CHAPTER

5THE

NETWORK

LAYERNetwork

Layer

Design

Issues（網(wǎng)絡(luò)層設(shè)計概述）Routing

Algorithms

(路由算法)Congestion

Control

Algorithms

（擁塞控制算法）Quality

of

Service(服務(wù)質(zhì)量)Internetworking(網(wǎng)絡(luò)互連)The

Network

Layer

in

the

Internet(互聯(lián)網(wǎng)中的網(wǎng)絡(luò)層）15.1

NETWORK

LAYER

DESIGN

ISUESStore-and-Forward

Packet

Switching(

轉(zhuǎn)發(fā)分組交換)Services

Provided

to

the

Transport

Layer（為傳輸層提供的服務(wù)）Implementation

of

Connectionless

Service（無連接服務(wù)的實現(xiàn)）Implementation

of

Connection-Oriented

Service（面向連接服務(wù)的實現(xiàn)）Comparison

of

Virtual-Circuit and

DatagramSubnets（虛電路或數(shù)據(jù)報子網(wǎng)的比較）25.1.1

Stored-and-forward

packet

switchingA

host

with

a

packet

to

send

transmits

it

to

the

nearest

router.The

packet

is

received,

verified,

and

stored.Then

it

is

forwarded

tothe next

router.This

step

can

be

repeated

many

times.Finally

the

packet

reaches

the

destination

host.(

轉(zhuǎn)發(fā)分組交換)35.1.2 Services

provided

to

the

transport

layerThe

network

layer

services

should

been

designed

with

thefollowing

goals

in

mind.The

services

should

be

independent

of

the

routertechnology.The

transport

layer

should

be

shielded

from

thenumber,

type,

and

topology

of

the

routers

present.The

network

addresses

made

available

to

the

transportlayer

should

use

a

uniform

numbering

plan,

even

crossLANs

and

WANs.Two

types

of

network

layer

services:–

Connection-oriented

service

vs

connected-less

servic4

e5.1.2

Services

provided

to

the

transport

layerTwo-type

of

services:Connection-less

services(無連接服務(wù)?Datagram數(shù)據(jù)報

):40+

years

of

experience

with

the

computernetwork,

unreliable

internet,

hosts ng

errorcontrol

and

flow

control.Connection-oriented

services（面向連接服務(wù)Virtual

Circuit虛電路）:100+

years

of

experience

with

the

worldwideephone

system,

quality

of

service.→

Connection-less

+

connection-oriented

services.55.1.2

Services

provided

to

the

transport

layerImplementation of

connectionless

servicesNo

advance

setup

isneeded.Packets

are

injected

into

the

subnet

individually

and

routedindependently

of

each

other.The

packets

are

frequently

called

datagrams

(in ogy

withegrams)

and

the

subnet

is

calle

tagram

subnet.Implementation of

connection-oriented

servicesA

path

from

the

source

router

to

the

destination

router

mustbe

established

before

any

data

packets

can

be

sent.This

connection

is

called

a

VC

(virtual

circuit),

similar

tophysical

circuits

set

up

by

the ephone

system,The

subnet

is

called

avirtual-circuit

subnet.65.1.3

Implementation

of

Connectionless

Service785.1.3

Implementation

of

Connectionless

ServiceP1on

H1

→

P2

on

H2P1:application

layer

→

H1:

transport

layerH1:transport

layer

→

H1:

network

layer

(disassemble)H1:network

layer

→

H2:

network

layerThe

routes

for

packets

1,2,3H1

→

A

→

C

→

E

→

F

→

H2The

routes

for

packets

4H1

→

A

→

B

→

D

→

E

→

F

→

H2H2:

network

layerH2:

transport

layer→

H2:

transport

layer

(assemble)→

P2:

application

layer5.1.4

Implementation

of

Connection-Oriented

Service95.1.5

Comparison

of

Virtual-Circuit

and

DatagramSubnets10115.2

ROUTING

ALGORITHMS（路由選擇算法）The

Optimality

Principle(最優(yōu)化原則)Shortest

Path

Routing(最短路徑路由)Flooding(擴散路由)Distance

Vector

Routing(距離向量路由)Link

State

Routing(鏈路狀態(tài)路由)Hierarchical

Routing(分層路由)Broadcast

Routing(廣播路由)Multicast

Routing(多點傳送路由)Anycast

Routing(任意路由)Routing

for

Mobile

Hosts(移動主機的路由)Routing

in

Ad

Hoc

Networks(特定網(wǎng)絡(luò)的路由)Routing

Algorithms:

