1. Classless and Subnet Address Extensions (CIDR)
Topics:
There are problems with the IP addressing scheme we’ve studied
We’ll study some ways to get around these problems

2. Review: IP Addresses

3. Problems with IP Addresses
The designers of IP addresses did not foresee the Internet’s tremendous growth
Higher overhead to manage network addresses
Larger routing tables
IP addresses might one day be exhausted

4. Solution to IP Addresses Problems
The same IP network prefix can be shared by multiple physical networks
A site can choose to assign and use IP addresses in unusual ways internally as long as:
All hosts and routers at the site honor the site’s addressing scheme
The site’s addressing scheme is transparent to other sites on the internet

5. Strategy 1: Transparent Routers
A network with a class A IP address can be extended: H1 H2 T H3 H4

6. Transparent Routers (cont)
Hosts on LAN are assigned IP addresses as if they were on WAN
LAN does not need its own network prefix
Traffic for hosts on LAN is multiplexed through T
Other hosts and routers on the WAN do not know T exists

7. Transparent Routers Advantages Disadvantages
Require fewer network addresses (LAN doesn’t need a separate network prefix)
Load balancing
Disadvantages
Require a large address space
Do not provide all the services of standard routers

8. Router running proxy ARP
Strategy 2: Proxy ARP
Using ARP, map a single network prefix into two physical addresses
Main network H1 H2 H3
Router running proxy ARP R H4 H5 H6
Hidden network

9. Proxy ARP (cont)
Gives the illusion that all hosts are on the same physical network
Router R answers ARP requests on each network for hosts on the other
R answers ARPs with its own hardware address (it lies)
When R receives a datagram it forwards it to the correct physical address

10. Proxy ARP Advantages Disadvantages Require fewer network addresses
Only the router running proxy ARP needs to know what’s going on
Disadvantages
Can only be used if the network uses ARP for address resolution
Allows spoofing

11. Strategy 3: Subnet Addressing
Hierarchical addressing
Network H1 H2
Rest of the internet R
Network H3 H4
All traffic to

12. Subnet Addressing (cont)
R receives all traffic for network
R routes the datagram to a physical network based on bits in the hostid field of the IP address
Another level has been added to the addressing hierarchy

13. Subnet Addressing (cont)
Regular (Class B) IP address:
New interpretation (locally only):
netid hostid
netid subnet hostid

14. Subnet Addressing (cont)
Advantages
Minimizes network address usage
Accommodates growth
Disadvantages
Added layer of complexity
Difficult to change once hierarchy is established

15. Subnet Addressing (cont)
Flexible
Allows 256 physical networks with 256 hosts each
Allows 8 physical networks with 8192 hosts each
netid subnet hostid
netid sub hostid

16. Subnet Masks 32 bits Example - a class B network: Subnet mask:
1 if the bit is part of the network address
0 if the bit is part of the host address
Example - a class B network:
Subnet mask:
netid subnet hostid

17. Subnet Masks Subnet bits do not have to be contiguous: = subnet id
= host id
netid

18. Representing Subnet Masks in Dotted Decimal Notation
Example - a class B network:
Subnet mask:
Dotted Decimal:
netid subnet hostid

19. Representing Subnet Masks in 3-tuple Notation
3-tuple notation
{<netid>,<subnet id>,<hostid>}
-1 means “all ones”
{-1,-1,0}

20. Routing in the Presence of Subnets
All hosts and routers must use a subnet routing algorithm
Net 1 (not a subnet address) R1 H R2
Net 2 (subnet of address N)
Net 3 (subnet of address N)

21. The Subnet Routing Algorithm
Recall the standard routing table:
(netid, next hop)
N = netid portion of IP address
Compare N with netid
Match = send datagram to next hop
Routing when subnets are in use:
(subnet mask, netid, next hop)
N = IP address & subnet mask

22. Using Subnet Masks for Routing
Host-specific routes
( , )
( , , )
Default routes
(default, )
( , , )
Standard, non-subnet class B network
( , )
( , , )

23. A Unified Routing Algorithm
Extract the destination IP address, D, from the datagram and compute the netid, N
If N matches any directly connected network address deliver the datagram directly over that network
else
for each entry (M,N,NH) in the routing table {
I = M&D
if (I == N) then send datagram to NH }
if no matches were found declare a routing error

24. Broadcasting to Subnets
IP address =
Broadcast to all hosts on network 128
What if network 128 has subnets?
Routers that interconnect the subnets must propagate the datagram to all physical networks
But the routers must take care not to route the datagrams in loops (reverse path forwarding)
Can you broadcast to just one subnet?
Yes: {network, subnet, -1}

25. Summary
Problem: IP v4 addresses (especially class B) would be exhausted
Solutions:
Subnet addressing - conserve network addresses by using the same network address for multiple physical networks
New version of IP (v6) with larger addresses
Supernet addressing - conserve class B network addresses by allowing a single organization to use multiple class C network addresses