1. IPv4 Addresses

2. Internet Protocol: Which version?
There are currently two versions of the Internet Protocol in use for the Internet
IPv4 (IP Version 4)
Specified in 1980/81 (RFC 760, 791)
Four byte addresses
Universally deployed
Problem: Address space almost exhausted
IPv6 (IP Version 6)
Specification from 1998 (RFC 2460)
Significant differences to IPv4, but not fundamental changes
16-byte addresses
Problem: Not widely used (yet?)

3. Slow adoption of IPv6
IPv6 is available since 15 years, and almost all operating systems and routers support it
But IPv6 is still a small fraction of the total Internet traffic
Measurements at Internet Exchange Point in Amsterdam:
IPv6 traffic as a percentage
IPv6 traffic (total volume)

4. Slow adoption of IPv6
IPv6 is available since 15 years, and almost all operating systems and routers support it
But IPv6 is still a small fraction of the total Internet traffic
Measurements at Internet Exchange Point in Amsterdam:
IPv6 traffic as a percentage
IPv6 traffic (total volume)

5. Slow adoption of IPv6
IPv6 is available since 15 years, and almost all operating systems and routers support it
But IPv6 is still a small fraction of the total Internet traffic
Measurements at Internet Exchange Point in Amsterdam:
IPv6 traffic as a percentage
IPv6 traffic (total volume)

6. ECE 461: IPv4 based
We use IPv4 throughout in lecture, labs, and assignment
IPv6 concepts are discussed in the lecture

7. IPv4 Addresses

8. IPv4 Addresses

9. IP Addresses
An IPv4 address is an address of the Internet Protocol Version 4. When the version is understood from the context we only say “IP address”
On the public Internet, an IP address is unique global address of a network interface
The IP address is used by hosts and routers for delivery of IP datagrams

10. Dotted Decimal Notation
IPv4 addresses are written using dotted decimal notation:
Each byte is identified by a decimal number in the range [0..255], separated by periods (`dots’).
Example:
1st Byte
= 128
2nd Byte
= 100
3rd Byte
= 11
4th Byte
= 60

11. Structure of an IP address
IP address consists of a network prefix and a host number
Network prefix identifies a particular network
Host number identifies an interface on the network
The delivery of a packet from a sender to the destination proceeds in 2 steps:
Use network prefix to deliver packet to the right network
Once the network is found, use the host number to deliver packet to the right interface
How can you determine the length of the network prefix?
network prefix
host number

12. Length of network prefix
The length of the network prefix must be provided in addition to the numerical value of the address
There are two conventions for indicating the length:
Add length of the network prefix to address (“prefix notation”, “slash notation”, “CIDR notation”)
Add bitmask of the network prefix in dotted decimal notation (“netmask”)
32 bits
network prefix
host number
Length of network prefix

13. Notation of IP address
Example: Suppose the network prefix is 16 bits long
Prefix notation: /16
Netmask notation:

14. Special IP Addresses IP address of a network Broadcast address
Host number is set to all zeros, e.g.,
Broadcast address
Host number is set to all ones, e.g.,
Broadcast goes to all hosts on the network
The address is a broadcast on the local network
Convention for default gateway (but not a reserved address):
Default gateway has host number set to ‘1’, e.g.,

15. Reserved Network Prefixes
Loopback interfaces: /8
Most systems use as loopback address
A host can send a packet to itself by sending it to address
Link Local Addresses: /16
Also referred to as auto-configuration
When a system cannot acquire an IP address, it selects a random address from the range (it tests whether the address is in use)
Packets from this address are not forwarded to other networks
Private networks: /8, /12, /16, /10
Private networks are IP networks that are not part of the public Internet
IP addresses need to be unique only within the same private network
Systems on different private networks can re-use addresses
When packets cross between private network and public Internet, IP addresses in header must be modified
Complete list is at:

16. Network Prefix and Address Block
The IP address of a network has the host number set to 0:
Network address with prefix length is called network prefix:
/ or /16
A network prefix is interpreted as a range of IP addresses:
/16  –
A network prefix is also called address block
It provides the range of addresses allowed on the network
The shorter the network prefix, the larger the address block

17. Network prefix length and address block size
To set up a network of a given size, one needs to acquire a address block with a sufficient number of addresses
Network Prefix # of IP Addresses
/32 1
/31 2
/30 4
… …
/24 256
/23 512
/22 1,024
/21 2,048
/20 4,096
/16 65,536
/8 16,777,216
/2 1,073,741,824
/1 2,147,483,648

