Internet as Virtual Network

• Network where all devices, servers, virtual machines, and data centers
that are connected are done so through software and wireless technology
• Enables communication between multiple computers, virtual machines
• Enables devices across many locations to function with the same
capabilities as a traditional physical network.
• This allows for data centers to stretch across different physical locations,
and gives network administrators new and more efficient options, like
the ability to easily modify the network as needs change, without having
to switch out or buy more hardware;
• Greater flexibility in provisioning the network to specific needs and
applications
• The capacity to move workloads across the network infrastructure
without compromising service, security, and availability.
 Virtual Network: How it Works
• A virtual network uses modern technology to create an extended
network that works wirelessly. This includes:
• vSwitch Software: Virtualization software on host servers that
allows you to set up and configure a virtual network.
• Virtual network adapter: Creates a gateway between networks.
• Physical network: Required as a host for the virtual network
infrastructure.
• Virtual machines and devices: Instruments that connect to the
network and allow various functionality.
• Servers: Part of the network host infrastructure.
• Firewalls and security: Designed for monitoring and stopping
security threats.
 Classes of Virtual Network
There are three classes of virtual networks:
1) VPN
2) VLAN
3) VXLAN
 VPN
• virtual private network
• uses the internet to connect two or more
existing networks
• This internet-based virtual network allows
users to log in from anywhere to access the
physical networks that are connected.
• VPNs are also used for masking internet use on
public WiFi and ensuring secure browsing
 VPN
• A VPN is created when data attached to packets
defines routing information that takes users to
the applicable address.
• a tunnel of addresses is created, encrypting the
browsing history and making it possible to
access information remotely
• VPNs provided a small-scope, fully virtual
network that uses the internet to allow for
people to connect.
 VLAN
• Uses partitions to group devices on a LAN network
into domains with resources and configurations that
are applied to each
• Using a VLAN allows for better security,
monitoring, and management of the devices and
servers within a specific domain
• This is especially true for large networks that may
be more vulnerable to attack when domains are not
used and monitored individually.
 VXLAN
• VXLAN means virtual extensible local area
network.
• In this network, your level 3 network
infrastructure provides a tunnel into level 2.
• Virtual switches create endpoints for each
tunnel, and another piece of technology, called
a physical or virtual base case, can route data
between endpoints.
 Benefits of Virtual Networking
• Remote work capabilities
• Digital security
• Streamlines hardware
• Flexibility and scalability
• Cost savings
• Productivity
Virtual Network
Network where all devices, servers, virtual
machines, and data centers that are connected are
done so through software and wireless technology
Enables communication between multiple
computers, virtual machines
Enables devices across many locations to function
with the same capabilities as a traditional physical
network.
Internet Architecture And Philosophy

Application Services

Reliable Transport Service

Connectionless Packet Delivery Service

The three conceptual layers of internet services
The Conceptual Service Organization

Internet software is designed around three conceptual

networking services arranged in a hierarchy
Much of its success has resulted because this
architecture is surprisingly robust and adaptable.
Advantages of this conceptual separation is that it
becomes possible to replace one service without
disturbing others
Research and development can proceed concurrently
on all three.
Connectionless Delivery System
Most fundamental internet service consists of a
packet delivery system.
Technically, the service is defined as :
An unreliable
Best-effort
Connectionless packet delivery system
Analogous to the service provided by network
hardware that operates on a best-effort delivery
paradigm.
Purpose of the Internet Protocol
Protocol that defines the unreliable, connectionless
delivery mechanism is called the Internet Protocol
IP provides three important definitions:
First, the IP protocol defines the basic unit of data
transfer used throughout a TCP/IP internet.
Second, IP software performs the routing function,
choosing a path over which data will be sent.
Third, formal specification of data formats and
routing, IP includes a set of rules that embody the idea
of unreliable packet delivery.
The Internet Datagram
DATAGRAM
Combination of the words data and telegram.
Message containing data that is sent from one location to
another.
Similar to packet, but doesn’t require confirmation that it
has been received
Ideal for streaming services, where the constant flow of
data is more important than 100% accuracy
Also called “IP Datagram's”
Data is transmitted via the Internet protocol, it is broken up
into packets or datagram's, which contain a header plus the
actual data transmitted.
DATAGRAM

Datagram header defines the source and destination of

the data
Other information, such as the total length (or size) of
the datagram, time to live (TTL), and the specific
protocol used to transfer the data.
Datagram's are sent via the UDP protocol, which is
used for media streaming and other services that do not
require confirmation that the data has been received.
Packets, on the other hand, are typically sent via TCP,
which guarantees all the data sent has been received.
Internet Datagram

DATAGRAM HEADER DATAGRAM DATA AREA

General form of IP datagram
Datagram Format
Datagram Type Of Service And Differentiated
Services
Differentiated services
 1lte ISO model, buih to
specific l1y« for intcmctwortit roudn1 in the same way TCP/IP pl'otocols do.

