Subnet & Classless Address Extensions

needleorderΔίκτυα και Επικοινωνίες

24 Οκτ 2013 (πριν από 4 χρόνια και 16 μέρες)

188 εμφανίσεις

1
Subnet & Classless Address
Extensions
Linda Wu
(CMPT 471 • 2003-3)
Notes-4CMPT 471 •2003-32
Content

Motivation

Transparent routers

Proxy ARP

Subnet addressing

Classless addressing
Reference: chapter 10
Notes-4CMPT 471 •2003-33
Motivation

Problem: network growth will
exhaust IPv4 address space
eventually

Solution: minimize the number of
addresses used

Avoid assigning netidwhenever possible

Share one netidamong multiple
networks: transparent router, proxy ARP,
subnet addressing

Arbitrary assignment: classless
addressing
Notes-4CMPT 471 •2003-34
Transparent Routers 
A special router T, called transparent router,
connects WAN and LAN

Other hosts and routers on the WAN do not
know T’s existence

The LAN does not have its own netid; the hosts
in the LAN are assigned unused addresses in
the WAN

T sends packets from the WAN to the
appropriate host in the LAN

T accepts packets from the LAN and routes
them across WAN to the destination
2
Notes-4CMPT 471 •2003-35
Transparent Routers
(cont.)

Advantages

One netidis shared by multiple networks:
fewer network addresses are required

Support load balancing

Disadvantages

Only work with WAN that has a large
address space

Do not provide all the same services as
conventional routers
Notes-4CMPT 471 •2003-36
Proxy ARP

ARP

Address Resolution Protocol

Maps IP addresses to physical addresses

Proxy

an application that closes a straight path between
2 networks and prevents the crackers from
obtaining internal addresses and details of a
private network
A
B
C
R
D
E
Main network
Hidden network
Router running proxy ARP
Notes-4CMPT 471 •2003-37
Proxy ARP (cont.)

How proxy ARP works?

Main network and hidden network: share the
same netid

A router, R, connects these 2 networks

R knows which hosts lie on which physical
networks, and maintains the illusion that only
one network exists

A in the main network sends packets to E in
the hidden network

A broadcasts ARP request for E’s physical addr.

R responses ARP request by sending back its own
physical addr.

A sends the packet destined for E to R

R forwards the packet destined for E over the hidden
network
Notes-4CMPT 471 •2003-38
Proxy ARP (cont.)

Advantages

One netidis shared by multiple networks

Proxy ARP can be added to a single
router without disturbing other hosts or
routers on the network

Disadvantages

The network must use ARP for address
resolution

Cannot be generalized to more complex
network topology

Rely on managers to maintain tables of
machines and addresses manually
3
Notes-4CMPT 471 •2003-39
Subnet Addressing

Subnet addressing, subnetting

A network is divided into several
smaller subnets

Each subnet has its own subnet
address

Subnets appear as a single network
to the rest of the internet

The router attached to the subnets
knows the network is physically
divided into subnets
Notes-4CMPT 471 •2003-310
Subnet Addressing
(cont.)

SubnettingExample
To the rest of the internet
141.14.128.1141.14.128.2
141.14.5.1141.14.5.2141.14.5.3
141.14.64.1141.14.64.5141.14.64.7
R1
R2
netid: 141.14.0.0
Class B
Subnet: 141.14.5.0Subnet: 141.14.64.0
Subnet: 141.14.128.0
Notes-4CMPT 471 •2003-311
Subnet Addressing
(cont.)

Subnettingaddress
Without subnetting
hostidnetid
191•10•141•14
192•192•141•14
hostidsubnetidnetid
With subnetting
ConnectionExchangeArea code
-
Hierarchy in telephone number
8902864)405(
Hierarchy addressing:
netid
subnetid
hostid
Notes-4CMPT 471 •2003-312
Subnet Addressing
(cont.)

Mask

a 32-bit binary number that gives the network
address when bitwise ANDedwith an IP address

e.g.
IP address: 123.24.3.1 (class B)
Mask: 11111111 11111111 00000000 00000000
IP & mask = 123.24.0.0 (network address)
Bitwise
AND
mask
IP
address
Network
address
4
Notes-4CMPT 471 •2003-313
Subnet Addressing
(cont.)

Default masks Masks for class A, B, C addresses

1s: preserve the netid

0s: set the hostidto 0

Number of 1s is predetermined: 8/16/24
255.255.255.011111111 11111111 1111111100000000C
255.255.0.011111111 1111111100000000 00000000B
255.0.0.01111111100000000 00000000 00000000A
Dotted-decimal
mask
Binary maskClass
Notes-4CMPT 471 •2003-314
Subnet Addressing
(cont.)

