data_20062pp_1161158.. - Communication Systems Laboratory

1

프로토콜

기술과

성능분석


발

표

:
김재석


발표일
: 2006
년

10
월

11
일

Part Ⅸ : Classless And Sbunet Address Extensions(CIDR)


Part Ⅹ : Protocol Layering

2

PART Ⅸ

Classless

And

Subnet Addresses Extention

Part
Ⅳ

䍉䑒

3


Introduction

Four Extension of the IP Address


Proxy Arp


Subnet Addressing


Anonymous Point
-
to
-
Point Networks


Classless Addressing

Part
Ⅳ

䍉䑒

4


Review Of Relevant Facts

Original address scheme (IPv4)


Divided into two parts : Network + Host


Unique network address (Each Physical Network)


Prefix (Each host on a network)


Advantage


reduce the size of routing table


keep one routing entry per network


Classful addressing



Class A
–

8bit network portion


Class B
–

16bit network portion


Class C
–

24bit network portion

Part
Ⅳ

䍉䑒

5


Minimizing Network Numbers

Reduce the number of network prefixes used


Weakness of Original IP addressing scheme


Growth ( From Mainframe Computer environment)


doubled in size every nine to fifteen months


Management overhead, Huge size of routing table, exhaustion


“
How can the technology accommodate growth without
abandoning the original calssful addressing scheme?
´


Transparent routers


unnumbered point
-
to
-
point links


proxy ARP


subnet addressing

Part
Ⅳ

䍉䑒

6


Proxy ARP

Single network prefix is used for two physical networks

H
1

H
2

H
3

H
4

H
5

H
6

Main Network

Hidden Network

Router running proxy ARP


Two network share single IP network


‘
R
’

keeps the location of hosts completed hidden


Hosts communicate


As if they are directly connected on a single network


TRUST


ARP based on COOPERTION & LEGITIMATION


spoofing warning Implementation

R

Part
Ⅳ

䍉䑒

7


Proxy ARP Cont

Host H
1

needs to communicate with host H
4

NO

OBJ

Action

비고


1

H1

Broadcast ARP REQ

To H4

2

R

Capture ARP REQ, Decide the LOC of H4

3

R


H1

Responds to the ARP REQ (R
’
s Physical Address)

4

H1

Receive ARP Response

Install the mapping ARP Table

5

H1


R(H4)

Sends Datagrams(Use above mapping to R)

6

R

Forward datagrams to H4

H
1

H
2

H
3

H
4

H
5

H
6

R

Part
Ⅳ

䍉䑒

8


Subnet Addressing

Subnet routing, Subnet forwarding, Subnetting

Standard, most general, most widely used

H
1

H
2

R

H
3

H
4

Network 128.10.1.0

Network 128.10.2.0

128.10.1.1

128.10.1.2

128.10.2.1

128.10.2.2

Rest Of The
Internet

(Another AS)

All Traffic to

128.10.0.0

Accepts all traffic for net 128.10.0.0

Chooses a physical network based on the third octet of the address

Part
Ⅳ

䍉䑒

9


Subnet Addressing Cont

Network Portion + Host Portion


Instead of Prefix and suffix


Internet Part

Local Part

Internet Part

Physical

Network

Host

Site, possibly with multiple
physical networks

Physical network and hosts at that
site


Hierarchical addressing


Leads to hierarchical routing



First level (Other Autonomous System) : use the first two octet



Next level : uses an additional octet



Lowest Level
…
.



Telephone System


Scalability

Part
Ⅳ

䍉䑒

10


Flexibility in Subnet Address Assignment

Allow Site Flexibility : assigning subnet address

Network 1

Part
Ⅳ

䍉䑒

To rest of Internet

Network 2

Network 3

Network 4

Network 5



Five Physical network



Three level

Subnet Bits

NO of Subnet

Hosts per
Subnet

0

1

65534

2

2

16382

3

6

8190

4

14

4094

5

30

2046

6

62

1022

7

126

510

Subnet Bits

NO of Subnet

Hosts per
Subnet

8

254

254

9

510

126

10

1022

62

11

2046

30

12

4094

14

13

8190

6

14

16382

2

Class B Subnetting

11


Variable
-
Length Subnets

Most site : Fixed
-
Length Subnets

Some Case Needs


Many networks with few hosts per networks


A few networks with many hosts



Flexibility


mixture of large and small networks


higher utilization of the address space


Address ambiguity



assigned carefully

Part
Ⅳ

䍉䑒

12


Implementation Of Subnets With Masks

Subnet Technology


either fixed or variable length


Standard : 32
-
bit Mask


Subnet prefix : set to
‘
1
’