11.5 X:i5 And Ila Relation To The ISO Model

Ahhough it was deslane,cl 10 provkle I conc:q,tual model and not an implemmtt.-
tion guide. the ISO layerina scheme has bttn the basis for several protocol implement,,-
Amoni the protocols commonIr associated wilh the ~ I . the suite was of~
toools known u X-25 is probably the best known and most widely u.,ed. X-15 es-
abhshc d as a ffi:Ommmdaliog of lhc lrtl~""""1ttol TdfflHfflflt11klllUHU U,rlo,, (ITU/,
f6nnffly the CCITT. an Of1anization that ~omm cnds standards for intcma1ional tcte-
scrvicss. X.25 has bttn adopted by ~lie da1a networks, and became cspccielly
popu ar in !_urop(. Considering X.U will help explain ISO layering.
In the X.25 view, a network opcnte s much like I tclcrbooc ua&c,n, An X.25 net-
work is assumed 10 consist of comtx packet switches that ~tain t h e ~
nccdod to roule packets. Host$ do J>1 attach dittetlY to communica1i9n Wires of "the
[Link] each hosl attaches to pgs gf the packet switches using a serial com-
rml'tication hne. In one sense, the connection between a host and an x.25 packet
fW'Hch 1s a miniature network consisting of onC serial link. The host must follow a
complicated procedure to transfer packets onto 1he network.
• PJ,ys,cil LIIJ'"• X.15 spccifi n a standard for the physical intereonncction
bdween host computers and nttwork packet switchC$, as well as the proeed wa used to
transfer packets from one machine to anothet. In the reference model, layer l specifies
the physicaJ intcrc:oMection including electrical characteristics of voltage and currenL
A corTCSpODdingprotocol, X.21, gives the details used by public data networks.
• Dalo J..iM Layn. The layer 2 portion of the x.25 ptotbcol spccif ia how data
travels between a host and the packet switch to which it connects. X.25 uses the tena.
frt1111e to refer to a unit of data as it passes between a host and a packet switch (it is im•
ponant td undetstand that the x.25 definition o f ~ differs slightly from the Way we
hive defined it). Because raw hardwa redeliv en only a stream of bits, the layer 2 prq-
tocol must define the fonnat of frames and specify how the two P,achin es recogn ize
frame bounda ljcs. Because transmission errors can destroy data. the layer 2 protocol in•
eludes error detection (e.a.. a frame checksum). Finally. because transmission is unreli•
able. the layer 2 protocol specifics an exchange of acknowledgements that allows the
two machines to know when a frame has been tnnsfer rcdsuc cessful ly.
One commonly used layer 2 protocol. named the Hlfh u.-.1 Dato IJ!olt Comnwni-
ctllion, is best known by its ac:.ronym, HDLC. Seven.I version s of HOLC'[Link], ·with the
most recent known u HDLCIUJ'B, It is impor1itnt to remember that successful
uansfer at layer 2 means a frame has been possed to the network packet switch for
delivery; it docs not guarwllee that the packet switch accepted the packet or wu able to
route it
• Ndwo rt Llly,r. The ISO reference model specifics that the third layer contains
functionality that completes the definition of the interaction between host and netwO<I<.
r Sec. 11.S X.25 And Its Relation To Th
e ISO Model 183

Called the network or co

. .
tio n subn et lay er, thi s lay er defines the basic unit of
:~ ca t- .
transfer across the ne tw oz m" 25mcludes the conc~pts_ of desti~ation addressing and rou h
ing. Remember that in the • world, com°!un1~at1on between host and packet switc
might
is conceptually isolated fro m the traffic that is being passed. Th~s, the netwo~k can
allow packets defined to be larger thft!l ~f -~ ~s that
lay er 3 pro toc ols m the net-
be transferred at I by
2 - Th e lay er 3 so ftw are assembles a packet 1n the for switch.
work , ayer ces) to the packet
811 ~ uses layer 2 to transfer it (possibly in pie . ·
La er ;~ ts ne tw or k co ng est ion pro ble ms .
Y • mu st also respond to vid es en d-t o-e nd reliability by having the de er
stina-
ye r 4 pro
. h [Link]:sport Layer. La the source host. The idea here is that even though low
f 0
n os t communicate with
ayers of protocols provide
rel iab le ch ec ks at ea ch transfer, the end-to-end lay
. •
er double