Subnet mask

Change some of the leftmost 0s in the
default mask to 1s to make a subnet mask

Preserve netidand subnetid, set hostidto 0

Contiguous subnet mask (recommended)11111111 11111111 11000000 00000000

Noncontiguous subnet mask11111111 11111111 00110000 001000000
Subnet mask: 255.255.224.0
00000 00000000
11111111111 11111111
Default mask: 255.255.0.0
00000000 0000000011111111 11111111
Notes-4CMPT 471 •2003-315
Subnet Addressing
(cont.)
Bitwise
AND
Subnet mask
255.255.224.0
141.14.72.24
IP address
141.14.64.0
Network address
72 010 01000
224 111 00000
010 00000 (64)
Notes-4CMPT 471 •2003-316
Subnet Addressing
(cont.)

Subnet design exampleA company is granted network address
200.16.64.0 (class C). It needs 6 subnets.
Design the subnet.

# of 1s in the default mask = 24 (class C)

6 subnets < 23: need 3 more 1s in the subnet
mask

Total # of 1s in the subnet mask: 24 + 3 = 27

Total # of 0s in the subnet mask: 8 –3 = 5
(hostidbits)

Mask is:
11111111 11111111 11111111 11100000
,
or, 255.255.255.224

# of hosts per subnet: 2
5 = 32
5
Notes-4CMPT 471 •2003-317
Subnet Addressing
(cont.)

Subnet address ranges
200.16.64.255~200.16.64.2248th :
200.16.64.223~200.16.64.1927th :
200.16.64.191~200.16.64.1606th :
200.16.64.159~200.16.64.1285th :
200.16.64.127~200.16.64.964th :
200.16.64.95~200.16.64.643rd :
200.16.64.63~200.16.64.322nd:
200.16.64.31~200.16.64.01st:
Notes-4CMPT 471 •2003-318
Subnet Addressing
(cont.)

Fixed-length subnetting

All 1s or all 0s subnet is not recommended

All 1s and all 0s host addresses are reserved

Variable-length subnetting

No single subnetidpartition works for all
organizations

An organization may select subnetidpartition on
a per-network basis; all hosts and routers
attached to the network must follow the partition
254 (28-2)254 (28-2)8
16382 (214
–2) 2 (22-2) 2
65534 (216
–2)1 (20)0
Hosts per subnet (class B)# of subnetsSubnet bits
Notes-4CMPT 471 •2003-319
Subnet Addressing
(cont.)

Variable-length subnettingexample A class C site has 5 subnets with host numbers:
60, 60, 60, 30, 30

2 bits in subnetid? No, only 4 subnets.

3 bits in subnetid? No, at most 32 hosts per
subnets.
62 hosts
62 hosts
30 hosts
62 hosts
30 hosts
First mask (26 1s)
255.255.255.192
Second mask (27 1s)
255.255.255.224
router
Notes-4CMPT 471 •2003-320
Subnet Addressing
(cont.)

Subnet broadcasting

Subnet broadcast address

hostidis all 1s

3-tuple form: {netid, subnetid, -1}, “-1”
means “all 1s”.

{netid, -1, -1}

Means “deliver packet to all hosts with
network address netid, even if they are in
separate physical subnets”

Operationally, such broadcasting make
sense only if routers that interconnect the
subnets agree to propagate the packets
to all subnets
6
Notes-4CMPT 471 •2003-321
Classless Addressing

Also called supernetting

Combine several address blocks to create
a larger address range: supernet

Instead of using a single netidfor
multiple subnets, it allows a network’s
addresses to span multiple netids

E.g., an organization that needs 1000 addresses
can be granted 4 class C blocks instead of 1 class
B block
X.Y.32.0 ~ X.Y.32.255
X.Y.33.0 ~ X.Y.33.255
X.Y.34.0 ~ X.Y.34.255
X.Y.35.0 ~ X.Y.35.255
Notes-4CMPT 471 •2003-322
Classless Addressing (cont.)

Address block assigning

Choose address blocks randomly

The routers outside of the supernettreat
each block separately

Each router has N entries in its routing table,
N = # of blocks; therefore, increase the size
of the routing table tremendously
Notes-4CMPT 471 •2003-323
Classless Addressing (cont.)