Host prefix : set to
‘
0
’

Part
Ⅳ

䍉䑒

11111111 11111111 11111111

00000000

Identify Network

Identify Hosts


recommendation


use contiguous subnet mask


routing Table trick

11111111 11111111 00011000

01000000

13


Subnet Mask Representation

In Binary


awkward and prone to errors


Alternative representations


dotted decimal representation (
ex) 255.255.255.0
)


3
-
tuple(
{<network number>,<subnet number>,<host number>}
)


‘
-
1
’

means
“
all ones(1s)
”


ex) 255.255.255.0 {
-
1,
-
1,0}


{128.0,
-
1,0}

Part
Ⅳ

䍉䑒

14


forwarding In The Presence Of Subnets

Subnet forwarding, Subnet routing


Modified standard IP forwarding algorithm

Part
Ⅳ

䍉䑒

乥琠N⡓畢(e琠潦 a摤牥ss 丩

H

R2

R1

Net 3(Subnet of address N)

Net 1(not a Subnet address)

H can send to either
‘
R1
¶

and
µ
R2
¶



Not Shortest Path



To activate optimal forwarding : user subnet forwarding


Modified standard IP forwarding algorithm


The subnet mask should be uniform across all networks


All machines should participate in subnet forwarding

15


The Subnet forwarding Algorithm

Used With subnet searches a table of routes

Routing table entries (network address, next hop address)


network address : Ip address of destination network

N


next hot address : address of a router to which datagrams


destined for
N

should be sent

Subnet Forwarding Algorithm


maintains additional information in the routing table


address mask : extract bits for the destination address for


comparison with the table entry


network address


Next hop address

Part
Ⅳ

䍉䑒

16


Maintenance Of Subnet Masks

How are subnet masks assigned by an administrator?


Each site is free to choose subnet masks for its networks


balance sizes of networks


Numbers of physical networks


Expected growth


Ease of maintenance


nouniform masks


flexibilities but lead to ambiguous route

Part
Ⅳ

䍉䑒

17


Broadcasting To Subnets

{network,
-
1,
-
1}


deliver a copy to all machines that have network as their networks
address even if they on separate physical networks


Reverse path forwarding (RPF)


Router can
’
t merely propagate a broadcast packet that arrives on one
interface to all interfaces that share the subnet prefix


discard the datagram unless it arrived on the inerface used to forward
to the source


{Network, subnet,
-
1)

Part
Ⅳ

䍉䑒

18


Anonymous Point
-
To
-
Point networks

To avoid assigning a prefix to each point
-
to
-
point connection.

Often applied when a pair of routers is connected with a leased
digital circuit.

Unnumbered network


no number on leased line


no host address to the router at each end


no hardware address


interface software



ignore the next hop address


arbitrary value can be used as the next
-
hop address


does not operate like shared
-
media hardware


only one possible destination
–

摯es no琠畳u 灨ys楣al 慤dress

Part
Ⅳ

䍉䑒

19


Anonymous Point
-
To
-
Point networks cont

Part
Ⅳ

䍉䑒

R2

R1

128.10.0.0

128.211.0.0

Leased serial line

128.10.2.250

128.211.0.100

TO REACH HOSTS

ON NETWORK

ROUTE TO

THIS ADDRESS

USING THIS

INTERFACE

128.10.0.0

DELIVER DIRECT

1

default

128.211.0.100

2

1 2

Routing Table of R1

The address of R2’s Ethernet connection

20


Classless Addressing And Supernetting

Attempt to conserve the IP address space


subnet addressing, unnumbered network


not enough to prevent Internet growth from exhausting


defining an entirely new Version of IP with large addresses


temporary solution to accommodate growth


Permit a network prefix to be an arbitrary length


CIDR
-

Classless Inter
-
Domain Routing


classful scheme did not divide network addresses into equal size


Class C : much smaller than demand for class B


not amenable to subnetting


Class B : would be exhausted quickly (ROADS)