18. How does one get an IP network address?
IANA
RIR
LIR (ISP)
allocates
allocates
assigns
assigns
End user
End user
To get the networks of UofT at ARIN:
Enter “University of Toronto”.
Click on UNIVER-36
Click on “Related Networks”
Internet Assigned Number Authorities (IANA) manages the global IP address space.
IANA allocates addresses to Regional Internet Registries (RIR)
IANA allocates addresses in /8 address blocks
All available blocks have been allocated. The last available address block was assigned in 2011

19. How does one get an IP network address?
IANA
RIR
LIR (ISP)
allocates
allocates
assigns
assigns
End user
End user
RIR allocates addresses to Local Internet Registries (LIR)
LIRs are generally Internet Service Providers
Block sizes are /20 or larger
End users obtain addresses from an RIR or LIR
Some countries have National Internet Registries (NIR), which are between the RIR and LIR levels

20. Regional Internet Registries
There are five Regional Internet Registries, each responsible for managing the address space of a large geographic area
Address blocks managed by the RIRs are listed at:

21. Available IPv4 addresses (Sep 2018)
RIRs are running low on IP addresses:
(one /8 address block ≈16.8 million IP addresses)
Current status of remaining IP addresses:
(2012)
Addresses in RIR Pool (/8s)
Approximate number of Addresses
APNIC
0.9138
≈ 15 million
RIPENCC
1.1984
≈ 20 million
ARIN
3.3656
≈ 56 million
LACNIC
3.2246
≈ 54 million
AFRINIC
4.1441
≈ 70 million
(Sep 2016)
Addresses in RIR Pool (/8s)
Approximate number of Addresses
APNIC
0.2681
≈ 4.5 million
RIPENCC
0.4544
≈ 7.6 million
ARIN
0.0004
≈ 6,700
LACNIC
0.1148
≈ 1.9 million
AFRINIC
0.4513
≈ 7.5 million
Source:

22. IPv4 address exhaustion
The vast majority of IPv4 addresses are used up
 IPv4 address exhaustion
Adoption of IPv6 is driven by IPv4 exhaustion:
Effectively, IPv4 exhaustion has occurred
In the past, IPv4 exhaustion has come with dire warnings
People have been creative with slowing down IPv4 exhaustion through policies and protocols
IPv4 exhaustion will occur gradually, but is increasingly urgent
As IPv4 addresses become less available, the use of IPv6 will increase
In the past, there was expectation that IPv6 will replace IPv4
What has happened is that IPv4 and IPv6 coexist

23. Classful IP Addresses
When Internet addresses were standardized (early 1980s), the Internet address space was divided up into classes
Three classes available for interfaces:
Class A: Network prefix is 8 bits long
Class B: Network prefix is 16 bits long
Class C: Network prefix is 24 bits long
Two additional classes:
Class D: Used (multicast) group addresses
Class E: Reserved for future use

24. Classful IP Addresses
The classes were identified by the bit values of the IP address
Class A: IP address starts with “0”
Class B: IP address starts with “10”
Class C: IP address starts with “110”
Class D: IP address starts with “1110”
Class E: IP address starts with “11110”
The length of the network prefix is implied by the address
There is no need to provide the length of the network prefix

25. Classful IP Addresses

26. Classful IP Addresses
We will learn about multicast addresses later in this course
Class E addresses have never been released for use

27. IP Allocation with Classful IP Addresses
Limited flexibility for obtaining a network address:
Class C: ≤ = 254 IP addresses:
Class B: ≤ = IP addresses
Class A: ≤ = 16,772,214 IP addresses
Too few network addresses for large networks
27 = Class A networks
214 = 16,384 Class B networks
Flat address space. Routing tables in the backbone Internet need one entry for each network address.
Class C networks: up to 221 = 2,097,152 entries
By 1993, the lookup time of routing tables had become a bottleneck for Internet performance

28. Allocation of Classful Addresses

29. From Classful to Classless
In 1993, there was a major shift for interpreting and allocating IP addresses:
until 1993:
Classful Addresses
after 1993:
Classless Interdomain Routing (CIDR)

30. CIDR - Classless Interdomain Routing
CIDR abandons the notion of classes
Key Concept: The length of the network prefix of IP addresses can be any value
Consequences:
Since the length of the network prefix is no longer implied by the address, the size of the network prefix must be provided with an IP address ( CIDR notation)
Hierarchical routing aggregation introduces dependency of IP addresses to service provider
Need for new lookup algorithms for routing tables

31. Division of IPv4 address space
00000…00
Class A
(231 addresses)
10000…00
Class B
(230 addresses)
11000…00
Class C (229 addresses)
11100…00
11110…00
Class D (228 addresses)
unused addresses
Class E (228 addresses)