B227 Lecture Overheads - Topic 5: The Network Layer in the Internet

[Back to Lecture Notes page]

The Network Layer in the Internet

Sub-topic Outline:

The IP Protocol

IP Addresses
Subnets

Internet Control Protocols

ICMP
ARP
RARP

IPv6

The Internet Protocol (IP)

The Internet is a collection of many different networks

The central protocol which holds all the different networks together is the Internet Protocol (or IP), especially the way addressing format.

The version of IP used for the last 20 years is version 4.

The aim of IP is to transport datagrams from source to destination inside and outside the same network.

The IP Header

Fields in an IPv4 header:

Version: which version of IP?

IHL: This header's length - minimum 160 bits.

Type of Service: What type of service this packet requires (eg. reliability, speed, etc). Each router can try and satisfy this.

Total length: This header's length - maximum 64Kb.

Identification, DF, MF, Fragment Offset: for fragmenting packets which are too large.

Time to live: value to limit packet lifetime, so it doesn't float around forever - set by host and decreased by each router.

Protocol: the protocol which pass IP the data (eg. TCP).

Checksum: checksum for the header, to ensure there are no errors.

Source and Destination Addresses: Sender and receiver of the data.

Options: Other useful header fields not included above.

IP Addresses

Every host on the Internet has a unique IP address.

All addresses are in 32-bits - used in the source and destination fields of the IP headers.

Usually written in dotted decimal notation:

Every 8-bits (1 byte) written as a decimal number between 0 to 255.

We end up with 4 decimal numbers separated by dots. Eg. 134.115.157.141

Four classes of networks (A, B C, D) with their own addresses: Figure 5-47 p416

Some addresses are also special addresses: Figure 5-48 p417

IP Subnets

IP allows subnets to be created to:

Allow individual sites to create a hierarchy in their local networks

Alleviate individual routers on a site from having to know about all host on the network.

How IP subnet numbers work:

Internet Control Protocols

Beside IP, which is used to transfer data, there are also other protocols used on the Internet for control.

The Internet Control Message Protocol (ICMP)

Address Resolution Protocol (ARP)

Reverse Address Resolution Protocol (RARP)

Like IP, each of these protocols also specifies a format for packets used.

Internet Control Message Protocols

Used to report anything unexpected detected by the routers.

Different types of ICMP packets: Figure 5-50 p420

Address Resolution Protocol (ARP)

Used to convert from an IP address to a LAN address (also called physical address, hardware address, or MAC address).

Necessary because most LAN interface cards only understands its LAN address format, but not IP addresses.

How it works:

A machine wanting to send data to another machine with a certain IP address (eg. 134.115.157.141) will broadcast an ARP packet to the whole LAN saying: "Who owns the IP address 134.115.157.141?"

The machine with IP address 134.115.157.141 will respond with a reply ARP packet with its own LAN address on it.

Reverse Address Resolution Protocol (RARP)

To convert from LAN addresses to IP addresses

Used in diskless workstations:

When the workstations boots-up, it loads an image of the OS from a file server.

It then broadcast an RARP packet containing its own LAN addresses.

An RARP server will respond with the machines IP address.

This way, all workstation can load up the same OS image from the file server.

IPv6

Deployed the middle of 1999.

Main purpose for change from IPv4:

Extend the IP address space

More efficiency in its operations compared to IPv4.

To extend IP so that it supports more multimedia protocols currently being developed

The IPv6 Address

Consists 128 bits (compared to32 bits for IPv4)

Using hexadecimal (based 16, 4 bits) digits (0, 1, …, 9, A, …, F) for representation

Digits grouped in four to represent a 16-bit (4 x 4 bits) numbers.

The address consist of eight such 16-bit numbers, separated by colons. Eg:

68DA:8909:3A22:FA64:68DA:8909:3A33:AB45

Some parts may not be needed, and will be set to 0. Eg:

68DA:0000:0000:0000:68DA:8909:3A33:AB45

68DA:0:0:0:68DA:8909:3A33:AB45

Consecutive 0’s can be omitted and replaced by a double colon. Eg from above:

68DA::68DA:8909:3A33:AB45

If there are two sections in the address with consecutive 0’s, only one can be replaced by a double colon.

IPv4 to IPv6 Address Conversion

IPv6 will represent IPv4 addresses by keeping the decimal dot format, and tacking 0’s in the beginning. Eg.

164.115.64.2

becomes

::164.115.64.2

IPv6 Address Hierarchy

The right-most four fields (64 bits) will represent the host portion of the address, and the left-most four fields will represent the network portion.

The IPv6 Datagram

Figure 5-56 p439

Version field (4 bits) – version of the protocol, ie. 6 (or 0110 in binary)

Priority field (4 bits) – to indicate the priority of the packet.

Packets with priority 0-7 indicates source provides congestion control.

The higher the priority, the important important it is for the packet to reach it’s destination.

Flow label field (24 bits) – used to indicate multiple packets "flowing" from one source to a destination.

Can be used in place of destination field.

Its use is implementation specific.

Payload Length field (16 bits) – The size of the data plus header. Limited to 64kb.

Next Header field (8 bits) – specified what the next header of the packet is.

Usually indicates TCP or UDP.

But can also be an IP header extension.

Hop Limit field (8 bits) – How many hops the packet can traverse before being discarded.

Controversies with IPv6

Fixed size address length too large

Hop limit maximum of 255 (for 8-bits) to small

Fixed packet sizes of 64Kb too small for supercomputers

No checksum

No support for mobile host

What security level to offer?

[Back to Lecture Notes page]