IntroductionA

router

performs

two

tasks:ing

packet

according

to

the

routing

tableTo

forward

the(Forwarding)To

fill

in

and

update

the

routing

table

(Routing)Desirable

properties

in

a

routing

algorithm:Correctness

(正確性),

simplicity

(簡單性):

no

commentRobustness(健壯性):The

routing

algorithm

should

copewithchanges

in

the

topology

and

traffic

without

requiring

all

processesin

all

hosts

to

be

aborted

and

the

network

to

be

rebooted

everytime

some

router

crashes.Stability(穩(wěn)定性):

A

stable

algorithm

reaches

equilibrium

andstays

there.–

Fairness

(公平性),

Efficiency

(高效率):

between

fairnessand

efficiency.12Routing

Algorithms:

IntroductionClasses

of

routing

algorithmsNonadaptive

algorithms(非自適應(yīng)算法)do

not

basetheir

routing

decisions

on

measurements

or

estimates

ofthe

current

traffic

and

topology.Instead,the

choice

ofthe

route

to

use

to

get

from

I

to

J(for

all

I.and

J)iscomputed

in

advance,

off-line,and

downloaded

to

therouters

when

the

network

is

booted.Adaptive

algorithms(自適應(yīng)算法),in

contrast,

changetheir

routing

decisions

to

reflect

changes

in

the

topology,and

usually

the

traffic

as

well.They

differ

inwhere

they

get

their

information,when

they

change

the

routes,

andwhat

metric

is

used

for

optimization.13145.2.1Routing

Algorithms:

Shortest

path

routingTo

build

a

graph

of

the

subnet,with

each

graph

node

representing

a

router

andeach

graph

edge

representing

a

communication

line.To

choose

a

route

between

a

given

pair

of

routers,

thealgorithm

just

finds

the

shortest

path

between

them.How

to

measure

path

length:Hops(結(jié)點數(shù)量),Physical

distance(物理距離),Bandwidth(帶寬),traffic(流量),cost(費用),measured

delay(測量延遲),mean

queue

length(平均隊列長度)andOther

factors

or

combinations

of

these

factors.155.2.3

Routing

Algorithms:

FloodingFlooding（擴散法）:

Every ing

packet

is

sent

out

onevery

outgoing

line

except

the

one

it

arrived

on.How

to

damp

the

flooding

process:One

is

to

have

a

hop

counter

contained

in

the

header

ofeach

packet,

which

is

decremented

at

each

hop.The

other

is

to

keep

track

of

which

packets

have

beenflooded,

to

avoid

sending

them

out

a

second

time.A

variation

of

flooding

that

is

slightly

more

practical

isselective

flooding.Flooding

is

not

practical

in

most

applications,

but

it

doeshave

some

uses

such

as military

applications.165.2.4Routing

Algorithms:

Distance

vector

routingDistance

vector

routing（距離矢量路由）(也叫Bellman-Ford

routing):Each

router

maintains

a

vector

or

table

givingthe

best

known

distance

to

each

destination

andwhich

line

to

use

to

get

there.These

tables

are

updated

by

exchanging

informationwith

the

neighbors.

To

update

these

tablesMeasure

its

distance

to

its

neighborsReceive

the

vectors

from

its

neighborsCompute

its

own

new

vector.17Routing

Algorithms:

Distance

vector

routingAssumethat

delay

is

used

as

a

metric(度量)and

that

the

router

knowsthe

delay

to

each

of

its

neighbors.Once

every??

msec,

each

router

sends

to

each

neighbor

a

list

ofits

estimated

delays

to

each

destination.

It

also

receives

a

similarlist

from

each

neighbor.Imagineone

of

these

tables

has

just

come

in

from

neighbor

X, with

Xibeing’s

estimate

of

how

long

it

takes

to

get

to

router

i.If

the

router

knows

that

the

delay

to

X

is

m

msec,Then

it

can

reach

router

i

via

X

in

Xi

+

m

msec.By

performing

this

calculation

for

each

neighbor,

a

router

canfindout

which

estimate

seems

the

best

anduse

that

estimate

and

thecorresponding

link

in

its

new

routing

table.Note

that

the

old

routing

table

is

not

used

in

the

calculation.Routing

Algorithms:

