Subnetting

1
Subnetting

• Dividing a network into multiple networks

2
Limitations of this approach

• Subnets are created in lumpy increments
• A class A network can have 1, 254, or 65,534
  subnets with 16,777,214, 65,534, or 254 hosts /
  subnet respectively
• A class B network can have 1 or 254 subnets with
  65,534, or 254 hosts / subnet respectively
• A class C network can only have 1 subnet (no
  subnetting)
• One result is wasted IP addresses due to use of
  lumpy increments
• What if?
• Your physical network supported 1000, 2000
  hosts/segment?
• If you had a class B network, youd be forced to
  create artificially small network segments.
• What if you had a class B (or class C) network
  and needed more than 254 subnets (or more than 1
  subnet)?

3
Use of partial octets for mask

• You don't need to allocate the whole octet for
  just one use.
• So far, octet has been all subnet ID or all host
  ID
• One octet in the mask can be defined as being
  used in part for subnet ID, and partly for host
  ID
• With a few restrictions, you can allocate any
  number of bits in this octets for subnet number
  and host ID
• Restriction
• At least two bits are needed to represent subnet
  ID
• At least two bits are used to represent host ID.

4
Restrictions on masks
class A
Dividing line between 1s and zeros can be
placed anywhere between these limits
Default mask
2 bit minimum for host ID (all 0s)
2 bit minimum for subnet ID (all 1s)
Class B
2 bit minimum for subnet ID (all 1s)
2 bit minimum for host ID (all 0s)
Default mask
Class C
Everything to the left of the dividing line will
be all 1s, all 0s are to the right
Default mask
5
Class B subnet example

• The first two octets represent the IP Network
  number.
• You can use between 2 and 14 bits of the host ID
  to represent subnet ID
• You can use between 2 and 14 bits to represent
  host ID
• Let's say you used 6 bits of the IP host ID area
  to represent subnet ID
• The resulting subnet mask would be 255.255.252.0

bit position within the octet
Default mask
Mask for subnet
Host ID
6
Determining number of subnets hosts/subnet
possible for a subnet mask

• To determine possible number of subnets and
  possible number of hosts per subnet, use the
  following
• subnets 2( number of bits used to store the
  subnet ID) - 2
• hosts / subnet 2( number of bits used to store
  the host ID) 2
• Example Class B mask 255.255.252.0 uses 6 bits
  to represent subnet ID, and 10 bits to represent
  Host ID within the subnet
• subnets 26 2 64 2 or 62 subnets
• hosts / subnet 210 2 1024 2 or 1022
  hosts
• Why minus 2?
• Subnet ID shouldnt be all 1s or all 0s
• Host ID cant be all 1s or all 0s

7
Possible octet mask values

• There are 9 possible values any given octet will
  be in a subnet mask
• 255 - all 8 bits are part of the subnet mask -
  11111111
• 254 - 1st 7 bits are part of the subnet mask -
  11111110
• 252 - 1st 6 bits are part of the subnet mask -
  11111100
• 248 - 1st 5 bits are part of the subnet mask -
  11111000
• 240 - 1st 4 bits are part of the subnet mask -
  11110000
• 224 - 1st 3 bits are part of the subnet mask -
  11100000
• 192 - 1st 2 bits are part of the subnet mask -
  11000000
• 128 - 1st bit is part of the subnet mask -
  10000000
• 0 - all of the octet represents host ID
  - 00000000

8
More examples

• Class A mask 255.255.192.0
• 10 bits used for subnet mask 1022 subnets
  possible
• 14 bits used for host ID (within subnet) 16,382
  maximum / subnet
• Class B mask 255.255.248.0
• 5 bits used for subnet mask 30 subnets possible
• 11 bits used for host ID 2046 hosts / subnet
• Class C mask 255.255.255.224
• 3 bits used for subnet mask 6 subnets possible
• 5 bits used for host ID 30 hosts / subnet

9
Defining a subnet mask

• Feasible masks are limited by
• How many network segments you need
• indicates minimum number of bits needed in subnet
  mask for subnet ID
• Expected maximum number of hosts / segment
• indicates minimum number of bits needed in subnet
  mask to represent host ID
• This should give you one or more options of masks
  that would provide the needed number of subnets
  and hosts/ subnet
• If not, youve specified an impossible solutions,
  e.g. 50 subnets with 200 hosts/subnet on a class
  C network.