th at no machine in the middle failed

ch ec ks to m ~e su re
ers of the IS O mo de l de scribe how protocol software
lay ms. The
• Sessio'! Layer. Hi gh er cti on ali ty ne eded by application progra
all the fun
,can be organized to hand
le
of rem ote ter mi na l ac cess so fundamental that
th e problem c
IS O co mm itt ee considered it. In fac t,., the ce ntr al se rvice offer~d by early publi
handle s -a spe-~
they as sig ne d lay er 5 to al to ho st int erc on nection. The carrier provide
ten nin '
d~ta .networks consisted of ck et 'As sem ble r An d·_ Di sassembler (PAD) on the
er called a Pa o carry their own compute
r
c1al pu rp os e ho~t co mp ut cri 6e rs, of len tra ve l~r s wh
ss. Subs ich
,network wi th di alu p acce ma ke a, ne tw or k co nn e1ction to the host with wh
PA D,
and .modem, dial up th e local log· in. M ~y carriers choose t~ ·make using the network
d
they wish to co mm un ica te, an than direct dialup.
tion less expensjve
for long .distance communica er 6 is int en de d to inqh;ide functions thal Jl)any ap
-
.IS O lay
• Presentation Layer. the ne tiiv ork . 'Ty,pical examples include sta
ndard
·w he n us ing
plication programs need ert grap'~ 1c s im ag es ,ntp bit streams for transmiss
ion
or co nv
routines that compress text IS O sta tjd ar~ kn ow n as 1_bs~ract Syntax N'!_t_ati!!!I
__1·
a rile an
across network. For exam ion of data that ~pplication programs use.·· One of the
~t
.(ASN.l)• provides a represe~ es ASN.1 to represent data;
b· us
TCP/1P protocols, SNMP. als , IS O lay er 7 inc lud es application programs that us
e
na lly
, • Application /Jaylr. Fi on ic ma il or file transfer programs. In particular
,
de ele ctr
the network. Examples inclu ele _c !fo ~ic ma il kn pw n as the ~-1Q9 stan~ard. In .-
ocol fo r
the ITU has devised a •prot on me s~ ag e ha nd ling system"s; the ISO version is
ed jointly_
fact,.(the ITU and ISO wor~ • • · •
called MO'/ZS. ,

yer Reference Model

\
\
'
11 .5.1 . The TCPn P 5-La
·,..
'' ---- . m od el did not arise·from a standards committee, but_
The s~cond major layering [Link] /IP proto col suite. With a little w_ork, ~e
that led to th e
came instead from research ribe th e TC P/ IP lay ering s·cheme, but the underlying
de sc
ISO model can be stretched to nt distinguishing the two.
different enou gh to wa rra
assumptions are
r
184
Protocol Layerin1 Oaap. 11
.
Broadly speaking TCP/IP softw are •is organiz .
software la ers th ,_ ed into fiye con~pt ual layers - four
al layers a/well att~u•!d on a fifth lay~r of hardware. Figure l l.5 shows the cOnceptu-
as e ,onn of data as it passes between them.

Conce ptual Layer Objec ts f?assed .,I

• Betwe en 1-'ayers

[ Appncat1on
I Messa ges or Stream s
[ Trans port
I .. , port Protoc ol Packe ts
Trans

l
Intern et
I ... /
, IP_ Datag rams
• .
lNetwork rnterfacej //
/

.........
. ......... ......... .. Netwo rk-Spe cific Frame s
.
.. Hardw are
_,/

Figure 11.S The· 4 con·ceptual layers of TCP/IP software above the hardware
layer, and, the form of objects passed between layers. The layer
1

labeled network,,interface is sometimes called the data link layer.

'
'I

' . \

• Applicat~on Layer. •At th~ highest layer, users invoke application programs that
access services available across a TCPDP internet. An application interacts with one of
I

the transport layer protocols ·~o send or receive data. Each application program chooses
'

the style of transpprt needed,··which can be either a sequence of individual messages or

a continuous stream of byte$. [Link] application program passes data in the required form
to the transport layer for delivery.
• Transport Layer. The primary. duty of the transport layer is to provide com-
muni;cation from one application program to another. Such communicatiop is often
called end-to-end. The transport layer may regulate flow of information. It may also
provide reliable transport, ensuring that data arrives without error and in sequence. To
do .so, transport protocol software arranges to have the receiving side send back ack-
nowledgements and the sending side retransmi~ lost packets. The transport software
divides the stream of data being transmitted into small pieces (sometimes called pack-
ets) and passes each packet along with a destination address to the next layer for
transmjssion.
, ayer, a gen-
Although 'Figure 11,.5 uses a single block to represent the application'l
eraJ purpose computer can have multiple application programs accessing .an internet at
one time. The ,transport layer must accept data ·from [Link] user programs and send it
to the next lowe.r layer. To do so, it adds additional information to each packet, includ-
"
 Sec..11..5 X. 25 W - llt]et,im To The ISO Model 185

ing codes that •h

identify wh •c • • . . . .
should .. . app11cat1on program sent 1t and which apphcatton program
um to
. ;.eceave at, as well as a checksum. The receiving machine uses the checks
the
ve~•fy at the packet arrived intact, and uses the destination code to identify
appli-
cation program to which it should be delivered.
. • WUIU !I Lay~r. As we have already seen, the Internet layer
handles communica-
the tzar
tion from one machi ne to another. ft accepts a request to send a packet from
should be
sport layer along with an identification of the machine to which the packet
, uses the
sen~ It enca~sulates the ,packet in an IP datagram, fills in the datagram h~der
send it to a
routin g algori thm to determine whether to deliver the datagram directly or
ission.
router, and passes the datagram to the appropriate network interface for transm
y, and uses
The Intern et layer also handle s incoming datagrams, checking their validit
locally or for-
the .routing algori thm to decide wheth er the datagram shoold be proces sed
t layer
warde d For datagr ams addr~ssed to the local machine, software in the interne
ols the
de1etes the datagr am header , and chooses from among severa [Link] port pro~oc
es ICMP er-
one that wilt handle the packet . Finally, the Intern et layer sends ~d recei~
rot and ·control messa ges as needed.
• Network Inte,face Layer. The lowest layer TCP/JP softwa re comp1-ises a net-
'
t
them ov,er
work interfa ce layer, respon sible for [Link] IP datagr ams and transm itting
.. ·when the
a specif ic ·network. A netwo rk interface may consi~t of a device driver (e.g
or a compl ex
netwo rk is a local area ~etwo [Link] which the,·maqhine attach es directl y)
ts of pack-
subsy stem that uses' its own data link protocol (~.g•i, [Link] netwo rk consis
et :switc h~.tha t comqn micate with h~·sts using HDLC)~ I '