Distance

vector

routingA

subnetInputfrom

A,I,

H,

K,and

thenewroutingtablefor

J.18Routing

Algorithms:

Distance

vector

routingThe

count-to-infinity

problem(無窮計數(shù)問題):It

reacts

rapidly

to

good

news,but

leisurely

to

bad

news.1920Routing

Algorithms:

Distance

vector

routingTwo

main

applications:RIP

protocolBGP

protocol215.2.5

Routing

Algorithms:

Link

state

routingProblems

with

distance

vector

routingThe

delay

metric

was

queue

length,

thus

it

did

nottake

line

bandwidth

into

account

when

choosingroutes.The

algorithm

often

took

too

long

to

converge.→

Link

state

routing:

Each

router

must

do

the

following:Discover

its

neighbors,

learn

their

network

address.Set

the

distance

or

cost

metric

to

each

of

its

neighbors.Construct

a

packet

containing

all

it

has

just

learned.Send

this

packet

to

and

receive

packets

from

all

otherrouters.Compute

the

shortest

path

to

every

other

router.Routing

Algorithms:

Link

state

routingStep

1: Learning

about

the

neighborsO

packet

on

eachOne

router

sends

a

specialpoint-to-pointline.The

router

on

the

other

end

is

expected

to

sendback

a

reply ling

who

it

is.These

names

must

be

globally

unique.→

The

info

about

the

neighbors

can

be

found

out.22Routing

Algorithms:

Link

state

routingDesignated

Router(指定路由器)andBackup

Designated

Router(后備指定路由器)Nine

routers

and

a

LAN.A

graph

model

of

(a)23Routing

Algorithms:

Link

state

routingStep

2:

Measuring

line

costTo

determine

the

delay

is

to

send

over

the

line

aspecial

ECHO

packet

that

the

other

side

is

requiredto

send

back

immedia

y.By

measuring

the

round-trip

time

and

dividing

it

by

2,the

sending

router

can

get

a

reasonable

estimate

ofthe

delay.Average

delay

value

can

be

better.24Routing

Algorithms:

Link

state

routingSteps

3:

Building

link

state

packetsEach

router

builds

a

packet

containing

all

the

data.The

packet

starts

with

the

identity

of

the

sender,followed

by

a

sequence

number

and

age

and

a

list

ofneighbors.To

build

them

is

easy, but

when

to

build

them

isdifficult

to

determine:To

build

them

periodically

(at

regular

intervals)To

build

them

when

some

significant

event

occurs,such

as

a

line

or

neighb oing

down

or

comingback

up

again

or

changing

its

propertiesappreciably.25Routing

Algorithms:

Link

state

routingA

subnet.The

link

state

packets

for

this

subnet.26Routing

Algorithms:

Link

state

routingSteps

4:

Distributing

the

link

state

packetsTo

use

flooding

to

distribute

the

link

state

packetsTo

keep

track

of

all

the

(source

router,

sequence)pairs

they

see.To

include of

each

packet

after

the

sequencenumber

and

decrement

it

once

per

second.27Routing

Algorithms:

Link

state

routingSteps

5:

Computing

the

new

routesOnce

a

router

has

accumulated

a

full

set

of

link

statepackets,

it

can

construct

the

entire

subnet

graphbecause

every

link

is

represented.Dijkstra’s

algorithm

can

be

run

locally

to

constructthe

shortest

path

to

all

possible

destinations.

Theresults

of

this

algorithm

can

be

installed

in

therouting

tables.Some

applicationsIS-IS

(Intermediate

System

–

Intermediate

System)OSPF

(Open

Shortest

Path

)28295.2.6

Routing

Algorithms:

Hierarchical

routingAs

networks

grow

in

size,

the

router

routing

tables

growproportional,

and

so

do

router

memory

and

computingpower.Two

level

routing:

Every

router

knowsall

the

details

about

how

to

route

packets

todestinations

within

its

own

regionbut

knows

nothing

about

the

internal

structure

ofother

regions.Multiple-levelrouting:Regions

?

clusters

?zones

?groups

?…Routing

Algorithms:

Hierarchical

routingReduction

of

routing

tables30315.2.7

Routing

Algorithms:

Broadcast

routingBroadcasting:

to

send

a

packet

to

all

destinationssimultaneously.The

source

simply

sends

a

distinct

packet

to

everydestination.Flooding.Multidestination

routing

(each

packet

containseither

a

list

of

destinations

or

a

bit

map

indicatingthe

desired

destinations.)