Supernetting


Assign a block of class C instead of a single class

Part
Ⅳ

䍉䑒

21


Advantage of Supernetting

Clear the issue of Class B address exhausting

Disadvantage : Increasing of routing information

CIDR clear the issue of increasing of routing Information

(Network address, Count)


Network address : minimum network address of the block


Count : Entire NO of network address of the block


192.5.48.0


ex) (192.5.48.0, 3) 192.5.49.0


192.5.50.0



Part
Ⅳ

䍉䑒

22


CIDR Address Blocks And Bit Masks

ISP assign each subscriber a block of addresses

Uses 32
-
bit address bit mask

Part
Ⅳ

䍉䑒

Dotted Decimal

32
-
bit Binary Equivalent

Lowest

128.211.168.0

10000000 11010011 10101000 00000000

highest

128.211.175.255

10000000 11010011 10101111 11111111

CIDR Mask

11111111 11111111 11111000 00000000





Example of CIDR block of 2048 addresses

continuous 1 bit for prefix,

continuous 0 bit for suffix

23


Address Blocks And CIDR Notation

Identifying CIDR Block requires : address, mask

CIDR Notation (Slash Notation)


ex) 128.211.168.0/21

Part
Ⅳ

䍉䑒

CIDR
Notation

Dotted
Decimal

CIDR Notation

Dotted
Decimal

CIDR Notation

Dotted Decimal

/1

128.0.0.0

/12

255.240.0.0

/23

255.255.254.0

/2

192.0.0.0

/13

255.248.0.0

/24

255.255.255.0

/3

224.0.0.0

/14

255.252.0.0

/25

255.255.255.128

/4

240.0.0.0

/15

255.254.0.0

/26

255.255.255.192

/5

248.0.0.0

/16

255.255.0.0

/27

255.255.255.224

/6

252.0.0.0

/17

255.255.128.0

/28

255.255.255.240

/7

254.0.0.0

/18

255.255.192.0

29

255.255.255.248

/8

255.0.0.0

/19

255.255.224.0

/30

255.255.255.252

/9

255.128.0.0

/20

255.255.240.0

/31

255.255.255.254

10

255.192.0.0

/21

255.255.248.0

/32

255.255.255.255

/11

255.224.0.0

/22

255.255.252.0

Dotted decimal mask values form all possible CIDR prefixes

24


A Classless Addressing Example

Complete flexibility in allocating blocks of various sizes

ISP can assign each customer a block of an appropriate size


128.211.0.0/21 for some customer


128.211.176.212/30 for some customer

Part
Ⅳ

䍉䑒

Dotted Decimal

32
-
bit Binary Equivalent

Lowest

128.211.176.212

10000000 11010011 10110000 11010100

highest

128.211.176.215

10000000 11010011 10110000 11010111

CIDR Block with 128.211.176.211/30

Summary
-

Classless Addressing


Treat IP Address as arbitrary integers


Partitioning addresses into contiguous blocks

25


Data Structure And Algorithms for Classless Lookup

Speed
–

fundamental criterion of Algorithm and data Structure


finding next hop, making changes to values in the table

1.
Hashing And Classful Address

Glassful Addressing :

self
-
identifying, hash table works well

router extracts network portion
N
and using it hashing Key

Classless Addressing :

non self
-
identifying, hashing doesn
’
t works well

Alternative must be used

2. Searching By Mask Length

LPM (Longest Prefix Match) : subscriber
’
s address mask > ISP
’

address
mask

Iterates approaching
–

extremely slow

Default route
–

performs 31 unnecessary lookups

Part
Ⅳ

䍉䑒

26


Data Structure And Algorithms for Classless Lookup

3. Binary trie Structure

Hierarchical data structure

Variants of binary trie

Many systems use

Part
Ⅳ

䍉䑒

32
-
Bit Address

Unique Prefix

00110101 00000000 00000000 00000000

01000110 00000000 00000000 00000000

01010110 00000000 00000000 00000000

01100001 00000000 00000000 00000000

10101010 00000000 00000000 00000000

10110000 00000000 00000000 00000000

10111011 00000000 00000000 00000000

00

0100

0101

011

010

10110

10111

A set of Binary address and the corresponding set of prefixes that uniquely identify each