10
Defining a subnet mask

• Determine the number of bits for subnet ID
• Add 2 to the number of subnets needed
• Convert this number into binary and determine how
  many binary digits are needed
• 4 subnets 4 2 (6) is 110 (requires 3 bits)
• 100 subnets 100 2 (102) is 1100120 (requires
  7 bits)
• 500 subnets 500 2 (502) is 111110100 (requires
  9 bits)
• After the default mask for the network ID,
  allocate that number of bits as 1s in the mask
• Why add 2?
• subnet ID shouldnt be all 0s or all 1s (see
  next slide)

11
Defining a subnet mask

• With a 3 bit subnet mask, there are 8 numbers
  that can be represented
• As shown below, zero and 224 (all 0s and all
  1s) would normally not be used for subnet IDs
• This results in only 6 usable subnet IDs being
  possible given 3 bits

12
Defining a subnet mask

• Assume a class A network
• 4 subnets
• 11111111.1110000.00000000.00000000 or
  255.224.0.0
• 100 subnets
• 11111111.11111110.00000000.0000000 or
  255.254.0.0
• 500 subnets
• 11111111.11111111.10000000.00000000 or
  255.255.128.0
• Assume a class B network
• 4 subnets 255.255.224.0
• 100 subnets 255.255.254.0
• 500 subnets 255.255.255.128

13
Defining a subnet mask

• Determine number of hosts needed per network
  segment
• Add 2 to this number
• Convert this number into binary and determine how
  many binary digits are needed
• 4 hosts is 110 (requires 3 bits)
• 100 hosts is 1100110 (requires 7 bits)
• 500 hosts is 111110110 (requires 9 bits)
• Your mask must have at least the required number
  of bits set to zero in the low-order (right-hand)
  positions of the mask
• Plus 2? host ID cannot be all 1s or all 0s

14
Defining subnet IDs and Host IDs

• The subnet mask is used to determine possible
  subnet IDs for the network
• These are assigned them to the networks
  subnetworks
• Given a subnet ID, that defines the range of
  possible host IDs
• Determining the network ID host ID from an IP
  address with such a partial-octet mask is NOT
  user-friendly

15
Defining subnet IDs

• Assume an octet mask of 224 11100000
• Start numbering IDs at 00100000
• End number of IDs at 11000000
• Convert each 8-bit number to decimal for
  interpretation

00000000 0 00100000 32 01000000 64 01100000
96 10000000 128 10100000 160 11000000
192 11100000 224
Not used
16
A shortcut for enumerating subnet IDs

• For a given mask such as 224
• Determine the decimal value of the 1st (lowest)
  possible subnet ID value in the octet.
• For example, with 224, this number was 32
• Add 32 to the current subnet ID to determine the
  next subnet ID 64 32 32
• Repeat the above until you enumerate all possible
  IDs
• 96 64 32
• 128 96 32 .. And so on

17
Defining the range of host IDs for each subnet

• The starting point for the range of host IDs is
  to set the right-most host ID bit to be one, and
  all other host ID bits to be zero
• The end-point would be the reverse the
  right-most host ID bit is set to zero, and all
  other host ID bits are ones.
• For example, given a 255.255.224.0 subnet mask

Invalid x.y.32.1 x.y.63.254 x.y.64.1
x.y.95.254 x.y.96.1 x.y.127.254 x.y.128.1
x.y.159.254 x.y.160.1 x.y.191.254 x.y.192.1
x.y.223.254 Invalid
Host ID Ranges
Subnet IDs
18
Interpreting the IP address