(One

router

pays

full

fareand

the

rest

ride

free.)To

make

use

of

the

sink

tree

for

the

router

initiatingthe

broadcast.Reverse

path

forwarding.Routing

Algorithms:

Broadcast

routingReverse

path

forwarding:When

a

broadcast

packet

arrives

at

a

router,

therouter

checks

to

see

if

the

packet

arrived

on

theline

that

is

normally

usedforsending

packets

tothe

source

of

the

packets.If

so,

there

is

an

excellent

chance

that

thebroadcast

packet

itself

followed

the

best

routefrom

the

router

and

is

therefore

the copy

toarrive

at

the

router.

Then

the

router

forwardscopies

of

it

onto

all

lines

except

the

one

it

arrivedon.If

no,

it

is

discarded

as

a

duplicate.32Routing

Algorithms:

Broadcast

routingReverse

path

forwarding.A

subnet.a

Sink

tree.The

tree

built

by

reverse

path

forwarding.335.2.8

Routing

Algorithms:

Multicast

routingMulticasting(組播):to

send

messages

to

well-definedgroups

that

are

numerically

large

in

size

but

smallcompared

to

the

network

as

whole.Group

management:

Someway

is

needed

to

createanddestroy

groups,

and

to

allow

processes

to

join

and

leavegroups.Computing

a

spanning

tree

covering

all

other

routers.Multicast

routing

is

to

prune

the

spanning

tree.When

a

process

sends

a

multicast

packet

to

a

group,The

router

examines

its

spanning

treeand

prunes

it,

removing

all

the

lines

that

donot

leadto

hosts

that

are

members

of

the

group.34Routing

Algorithms:

Multicast

routing(a)

Anetwork. (b)

A

spanning

tree

for

the

leftmost

router.A

multicast

tree

fA

multicast

tree

froup

1.roup

2.355.2.9

Routing

Algorithms:

Anycast

routingIn

anycast,

a

packet

is

delivered

to

the

nearest

memberof

a

group.

Schemes

that

find

these

paths

are

calledanycast

routing.36Routing

Algorithms:

Routing

the

mobile

hosts37385.3

CONGESTION

CONTROL

ALGORITHMS(擁塞控制算法)Approaches

to

Congestion

ControlTraffic-Aware

RoutingAdmission

ControlTraffic

ThrottlingLoad

SheddingCongestion

Control

Algorithms:

IntroductionCongestion:

When

too

much

traffic

is

offered,congestion

sets

in

and

performance

degrades

sharply.3940Congestion

Control

Algorithms:

IntroductionCongestion

causes(擁塞原因):Burst

packets

on

one

output

line.

If

all

of

a

sudden,

streams

of

packets

begin

arrivingon

three

or

four

input

lines

and

all

need

the

sameoutput

line,

a

queue

will

build

up.

If

there

isinsufficient

memory

to

hold

all

of

them,

packets

willbe

lost.

Adding

more

memory

will

not

help

a

littlebut

not

much.Mismatch

between

parts

of

the

systemInsufficient

memory.Slow

CPU.Low

bandwidth.Congestion

Control

Algorithms:

IntroductionCongestion

control

and

flow

controlDifferences:Congestion

control:

Subnet,

global.A

store-and-forward

network

with

1-Mbps

linesand

1000

large

computers,

half

of

which

aretrying

to

transfer

files

at

100kbps

to

the

other

half.Flow

control:

end-to-end

traffic,

local.An

example:

1000Gbps(

puter)

→1Gbps

(

al

computer).Similarity:

Both

congestion

control

and

flow

control

canl

the

sender

to

slow

down–

because

the

receiver

cannot

handle

the

load

orbecause

the

network

cannot

handle

it.415.3.1

Congestion

Control

Algorithms:Approaches

to

Congestion

ControlTime

scales

of

approaches

to

congestion

controlNetwork

provisioning(網(wǎng)絡(luò)供應(yīng))：monthsTraffic-aware

routing:

hoursAdmission

control:

minutesTraffic

throttling:

secondsLoad

shedding:

seconds42435.4

QUALITY

OF

SERVICE(服務(wù)質(zhì)量)Requirements

of

QoSTechniques

for

QoSIntegrated

ServicesDifferentiated

ServicesLabel

Switching

and

MPLS44Quality

Of

Service:

RequirementsA

flow

is

a

stream

of

packets

from

a

source

to

adestination.In

a

connection-oriented

network,

all

the

packetsbelonging

to

a

flow

follow

the

same

route;In

a

connection-less network,

they

may

followdifferent

routes.The

needs

(or

requirements)

of

each

flow

can

becharacterized

by

four

primary

parameters:Reliability(可靠性),Delay(延遲),Jitter(抖動),Bandwidth(帶寬).5.4.1

Quality

Of

Service:

RequirementsHow

stringent

the

quality-of-service

requirements

are.455.4.2

Quality

Of

Service:

Traffic

sha(a)

Sha

packets.