Choose address blocks based on a set
of rules so that each router has only
one entry in the routing table: required
by CIDR (Classless Inter-Domain
Routing)

# of blocks is a power of 2 (1, 2, 4, 8 …)

The size of each block is a power of 2

The blocks are contiguous in the address
space (no gaps between the blocks)

The size of supernet= (# of blocks) *
(size of each block): a power of 2

The first address can be evenly divisible
by supernetsize
Notes-4CMPT 471 •2003-324
Classless Addressing (cont.)

Blocks defining in different addressing
schemesBlock: [first address, last address]

Classfuladdress

one block, default mask is always known

the first address only can define the block

Subnetting

the first address in the subblock(subnet) and
subnet mask define the subblock

Supernetting

the first address of the supernetand supernet
mask define the superblock

IP address & supernetmask = first address
(network address)
7
Notes-4CMPT 471 •2003-325
Classless Addressing (cont.)

Supernetmask

The reverse of a subnet mask

Has less
1s than the default mask for this class
00000
11111111 11111111 11111111 111
Subnet mask
Divide 1 network into 8 subnets
000 00000
11111111 11111111 11111111Default mask (class C)
000 000 00000
11111111 11111111 11111
Supernetmask
Combine 8 networks into 1 supernet
Subnetting
Supernetting
Notes-4CMPT 471 •2003-326
Classless Addressing (cont.)

Supernetmask examples

A supernetis made out of 16 class C
blocks, what is its supernetmask?
Block #: 16 = 2
4
Change the last 4 1s in the default mask (class
C) to 0s to get the supernetmask:
11111111 11111111 11111111
00000000
11111111 11111111 11110000
00000000
Notes-4CMPT 471 •2003-327
Classless Addressing (cont.)

A supernetwith mask 255.255.248.0
includes an address 205.16.37.44, what is
the address range?

First address
205.16.37.44 AND 255.255.248.0
= 205.16.32.0
(11001101 00010000 00100000 00000000
)

Mask 11111111 11111111 11111000 00000000
,
1s: 21, 0s: 11

Last address: 205.16.39.255 (11001101
00010000 00100111 11111111
)
Notes-4CMPT 471 •2003-328
Classless Addressing (cont.)

Slash notation (CIDR notation):
A.B.C.D/n

For identifying a CIDR block

A.B.C.D: an IP address

n: # of bits that are shared in every
address in the block, i.e., # of 1s in the
mask

Prefix: common part of the address range
(similar to netid), prefix length = n

Suffix: varying part of the address range
(similar to hostid), suffix length = 32 -n
8
Notes-4CMPT 471 •2003-329
Classless Addressing (cont.)

Relationship between mask and prefix
length
255.255.255.0 (class C)/24224.0.0.0/3
255.255.255.255/32255.0.0.0 (class A)/8
255.255.0.0 (class B)/16192.0.0.0/2
255.240.0.0/12128.0.0.0/1
Mask/nMask/n
Class A: a.b.c.d/8
Class B: a.b.c.d/16
Class C: a.b.c.d/24
Notes-4CMPT 471 •2003-330
Classless Addressing (cont.)

Subnettingwith classless addressing

Increase supernetprefix length (n) to
define the subnet prefix length
Example
: an organization is granted the block
130.34.12.64/26. It needs to have 4 subnets.
What is the subnet address and address range
for each subnet?

Prefix length = 26, suffix length = 6 # of
addresses in the block: 2
6
= 64

4 subnets 16 addresses per subnet

4 subnets subnet prefix /28 (2 more 1s
in the mask)
Notes-4CMPT 471 •2003-331
Classless Addressing (cont.)

Subnet address ranges

1st: 130.34.12.64/28 ~ 130.34.12.79/28

2nd: 130.34.12.80/28 ~ 130.34.12.95/28

3rd: 130.34.12.96/28 ~ 130.34.12.111/28

4th: 130.34.12.112/28 ~ 130.34.12.127/28
130.34.12.64/28
130.34.12.80/28
130.34.12.96/28
130.34.12.112/28
R3
R1
R2
Site: 130.34.12.64/26
Notes-4CMPT 471 •2003-332
Classless Addressing (cont.)

Reserved CIDR blocks

Private addresses, unroutableaddresses

Used with private networks

Never assigned to networks in the global
Internet

Router in the global Internet knows they
are reserved addresses, and can detect it
if a packet destined to the reserved
address accidentally reaches the Internet
9
Notes-4CMPT 471 •2003-333
Classless Addressing (cont.)

Reserved CIDR blocks: list
169.254.255.255169.254.0.0169.254/16
192.168.255.255192.168.0.0192.168/16
172.31.255.255172.16.0.0172.16/12
10.255.255.25510.0.0.010/8
Last addressFirst addressPrefix