• Following from the last slide
• x.y.32.1 is x.y.001 00000.00000001
• x.y.63.254 is x.y.001 11111.11111110
• 00100000 (32) 11111 (31) equals 63
• x.y.64.1 is x.y.010 00000.00000001
• x.y.95.254 is x.y.010 11111.11111110
• 01000000 (64) 11111 (31) equals 95

19
Bitwise anding

• A prior explanation for mask use said that for a
  given bit
• "1" over a number says "Look at the number
  underneath" a "0" says "Don't look.
• What hosts really do with a mask is to perform
  bitwise anding of mask bits with IP address bits
• Logical anding
• 1 and 1 yields 1
• 1 and 0 yields 0
• 0 and 1 yields 0
• 0 and 0 yields 0

20
Bitwise anding examples
Remaining Values
32-32 0
63-32 31
45-32 13
73-64 9
21
Example of two class B subnetworks with a
255.255.224.0 mask
Network 131.210.64.0 Mask 255.255.224.0
Network 131.210.32.0 Mask 255.255.224.0
131.210.32.100
131.210.64.56
131.210.73.123
131.210.45.150
To the Internet
22
Example 1 use of the subnet mask
Source host

• Suppose host 131.210.32.100 wants to send a
  packet to 131.210.45.150
• 1st, the host uses its subnet mask to determine
  its own subnet ID
• Then the host applies its mask to the destination
  address
• If these two numbers are the same, it indicates
  that the destination is on the same subnet as the
  source it doesnt need to be routed.

Destination host
23
Example 2 use of the subnet mask
Source host

• Suppose host 131.210.32.100 wants to send a
  packet to 131.210.73.123
• The host uses its subnet mask is used to
  determine its subnet ID
• Then the host applies its mask to the destination
  address
• If these two numbers are not the same, it
  indicates that the destination is on a different
  subnet than the source it needs to be routed.

Destination host
24
Example 3 use of the subnet mask
Source host

• Suppose host 131.210.32.100 wants to send a
  packet to 60.5.10.100 (somewhere on the Internet)
• The host uses its subnet mask is used to
  determine its subnet ID
• Then the host applies its mask to the destination
  address
• If these two numbers are the same, it indicates
  that the destination is on a different subnet
  than the source it needs to be routed.
• The router sends it onto the Internet since its
  not a local subnet, its on a network of a
  different organization

Destination host
25
Alternate way of expressing mask

• 131.210.32.100/19
• This says that the mask is 255.255.224.0 (19
  bits)
• Subnet 131.210.32.0, host 0.0.0.100
• 131.210.73.123/19
• This says that the mask is 255.255.224.0 (19
  bits)
• Subnet 131.210.64.0, host 0.0.9.123
• 10.1.53.100/20
• The mask is 255.255.240.0 (20 bits)
• Subnet 10.1.48.0, host 0.0.5.100

26
Designing a Variable-Length Subnet Mask

• Analyze requirements for individual subnets
• Aggregate requirements by their relationships to
  the nearest power of two
• Use subnets that require largest number of
  devices
• To decide the minimum size of the subnet mask
• Aggregate subnets that require fewer of hosts
• Define VLSM scheme that
• Provides the necessary number of subnets of each
  size to fit its intended use best

27
Calculating Supernets

• Supernets
• Steal bits from network portion of an IP
  address to lend those bits to the host
• Permit multiple IP network addresses to be
  combined
• Allow an entire group of hosts to be reached
  through a single router address

28
Classless Inter-Domain Routing

• Limitations
• Network addresses must be contiguous
• When address aggregation occurs
• CIDR address blocks work best when they come in
  sets that are greater than 1 and equal to some
  lower-order bit pattern that corresponds to all
  1s
• Addresses commonly applied to Class C addresses
• To use a CIDR address on any network
• Routers in routing domain must understand CIDR
  notation

29
Public Versus Private IP Addresses

• Private IP addresses ranges
• May be in the form of IP network addresses
• Address masquerading
• May be performed by boundary devices that include
  proxy server capabilities
• Private IP address limitation
• Some IP services require a secure end-to-end
  connection

30
(No Transcript)