(b)

A

leaky

bucket.

(c)

A

token

bucketLeaky

bucket

algorithmToken

bucket

algorithm46475.4.3

Quality

Of

Service:

Packet

SchedulingResource

reservationSuppose

there

is

a

specific

route

for

a

flow,

ites

possible

to

reserve

resources

along

theroute

to

make

sure

the

needed

capacity

is

available.What

resources

to

reserve?BandwidthBuffer

spaceCPU

cyclesQuality

Of

Service:

Packet

SchedulingRound-robin

Fair

Queuing485.4.4

Quality

Of

Service:

Admission

ControlAn

example

flow

specification49Quality

Of

Service:

Admission

ControlBandwidth

and

delay

guarantees

with

token

bucketsand

WFQ(Weighted

Fair

Queueing).50515.4.5 Quality

Of

Service:

Integrated

servicesHow

to

stream

multimedia?Integrated

services

(Flow-based

algorithms)Differential

services

(Class-basedalgorithms)How

about

having

the

senders

reserve

bandwidth

inadvance?

Too

many

destinationsRSVP(Resource

reSerVation

Protocol資源預(yù)留協(xié)議):toallow

multiple

senders

to

transmit

to

multiple

groups

ofreceivers,permits

individual

receivers

to

switchchannels

freely,and

optimizes

bandwidth

use

while

atthe

same

time

eliminating

congestion.Quality

Of

Service:

Integrated

services(a)

Host

3

requests

a

channel

to

host

1. (b)

Host

3

then

requestsasecond

channel,

to

host

2. (c)

Host

5

requests

a

channel

to

host

1.525.4.6

Quality

Of

Service:

Differential

servicesExpedited(暢通的,迅速的)packets

experience

atraffic-free

network.(RFC

3246)53Quality

Of

Service:

Differential

servicesAssured

Forwarding(確定推進)(RFC2597)545.4.7

Quality

Of

Service:Label

switching

and

MPLSTransmitting

a

TCP

segment

using

IP,

MPLS,

and

PPP.55Quality

Of

Service:Label

switching

and

MPLSForwarding

an

IP

packet

through

an

MPLS

network56575.5

INTERNETWORKING(網(wǎng)絡(luò)互連)How

Networks

DifferHow

Networks

Can

Be

ConnectedTunnelingInternetwork

RoutingPacket

Fragmentation5.5.1

Internetworking:

How

Networks

DifferSome

of

the

many

ways

networks

can

differ585.5.2

Internetworking:

How

Networks

Can

BeConnectedWhich

device

is

in

which

layer.Frames,

packets,

and

headers.59Internetworking:How

Networks

Can

Be

ConnectedA

packet

crossing

different

networks.Network

and

link

layer

protocol

processing.605.5.3

Internetworking:Tunneling(隧道)Tunneling

a

packet

from

Paris

to

London.61Internetworking:

TunnelingTunneling

a

car

from

France

to

England.625.5.4

Internetworking:

Internetwork

RoutingTwo-level

routingateway

protocolateway

protocolInternet

routing:

ExteriIntranet

routing:

InteriTo

route

a

packetA

typical

internet

packet

starts

out

on

its

LAN

addressed

to

thelocal

multiprotocol

router.After

it

gets

there,

the

network

layer

code

decides

whichmultiprotocol

router

to

forward

the

packet

to,

using

its

own

routingtables.Direct

forwarding

using

native

network

protocolTunneling

using

the

intervening

network

protocolThis

process

repeats

until

the

packet

reaches

the

desinationnetwork.AS

and

BGP635.5.5

Internetworking:

FragmentationEach

network

imposes

some um

size

on

itspackets.