f
.
•
.
I
I

11.6 Differences Between ~·s9 An~: Internet Layering

'
<

There ar~ two subtle and imp.~rtant differe nces betwe en the TCP/IP layeri ng
at-
scheme and the ISO/X.25. scheme. The first difference revolves around the focus of
the
tention on reliability, while .the 8eCond invqlves' the location of intelligence in
overall system.

11.6.1 Link-Level vs. End-T6-End Reliability

One major'ditference between the TCP/IP protocols and the X.25 protocols lies in
'

their appro,aches to· providing reliable data transfer services. In the X.2S model, proto-
1

·COJ software detects and handles errors at all layers. At the

link level, complex proto-
cols guarantee that the transfer between a host and the packet switch to which it con-
nects wiJJ correct. Ch~ksums ~cc<;>mpany each piece of data transferred, and the i:e-_
ceiver acknowledges each piece of data received. The link .Jayer protocol intlu,des
-
ti.~eo~ and [Link] algorithms that prevent data loss .and provide automatic
recovezy_after hardware fails and •restarts .
Successive layers of X.25 provide reliability of their own. At layer 3, X~25 also
provides error detection and recovery for packets transferred onto the network, using
checksums as weU as timeout and retransmission techniques. Finally, layer 4 must pro.-
I
r 186 Protocol Layerin1

vide end-to-end reliability ' h avi. ng the source correspond wit h the ultimate destination
Cb~p. 11

to
verify del ive ry.
In con tra st to such a s cheme, TCP/IP bases its protocol layering on the idea that re-
liability is an d-
e nd pro ble m. Th e arc hit ect ura l philosophy is simple: construct the
internet ·t en to-
I the exp ect ed loa d, but allo w individual 1inks or . [Link] to
so can handle le or no
Iose d ata or co rru pt it with out t ry1•ng to repeatedly recover. In fact, there 1s htt
r18b 1·r . net wo rk int erf ace layer software. Instead, the tran $pm :t layer
re •ty •n mo st TC P AP -
recovery problems.
handles mo st e'! or detection ahd er verification makes TC P/I P sof
tware
s ·freedo m fro m int erf ace lay
Th e ~ uit ing rectly. Intermediate routers can
discard
un der sta nd and im ple me nt cor
much <easter to transmission errors or tha t ean
no t be
be co me co ~p ted bec aus e of
d a ~ s tha t wh en the arrival rate exc eed s ma chi
ne capacity,
ey can dis car d dat agr am s
delivered. Th ut informing
ou te· dat itg ram s thr ou gh pat hs with ~horter or lon ger delay witho
an d can rer
the sou rce or des tin ati on . am s do no t arrive. De tec tio n and
ks me ~s - tha t· som e dat agr
Ha vin g un rel iab le .lin the sou rce ho st and the ultimate desti-
car rie d ou t bet we en
rec ov ery of da tag ram ,loss is ver ific ati on . Th e end -to -en d sof tw are loc ate
d
cal led en· d~t o-e nd
na tio n and, ·is, the ref 9re , no wl ed ge me nts ,an d timeouts to co n-
e~, :us es· che cks um s, ack .
in the TCP/IP tra nsp ort lay n-o rie nte d :X.25 pro toc ol lay eri ng , the •
/
lik e the "co ~n ect 1o
trol tra nsm iss ion . ··Thus, un
its reliability~control i~ on e lay er.
'TCP/IP so ftw are focuse~ mo st ~f , ..

d·D~cision Making
11~6.2 Locus of lntellig'encean
1

tw een the X.2 5 mo del and the TCPAP model emerges when
An oth er difference be
au tho rity and co ntr ol. As a .general ru l~ networks us ing
one considers the locus of that provides a transport service.
Th e
ide a tha t a,n etw ork is a uti lity
X.2 5 ad he re ta the acc ess and monitors tra ffi c to keep
vic e co ntr ols ne tw 0rk
vendor that offers the ser ork ve nd or als o ha nd les pro ble ms like
g. Th e ne tw
records ·.for ac co un tin g and billin int ern ally, making transfers reliable. This
ack no wl ed ge me nts
routirtg, flow control, an d
do . In short, the network is a complex,
le that the hosts can (or nee d to)
ves litt
1vJew lea nt,system to which one can attach relatively simple host computers; tlte hosts
independe
network operation.
(themselves participate minimally in the participate in almost all of the network proto-
to
. In contrast, 'ECPAP requires hosts ely im ple ment end-to-end error detec-
me ntion ed tha t ho sts a:ctiv
,- cols. We have already rou tin g because they must choose .a router
. Th ey also pa rticipa te in
tion and recovery ipa te in ne tw ork co ntrol because they must
d the y pa rtic
when sending datagrams, an mp are d to an X.25 network, a TCPAP
ssa ges. Th us, wh en co
handle ICMP control me et de liv ery system to which intelligent
as a rel ative ly sim ple pa ck
:~ internet can be viewed
.i ~(!~ts attach.
 • ' I
·I
) If
4 ,,
187
':!flePn«[Link]
,1