27


Data Structure And Algorithms for Classless Lookup

Part
Ⅳ

䍉䑒

0

0

0

0

0

0

0

1

1

1

1

1

1

1

A Binary trie for seven binary prefixes listed previous

Prefix 0101

Stop when

reaches an exterior node

no path exists for the specified prefix

28


PATRICIA And Level compressed Tries

Binary Tries Omit details relate to optimization of lookup

“
skipping
”

levels in the trie that do not distinguish among routes

Examine all bits of a destination address at once rather than extracting
bits one at a time

PATRICIA tree

allows each node to specify a value to test long with a number of
bits to skip

Level compressed trie

provides additional optimizations by eliminating one or more levels
In the trie that can be skipped along any path

Trade Off

improve search speed but require more computation



Part
Ⅳ

䍉䑒

29


CIDR Blocks Reserved For Private networks

Reserved prefixes for Private Networks

private addresses

nonroutable addresses

Routers in the internet understand that the addresses are reserved

Part
Ⅳ

䍉䑒

Prefix

Lowest Address

Highest Address

10.0.0.0/8

172.16.0.0/12

192.168.0.0/16

169.254.0.0/16

10.0.0.0

172.16.0.0

192.168.0.0

169.254.0.0

10.255.255.255.255

172.31.255.255

192.168.255.255

169.254.255.255

The prefixes reserved for use with private internets

30


Summary

Five techniques to conserve IP addresses

Transparent Router

Proxy ARP

Subnet addressing

Anonymous point
-
to
-
point connection

CIDR

Part
Ⅳ

䍉䑒

31

PART Ⅹ

Protocol Layering

Part
Ⅳ

䍉䑒

32


Introduction

Structure of the software found in hosts and routers


presents the general principle of layering


easy to understand, build and trace

Part
Ⅳ

P牯瑯c潬⁌oye物湧


The Need for Multiple Protocols

Protocol allow one to


specify or understand communication without knowing the details

Complex data communication



protocol family, protocol suite
-

require a set of cooperative protocols


Why


Hardware failure, Network Congestion


Packet Delay (Loss), Data Corruption


Data Duplication or Inverted Arrival

33


The Conceptual Layers Of Protocol Software

Part
Ⅳ

P牯瑯c潬⁌oye物湧

Sender

Layer n

…

Layer2

Layer1

Receiver

Layer n

«

Layer2

Layer1

Network

The conceptual organization of protocol software in layers

In practice, the protocol software is much more complex.

High
-
Level Protocol Layer

Internet Protocol Layer

Network Interface Layer

Protocol 1

Protocol 1

Protocol 1

Interface 1

Interface 2

Interface 3

IP Module

Conceptual protocol layering Realistic view of software organization

Multiple network interfaces below IP and multiple protocol above it

34


The Conceptual Layers Of Protocol Software Cont

Part
Ⅳ

P牯瑯c潬⁌oye物湧

Message traversing from the sender through two intermediate routers to the receiver.

Intermediate only send the datagram to the IP software layer

Sender

Other
…

IP Layer

Interface

Net 1

IP Layer

Interface

IP Layer

Interface

Receiver

Other
…

IP Layer

Interface

Net 2

Net 3

35


ISO 7
-
Layer Reference Model

Part
Ⅳ

P牯瑯c潬⁌oye物湧

The ISO 7
-
Layer reference model for protocol software

ISO
’
s Reference Model of Open System Inerconnection

ISO Model

Contains 7 Conceptual layers

Application

Presentation

Session

Transport

Network

Data Link

(Hardware Interface)


Physical Hardware

Connection

Layer Functionality

1

2

3

4

5

6


7

36


X.25 And Its Relation To The ISO Model

Recommendation of ITU, most recognized and widely used

Adopted by public data network, like a telephone system


Physical Layer


Standard for the physical interconnection


Procedures used to transfer packets


Data Link Layer
-

How data travels


Network Layer
–

Network or communication subnet layer


Defines the basic unit of transfer,


The concepts of destination addressing and forwarding


Transfer Layer
–

Provide end
-
to
-
end Reliability


Session Layer
±

Remote terminal access


Presentation Layer
-

compress text, convert graphics images into
bit stream


Application Layer
±

application programs (E
-
mail, FTP
«
.)