These

limits

have

various

causes,

among

them:Hardware

(e.g.,

the

width

of

a

TDM

transmission

slot).Operating

system

(e.g.,

all

buffers

are

512

bytes).Protocols

(e.g.,

the

number

of

bits

in

the

packet

lengthfield).Compliance

with

some

(inter)national

standard.Desire

to

reduce

error

induced

retransmissions

tosome

level.Desire

to

prevent

one

packet

from

occupying

thechannel

too

long.64Internetworking:

FragmentationTransparent

fragmentation.

(ATM)Nontransparent

fragmentation.

(IP)65Internetworking:

FragmentationPath

MTU

Discovery665.6

THE

NETWORK

LAYERINTHEINTERNETThe

IPv4

ProtocolIP

AddressesThe

IPv6

ProtocolInternet

Control

ProtocolsOSPF

–

The

Interiateway

Routing

Protocolateway

Routing

ProtocolBGP –

The

ExteriInternet

MulticastingMobile

IP67The

Network

Layer

in

the

Internet:The

Internet:

Collections

of

Subnetworks

or

ASes685.6.1

The

Network

Layer

in

the

Internet:The

IPv4

ProtocolAn

IP

datagram

consists

of

a

header

part

and a

text

part.The

IPv4

(Internet

Protocol)

header.69The

Network

Layer

in

the

Internet:The

IPv4

Protocol:

Header

fieldsVersion:

to

keep

track

of

which

version

of

the

protocol

the

datagrambelongs

to.IHL:

to l

how

long

the

header

is,

in

32-bit

words.

5

<=

IHL

<=

15.Differentiated

service:Type

ofservice服務(wù)類型:3

for

priority,3

for

D,T,and

R,

and

2unused.Differentiated

services區(qū)分服務(wù):Defined

by

IETF

in

1998，6

forservice

class,

2

for

congestion.um

length

isTotal

length:

the

length

of

header

and

data.

The65,535

bytes.Identification:

the

ID

of the

datagram.DF:Don‘t

Fragment

只有當(dāng)DF=0時才允許分段(最小MTU=576，即假設(shè)512（傳輸層）+60（最大IP頭）+4（冗余）)MF:

More

Fragment

(=1

for

all

fragments

except

last

one=0

for

last

fragment)70The

Network

Layer

in

the

Internet:The

IPv4

Protocol:

Header

fieldsFragment

offset:

to l

where

in

the

current

datagram

thisfragment

belongs.

All

fragments

except

the

last

one

in

adatagram

must

be

a

multiple

of

8

bytes,

the

elementaryfragment

unit.Time

to

Live:

a

counter

used

to

limit

packet

lifetimes.Protocol:

which

transport

process

to

give

this

datagram

to.(

)Header

checksum:

to

verify

the

header

onlySource

and

destination

address:

to

indicate

the

networknumber

and

host

number.Options71725.6.2

The

Network

Layer

in

theInternet:IP

AddressesEvery

Internet

interface

has

an

IP

address,

whichencodes

its

network

number

and

host

number.

Thecombination

is

unique:

no

two

interfaces

have

thesame

IP

address.All

IP

addresses

are

32

bits

long

and

are

used

in

thesource

address

and

Destination

address

fields

of

IPpacketsIP

addressingPrefixesSubnets

(division),CIDR(mergement),Classful

and

Special

AddressingNATThe

Network

Layer

in

the

Internet:IP

Addresses:

PrefixesAn

IP

prefix.73The

Network

Layer

in

the

Internet:IP

Addresses:

PrefixesSplitting

an

IP

prefix

into

separate

networks

withsubnetting.74The

Network

Layer

in

the

Internet:IP

Addresses:

CIDRA

set

of

IP

address

assignments75The

Network

Layer

in

the

Internet:IP

Addresses:

CIDRAggregation

of

IP

prefixessupernet（超網(wǎng)）vs

subnet76The

Network

Layer

in

the

Internet:IP

Addresses:

CIDRLongest

matching

prefix

routing

at

the

New

Yorkrouter.77The

Network

Layer

in

the

Internet:IP

Addresses:CIDR試題：A

router

has

the

following(CIDR)

entries

in

its

routing

table:Address/mask

Interface Next

hop21:55:4878/22/22/23defaultInterface

0Interface

1router

1router

2for

each

of

the

following

IP

addresses,

which

one

does

the

router

selectfor

the

next

hop

if

a

packet

with

that

address

arrives?

(5

bonus)(1)

0(4)

(2)

4

(3)

(5)

The

Network

Layer

in

t