17
• ,~';, ~- ·•, •
{ '(''
inciple
1( 7 Th 8. Pr ot oc ol Layering Pr
layers,
of the par tic ula r lay eri ng sch eme used or the functions of the
lnde~endent idea, called the
n of lay ere d pro toc ols is bas ed on a fundamental idea. Th e
the operatio
layering principle, can be sum
marized succinctly:

re-
ed pro toc ols are des ign ed so that layer n at the destination
La yer
sen t by lay er n at the source.
ceives exa ctl y the sam e object
a. It allows the
pri nc ipl e exp lai ns wh y layering is such a powerful ide
Th e lay eri ng
att ent ion on on e lay er at a tim e, without wo rry ing about how
protocol de sig ne r to focus en bu ild ing a file tra nsf er ap pli cat ion
, the
. Fo r ex am ple , wh
oth er layers perfonn pli cat ion pro gra m ex ecu tin g on tw o
co m-
ly tw o ·co pie s of the ap
de sig ne r co nsi de rs on the y need to ex ch an ge for file transf
er. Th e
tra tes on the me ssa ge s
pu ter s, and co nc en e host receives ex act ly the da ta
tha t the ap-
tha t the ap pli cat ion on on
de sig ne r ass um es
ds.
pli cat ion on the oth er ho st sen
layering pri nc ipl e works:
Fig ure 11.6 illustrates ·how the
, I
Ho ste
Ho st A

AppHcation..
Ap pli ca tio n. ,; 1/,
,'
I

----------
1-• ' i )I

· identical
·'•
message•\
II
tra ns po rt
Tr an sp ort • I

f'
'

----------
I

identical
packet
I I Int er ne t
·Internet,
identical
-----------., datagra") .
-- . - . .,..,
Ne tw or k·
.,

Ne tw or~· k
Int er fac e
Int er fac e
--- jratM
----
',
identical

-- -- -- -- --
1

..

Physical Ne t

11.6 The path of a me ssa ge as it passes from an application on one

Figure yer n on host B receives
host to an ap plication on an other. La
host A sent.
exactly the same object that layer n on
188 [Link]. 11
Protocol Layering

t I
11.7.1 Layering In a TCPnP Internet Environmen
e, and the illustration in
. Our stat e~e nt of the layering principle is somewhat vagu distinguish between
use it fails to
Figu re 11.6 skim s ove r an important issue beca
transfers across multiple networks.
transfers from sour ce to ultimate destination and
of a message sent from an appli-
Fi~ure 11.7 illustrates the dist ~cti on, showing the path
her through a router.
cati on program on one host to an application on anot
network frames, one for
As the figure shows, message delivery uses two separate R to host B. The net-
from router
the tran smis sion from host A to router R, and another
d to R is identical to the frame sent
wor k layering principle states that the frame delivere
layers deal with end-to-end issues
by host A. By contrast, the [Link] transport
municates with its peer at the ulti-
and are desi gned so the software at the source com
s that the packet received by the
mat e destination. Thus, the layering principle state
to the packet sent by the transport
tran spor t laye r at the ultimate destination is identical
laye r at the original source.

HostB
HostA
identical Application
-- ---. -- --- ---- --- --
_ #

Apple Ilio n
I

--- --- message

-- ---
'

identical • Transport
Tra nsp ort
pack et

,RouterR
'
Inte rnet
(
Inte rnet
Inte rnet

- - - -- datagram -- ....
identical
- - - identical
data gram
--- - -
I

Network Net wor k

Net wor k
identical Interface identical Inte rfac e
Inte rfac e
fram e fram e

Physical Net 1 Physical Net 2

is used. The frame

Flgu re 11.7 The layering principle when a router
but
delivered to router R is exactly the frame sent from host A,
differs from the frame sent between R and B.
 Sec. 11.7
The Protocol Layering Principle
189
It is easy to understand that in h. . l
end-to-end transfers and th t tgher layers, the layering principle applies across l
st l
It is not as easy to~ h athat te lowe layer it applies to a single machine transfer.
hand, we have said th t ~w e ayering princ!ple applies to the Internet layer. On one
0sts
network, with the IP d: a~hed to an internet should view it as a large, virtual
tagrams travel & . . ~ram taking the place of a network frame. In this view, da- 1'

guaran t om ~•ginal source to ultunate destination, and the layering principle

0
J
ees tbat the ultimate destination receives exactly the datagram that the original
so~ sent•. On the other hand, we know that the datagram header c~ntains fields, like
counter, that change each time the datagram passes through a router.
'j
a •oe to ,.•
Thus, the ultimate destination will not receive exactly the same datagram as the source ll
~ent. We conclud e that although most of the datagram stays intact as it passes across an
I

mteme ~ the layering principle only applies to datagrams across single machine
transfers. To be accurate~ we should not view the Internet layer as providing end-to-end
service.
'