Part
Ⅳ

P牯瑯c潬⁌oye物湧

37


The TCP/IP 5
-
Layer Reference Model

Five conceptual layers

Part
Ⅳ

P牯瑯c潬⁌oye物湧

Application

Transport

Internet

Network Interface

Hardware

Conceptual Layer

Object Passed

Between Layers

Message or Streams

Transport Protocol Packets

IP Datagrams

Network
–
Specific Frames


Application Layer
–

Invoke application programs


Transport Layer


Provide communication from one application program to another


Regulate flow on information, Reliable transport



Internet Layer
–

handle communication from one machine to another


encapsulation, forwarding algorithm


Network Layer
–

responsible for accepting IP datagrams and


transmitting them over a specific network

38


The Protocol Layering Principle

Layered protocols are designed so that layer N at the destination
receives exactly the same object sent by layer N at the source


Allows the protocol designer


-

to focus attention on one layer at a time,


without worrying about how other layers performs

Part
Ⅳ

P牯瑯c潬⁌oye物湧

Application

Transport

Internet

Network

Interface

Application

Transport

Internet

Network

Interface

Identical
message

Identical
packet

Identical
datagram

Physical Net

Identical frame

Host A

Host B

39


Layering in a TCP/IP Internet Environment

Part
Ⅳ

P牯瑯c潬⁌oye物湧

Application

Transport

Internet

Network

Interface

Application

Transport

Internet

Network

Interface

Identical
message

Identical
packet

Host A

Host B

Internet

Network

Interface

Physical Net 1

Identical
datagram

Identical
datagram

Identical
frame

Physical Net 2

Identical
frame

Router R

The ultimate destination will not receive exactly the same datagtams
as the source sent


-

datagram header contains fields like TTL
…

The Layering principle only applies to datagrams across single
machine transfers.

40


Layering in The Presence Of Network Substructure

Part
Ⅳ

P牯瑯c潬⁌oye物湧

Transport

Internet

Network

Interface

Interanet

Protocol 1

Protocol 1

Protocol 1

Interface 1

Interface 2

Interface 3

IP Module

Point
-
to
-
Point

(Intranet)

Software Organization

Conceptual Layer

41


Two Important Boundaries In The TCP/IP Model

Part
Ⅳ

P牯瑯c潬⁌oye物湧

Protocol address boundary


high
-
level address (IP Address) :


from the internet layer upward


Application Programs


low
-
level address (Physical Address) :


network interface layer


Operating System Boundary


considered part of the operating system


and software that is not

42


Two Important Boundaries In The TCP/IP Model

Part
Ⅳ

P牯瑯c潬⁌oye物湧

Application

Conceptual Layer

Boundary

Transport

Internet

Network

Interface

Hardware

Software outside the operating system

Software inside the operating system

Only IP addresses used

Physical addresses used

43


The Disadvantage Of Layering

Part
Ⅳ

P牯瑯c潬⁌oye物湧

Strict layering can be extremely inefficient

Usually, relax the strict layering scheme

44


The Basic Idea Behind Multiplexing and Demultiflexing

Part
Ⅳ

P牯瑯c潬⁌oye物湧

䥐⁍潤畬u

A剐 M潤畬o

剁剐 M潤畬o

Demultiflexing Based

On Frame Type

Frame Arrived

Multiplexing and Demultiplexing


Communication protocol use throughout the layered hierarchy.

ICMP Protocol

UDP Protocol

TCP Protocol

IP Module

Datagram Arrived

Demultiplexing at the network interface

Demultiplexing of incoming frames based on the type field

found in the frame header

Demultiplexing at the internet layer

IP Software chooses an appropriate procedure

to handle a datagram base on the protocol type field

in the datagram header