The checksum in the IP packet covers only the header! not the data#

Two reasons:Two reasons:

" ll higher level protocols have their own checksum#" ll higher level protocols have their own checksum#

" The header of the IP packet changes with each visited router! but the" The header of the IP packet changes with each visited router! but the

data doesnDt#data doesnDt#

The topics discussed in this section include:The topics discussed in this section include:Checksum Calculation at the 'enderChecksum Calculation at the 'ender

Checksum Calculation at the *eceiverChecksum Calculation at the *eceiver

Checksum in the IP PacketChecksum in the IP Packet


54/68

TCP/IP Protocol Suite $#

To create the checksum the sender does the following:

The packet is divided into k sections! each of n bits#

ll sections are added together using 2Ds complement

arithmetic#

The final result is complemented to make the

checksum#

Note:Note:

Figure 8.22 Checksum concept


55/68

TCP/IP Protocol Suite $$

Figure 8.23 Checksum in oneDs complement arithmetic


56/68

TCP/IP Protocol Suite $%


57/68

Figure 8.24 )5ample of checksum calculation in binary


58/68

TCP/IP Protocol Suite $'


59/68


-et us do the same e5ample in he5adecimal# )ach row has fourhe5adecimal digits# 6e calculate the sum first# .ote that if an

addition results in more than one he5adecimal digit! the right"

most digit becomes the current"column digit and the rest are

carried to other columns# @rom the sum! we make thechecksum by complementing the sum# ,owever! note that we

subtract each digit from 2 in he5adecimal arithmetic 8Eust as

we subtract from 2 in binary arithmetic9# This means the

complement of ) 82+9 is 2 and the complement of + is B 8229#

@igure 1#$ shows the calculation# .ote that the result 81BB29is e5actly the same as in )5ample 2A#

Example18

See Next Slide

Figure 8.25 )5ample of checksum calculation in he5adecimal


60/68

TCP/IP Protocol Suite %


61/68

TCP/IP Protocol Suite %1

Checkppendi5 Cfor a detailed

description of checksum calculationand the handling of carries#

Note:Note:


62/68


8.5 IP PACKAGE

6e give an e5ample of a simplified IP software package to show its6e give an e5ample of a simplified IP software package to show itscomponents and the relationships between the components# This IPcomponents and the relationships between the components# This IP

package involves eight modules#package involves eight modules#


,eader"dding 4odule,eader"dding 4odule

Processing 4oduleProcessing 4odule

FueuesFueues

*outing Table*outing Table

@orwarding 4odule@orwarding 4odule

4TU Table4TU Table@ragmentation 4odule@ragmentation 4odule

*eassembly Table*eassembly Table

*eassembly 4odule*eassembly 4odule

Figure 8.26 IP components


63/68


5 d ddi d l


64/68

5eaderaddin! module

6ecei*e 4 data destination address

1. 7ncapsulate the data in an IP data!ram by addin! the IP

header

2. Calculate the checksum and insert it in the checksum ,eld3. 8end the data to the correspondin! input 9ueue

#. 6eturn

TCP/IP Protocol Suite %#

P i d l


65/68

Processin! module

1. 6emo*e one data!ram "rom one o" the input 9ueues2. I"(destination address is 12&.:.;.< or matches one o" the

local addresses (hostids))

1. 8end the data!ram to the reassembly module

2. 6eturn

3. I"(machine is a router)1. ecrement TT

#. I"(TT less than or e9ual to -ero)

1. iscard the data!ram

2. 8end an IC=P error messa!e

3. 6eturn$. 8end the data!ram to the routin! module

%. 6eturn

TCP/IP Protocol Suite %$

> t ti d l


66/68

>ra!mentation module? 6ecei*e 4 an IP data!ram "rom the routin! module

1. 70tract the si-e o" the data!ram2. I"(si-e @ =TA o" the correspondin! network)

1. I"((do not "ra!ment) bit is set)

1. iscard the data!ram


3. 6eturn2. 7lse

1. Calculate the ma0imum si-e

2. i*ide the data!ram into "ra!ments

3. +dd header to each "ra!ment

#. +dd re9uired options to each "ra!ment

$. 8end the data!rams

%. 6eturn3. 7lse

1. 8end the data!ram#. 6eturn

TCP/IP Protocol Suite %%

6 bl t bl


67/68

6eassembly table

? Ased by the reassembly module? >i*e ,elds

8tate(>677 or IBA87) 8ource IP address(source IP o" the data!ram) ata!ram I(uni9uely de,ne a data!ram) Timeout(predetermined amount o" time in which all

"ra!ments must arri*e) >ra!ments(a pointer to a linked list o" "ra!ments)

TCP/IP Protocol Suite %&

6 bl d l


68/68

6eassembly module

6ecei*e 4 an IP data!ram "rom the processin! module

1. I"(oset *alue is -ero and the = bit is )1. 8end the data!ram to the appropriate 9ueue

2. 6eturn.

2. 8earch the reassembly table "or the correspondin! entry

3. I"(not "ound)

1. Create a new entry#. Insert the "ra!ment at the appropriate place in the link list1. I"(all "ra!ments ha*e arri*ed)

1. 6eassemble the "ra!ments

2. eli*er the data!ram to the correspondin! upper layer protocol

3.6eturn

2. 7lse1. Check the timeout

2. I"(timeout e0pired)

1. iscard all "ra!ments


$. 6eturn

Documents

Chap 08 Modified PP2003