I .i
11.8 Laye ring In The Pr~sence•Of·Network Subs tructu re
~ecall from Chapte r 2 that some wide area networks contain multiple packet
switches. For example, a WAN can consist of routers that connect to a local network at
a
' each site as well as to oth~r routers using leased serial lines. When a router receives
clatawam, it either delivers the datag~am to .its destination on the local network, or
transfe rs the datagra m across a serial li~e to· ane~er
,,,,, touter. The question arises: ''H~w
do the protoco ls used on serial l~qe~ fit into the TCP/IP layering scheme ?'' The answer
depend s on how the designe r views the serial lip.e interconnections.
From the perspective of IP, the set of point-to-point connec tions among routers can
either function like a set of independent .physical networks, or they can functio n collec-
tively like a single physical network. In· the first ~ase, each physica l link is treated ex-
actly like any other networ k in t;he ,internet. The link is assigne d a unique network
number, and the two hosts that share the link each have a uriique IP address assigned
for their connecµont. Routes are ad<;fed to the IP routiqg table as they would be for any
other network. A new softwar~ module is added at the network interface layer to con-
trol the new link hardware, but no substantial changes are made to the layering scheme.
The main disadvantage of the independent network approach is that it proliferates net-
work numbers (one for each connection between two machines) and causes routing
t'flbles to be, larger than necessary. ·Both $!._riql Line IP (SLIP) and the Point to Point
Protocol (PPP) treat each serial link as a separate network.
-- '
The second approach to accommodating point-to-point connections avoids assign-
l
ing· ·multiple IP addresses to the physical wires. Instead, it treats all the connections col-
lectively as a single, independent IP network with its .own frame format, hardware ad-
dressing scheme, and data link protocols. Routers that use the second approach need
only one IP network number for all point-to-point connections.
Using the single network approach means extending the protocol layering scheme
to add a new intranetwork routing layer between the network interface layer and the

tlbe only exception arises when using the anonymous network scheme described in Chapter 10; leaving
the link unnumbered does not change the layering.
 190

. Procowl Layering Chap. 11

hardware devices. For ma"h'· - . •
,
1aye~.•seern~ \II ines with only on •
,unnecessary. To see wh . . e po1nt-to-p~1nt connecti~n, an. additional
phys.1cal p<nnt-to-point conn ct. y it ts needed, consider a machine with several
face fayer is divided. into m:iti •~ns, llnd recall from Figure 11.2 how the network inter-
need to add otre ·t1etwork . t rt'iP e software modules that each control one network. We
face must control rtlulti ~: : ~e for tbe ~ew point-to-point n~ork, but the new inter-
the new int ti . P ar ware devices. Furthermore, given a datagram to send,
Fi'Ol·,.-e t
Serhace tnUS choo~e ~e correct link over which the datagram should be sent.
e>-- 11 • s ows the organ1zat1on.
shou::e Internet layer ,sotm,~ passes . to the network interface all. datagrams that
.. , . ~~nt on any of the p~1nt-to-po1nt connections. The network interface passes
tbem \o the_ thttilhet routing module that must further distinguish among multiple physi-
cal ·Ce~necbons al)d route the datagram across the correct one.
The p~o~[Link] designs the intranet routing software determines exactly how
the software choos~s a physical link. Usually, the algorithm relies on ari intranet rout-
ing table. The intranet routing table is analogous to the internet routing table in that it
specifies a mapping of destination address to route. The table contains pairs of entries,
(0, L}, where D is a destiriation host address and L specifies the physical line used to
reach that destination.

Conceptual Layer SofJware· Organization

, '

l . Prot~ol 1 Protocol 2 Protocol 3

Transport
•
I
'

IP Module
• I

' Internet

Network
I
Interface Interface 1 Interface 2
'

Point-To-Point
'
Intranet (Intranet)

(a) (b)

Figure 11.8 (a) Conceptual position of an ~tranet protocol for point-to-point

connections when IP treats them as a single IP network, and (b)
detailed diagram of corresponding software modules. Each ar-
row corresponds to one physical device.

The difference between an internet routing table and an •intranet routing table is
that intranet routing tables are quite small. They only contain routing information for
hosts directly attached to the point-to-point network. The reason is simple: the Internet
layer maps an arbitrary destination address to a specific router address before passing
r
Sec. 11.8 • I Layering lb The p
{. I I l"eSence or Network Su~ture
the· datagratn t . 1s1
• • '' o a network interface Th .
among ~achmes . enetwork.
on a single poinHo~point as ed only to distinguish
mtranet laye7is
, '. , I •

• ' . I ' I

·11.,9 Two Im"0 rta·. . . .

• , .,,. 1
• ' • ,I \
•. 1'1_ • )1 I ',
nt Boundanes In The TCPnP Model
•

' :,. ·The (i<)Dceptµa1 protocol layeri o • · 1 • .. .

a ::proto~Lad~ ~ ~und th. n~ me ud~sitwo ,boundaries that may not ~e obvious: .
•pperatini~:Ystem bound .. at separates h•gh-level and low_-level addressing, and an
,• . , aty that separ~es the system from apphcation programs.
1 91
• t • • Hi9h-L~Yel
•' . ' . . Protocol
,. •~ Boundary
. . , Add)itft"'.:.. '

,,

I
I
,• • • I
I ,' '.

J •
• I
•, , ,Hardware ..•
,.
'' ••••..• , ••• ~···••1
'•'.,
'·"••1•••\ • ) '
I I
• • I

' rI '1 '

Flgufe. ui T~e relati~nsh!P between c~nceptual layering and the boundanes

f I' \f •
' •

•for op,erating system and high-level pro~ocol addresses.

 11.9.2 o~
• ..-vrat1ng S\lete Prot •
IQ rn Boundary oc:01 l..ayertna Chap. 11

.Figure 11 9
software th . • shows anoth ,
. at is geheral I .er important bound
not. WhtJe each imp! ary as well, the division betw
y considered Part of th
the. scheme hernentation of
follbw.
[Link]
ak
and
TcP/IP ch e system s o f t w a r e thaete.n

b ·tw s own 8 ooses how t ts

e een lower layers of. ecause they lie inside the m. o e the distinction, many

between an applicati protocol software is m h Ioperating system, passing data

lem in more deta'l on program and a transport I uc ess expensive than passing it
provide t and describes an example of' thaY_er. Chapter 20 discusses the prob-
• e interface an operating system might
.

• • I

11.10 The Oisadvanta' .' Of. . .

. • .· ge .. Layering
W~ have said that layerin is a fi •. • .·, . ,, . . .
col design. , It ,allows the design un~~enta l. idea• !h~t provides the basis for proto-
and solve .each One inde Chd g t t~.,~•vide a ~omphc11,t~d problem into subproblems
IB¥ering can be extrerriet . 1 e~t.t Unfor,tuna!e,ly, the software that results from strict
layer. It tn t . Y U ~~Cten~., As ex~j)le, [Link] the job of the transport
us accept a str~am O~,bytes,. from an [Link] program divide the stream
. . . ._,
1nlU pac ets and
.
d
k. •
h • .. •· · 1 ·~ , '

'sen ~ac pack~t across the internet. To optimize transfer, the tran-
I • , J

SPQrt l~yer sh~uld choose the larges,t: ~[Link],I~ Pa¢ket size that will allow one packet to
travel m one network frame. ,In. partic4,lil,r;tif:~e destillation machine attaches directly
to one of the same .netw~rks as the ,~o,urc'~, onl~hone physical net will involved in the be

traq~fer, so the sender c~ ·optimize :packet,1:1siz~ for that network. If the software
preserves strict layering, hc;>wever,. the···transpcirt layer cannot know how the Internet
rnodµJe will route traffic or·,.~hi<?h '·n,etworks attach directly. Furthermore, the transport
' layer will 1not understand 'the tlatagram dr fhu;n-e' formats nor 'will it be able to determine
1

·how .many octets of header·wiJI be added to a ~acket. Thus, strict layering will prevent
the transport layer from [Link] tran~fers .
Usually, implementors [Link] the strict layering scheme when building protocol
software. They allow information like route selection and network MTU to propagate
upward .• When: ~)locating buffers, they often leave space for headers that will be added
by lower layer protocols and may retain headers on incoming frames when passing them
to higher layer protocols. Such optimizations can ,make dramatic improvements in effi-
ciency while retaining the basic layered structure.

' I I

11.11 [Link] Idea Behind Multiplexing And Demultiplexing

Communication protocols use techniques of multiplexing anq demultiplexing
throughout the layered hierarchy. When sending a message, the source computer in-
cJudes extra bits that encode the message type,· originating program, and protocols used.
 T sec. 11.1 I • iT,)i~
1
•j
Deanutt1p
l\fulttplexlng And
lexhta
Eventually, itlJ. ine ssa es
a str ea m of pa Ck ets are pla ced
•. t. the receivin
int !
o
network frames for tr
193

and co mb ine d into

ma tio n-t o gu ide . g en d, the destination m ~~sferuse
Co ns ide r, Pro ces sin g. ac ane s the ex tra infor-
, an ex am ple of d .
., ernu1t1plexin sh . . ure 11.10. own in Fig
g

IP Module
ARP [Link]. RARPMoclule

' I

·[Link]~itiplexlng -ased
I 1 1 On. Ffit1n1t' typ
..
e
I
I
I
•
I II
•~ I
1,
., , I
I
j
l ,' ,f
'f
. ,,, '
,, . 'I ' ' I
I ' ,,, I·

.:;;. --:~ ..__ __.

I
_
t ,
I
, r ' . I ,. i •,

am.·e·•· rv
1
1. 1 1
1 11
J i.

1 1,

'•: '
·"r ,,,',1,
,l I I. l I! ' ' ' .
~1., ' I '
'

.
Il I
; ,: '
., I 1

'
j f } \
,,
J
II
I
, I .,
\
I/•
I )

'

type field
o~f~~i'."fra~~s based' o~· the
) fl

,,
bem uit i~l e~i llg i·,, ~f; l~~
t I
' 'F igu re 11 .10 ·.-:.. t,11 .,.
fou nd in,
1he fraI me ,bea11' 11t~
l
• , i ·, , • 1' \
1
f

I
•
,•'!•~ '"
, .£.

" w j • ) \'.J• '\ ,.• •

1, 1
; '
•
I ,· • { I
. ,I ,

• ,~ '

. . ·t
,, I l ' I '

the frame type to

~f
.. . \,/1 . '
I J

I ' ''I/

1rtetwdik 'interface lay er u~es

J I
•

so ~a re in ·~he
Th e ~g ur e ill us tra tes ho w tha t the network interface
1

'the inc o~·iQ '.g. fra me . W e say

o~oose a pr oc ed ur e to hand~e its '·,'Po n:i ~e su ch a choice possible, software in ~p~e~.
•'on
demaltiplexes the fra me based
I

transmission. Th us , each
1

fra me typ e •fie ld be for e

.the so ur ce mac~ine ~u st set
the
ify frame contents.
software mo~ule that sends frapies uses Jhe type field to spec J:l.,.
t every protocol layer. For exm
1
•

de mu ltiplex .in~ oc cu r at alm o~

Multiplexing and es ·[Link] an d pa sses those frames that con-
e de mu ltiplex
.~pie, after the network. interfac H> so ftw are ex tracts the datagram and demu\ti- '
d~ le, the
tain IP dat~grams to the IP md proto co l.· Fig ure 11 .11 demo nstrates dem~ltiplex-
port
plexes further· based on the trans
ing at the Internet layer.
 cal ,

Plotooo1 LaJerinc Claap.11

ICMP Proto eo,
UDP Protoc01
TCP Protocol

IP Module

Pisiure [Link] De . .
DlUltiplcxmg at the Inte
prop nate J)roced met layer. IP software chooses an ap-
col
field in th udretato handle a datagram b~sed on the proto
e a gram heade r.

T~ decide how to handle ad

/

software examines the header of a da-

I. •

' t > a ~ , Inter'net

tagram and selects a pro • \ the datagram type. In the example, the
oco1 hand ler· based
.
poss ible datagram s , e· rj .
and
TCP, which \\e wi l~ ~< :_.I\.; ;MP, which!we have already examined, and UDP,
. , examm.e 1n l_ater chapters.

11. 12 Su mm ary '·. ',.

I,
f
•
• '

L
. i ..
l
!
Protocols ar~ the 'standards that spedify how data . is represented when being
I
specify how the transfer occurs
• transferred ,from one machine to another. !Protocols
passed. To simplify protocoi
ho~ errors '!1'e detected1. and how ackn?wledgements are segregated into subproblems
desi gn and 1mplementat1on, communication prob
lems are I
ned a separate protocol. •f
that, can [Link] independently. Each subproblem is assig a conceptual framework
The idea· of layering is' fundamental because it provides
les one part of the cQmmunica-
for protocol design. In a layered model? e3.9~ layer hand Protocots follow tbe layering
I
tion problem and •usually correspond~· to· C>tie protocol.
layer n on the destination
principle, which states that the software iqtplementing
layer n on the source machine
machine receives exactly what the software implementing
sends. well as the older ISO 7-layer
We examined the 5-layer Internet reference model as
only a conceptual frame-
reference model. In both cases, the layering mooel provides the ISO reference model
s follow
work fur protocol software. The ITU X.25 protocol
offered by a commercial util-
and• provide an example iof reliable communication service
a different ~aye~g sche~~- .
ity, ~bile the TCP/IP protocols provide an ex~ pl~ of demultiplexmg to dtstingwsh
Jn practice, protocol software uses multiplexmg and
protocol software more complex
among multiple protocols within a given layer, making
than the layering model suggests.

• I
 .
I
- •.
-~~--:
,
/
I

~ '

~ - •
_/ j ~
V ,..,,._\ v~
-t& /
aJ1 ~'\A
,,Q,
C-

U4Q_.. a_
0- I d
1-
~;;;;:=
'
..
~- CL
--..
--..r
-~A

· ~
-
~

.
~,

'
~f-'- c S > .. _ ~
. L$AA : -
• ~o..cltik I

,
(V"-,,
:- r ,t L - t- .
4
\.
_

• ~v4Q.
;..
'~

--
~ "'"""~

I
 I,

, I r

\
\

L~L
L~l
L~l

*i~.~~ •
---
 J'

/ L·

lP M ~

• \f \__,
r I

7 (
:} . .
·,.,
\
.
r'\
'

6 /
I
j

.
; ' \
I

5· '\s~· l
\v.- ]

\
I