Internet architecture, addressing, encapsulation, reliable transport and the TCP/IP protocol suite

To introduce the concept of an address space in general and the address space of IPv4 in particular.

To discuss the classful architecture, classes in this model, and the blocks of addresses available in each class.

To discuss the idea of hierarchical addressing and how it has been implemented in classful addressing.

To explain subletting and super netting for classful architecture and show how they were used to overcome the deficiency of classful addressing.

To discuss the new architecture, classless addressing, that has been devised to solve the problems in classful addressing such as address depletion.

Interconnect heterogeneous networks and provide universal service

Hardware: routers connect different networks

Internet protocols: provide universal service by creating a single virtual network

• Although a single router can connect many networks, most organisations use multiple routers

The illusion that there is a single universal network

The TCP/IP Internet Protocols

begun in the 1970s

The Internet has emerged into the public domain in the 1990s

Application Layer The application layer of the TCP/IP model corresponds to the application layer of the OSI reference model. Some well known examples of application level entities within the TCP/IP domain are:

• FTP/Telnet • HTTP/Secure HTTP (SHTTP) • POP3/SMTP • SNMP

Transport Layer

The transport layer of the TCP/IP model maps fairly closely to the transport layer of the OSI model. Two commonly used transport layer entities are TCP and User Datagram Protocol(UDP)

Internet Layer

The Internet layer of the TCP/IP model maps to the network layer of the OSI model. Consequently, the Internet layer is sometimes referred to as the network layer. The primary component of the Internet layer is the Internet Protocol (IP).

Network Access Layer The lowest layer of the TCP/IP protocol stack is the network access layer. The network access layer contains two sublayers, the media access control (MAC) sublayer and the physical sublayer. The MAC sublayer aligns closely with the data link layer of the OSI model, and is sometimes referred to by that name. The physical sub layer aligns with the physical layer of the OSI model.

Note: Some references divide the TCP/IP model into 5 layers, with the MAC and physical layers occupying the lowest two layers.

Examples of the network access layer : • Ethernet

• Wireless Fidelity (Wi-FI)/WiMAX

• ATM/Frame Relay

Uniform addressing, the IP address hierarchy, address classes, dotted decimal notation, special addresses, routers and addresses, address resolution

Internet protocols deal in packets and provide uniform addressing

Internet addressing is specified in the IP protocol

Each host is assigned a unique 32 bit address

Each 32 bit address is divided into two parts

prefix: physical network to which the host is attached - the network number

suffix: a host attached to a given physical network

Prefixes are coordinated globally and suffixes locally

The address space of IPv4 is 4,294,967,296. is it enough?

IPv4 addresses are unique. They are unique in the sense that each address define one, and only one, connection to the Internet. Two devices on the Internet can never have the same address at the same time. However, if a device has two connections to the Internet, via two networks, it has two IPv4 addresses. The IPv4 addresses are universal in the sense that the addressing system must be accepted by any host that wants to be connected to the Internet.

Size of prefix and suffix determines maximum number of networks and maximum number of hosts per network

IP defines different classes of address with different sized prefixes and suffixes

The first four bits of the address specify its class

Makes it easier to for humans to use addresses (names are also possible)

Public Internet network numbers are assigned by Internet Service providers (ISPs)

The idea of network mask in classless addressing is the same as the one in classful addressing. A network mask is a 32-bit number with the n leftmost bits all set to 0s and the rest of the bits all set to 1s. Example The following addresses are defined using slash notations: a. In the address 12.23.24.78/8, the network mask is

255.0.0.0. The mask has eight 1s and twenty-four 0s. The prefix length is 8; the suffix length is 24.

b. In the address 130.11.232.156/16, the network mask is 255.255.0.0. The mask has sixteen 1s and sixteen 0s.The prefix length is 16; the suffix length is 16.

The number of addresses in the block can be found as: Note: in which n is the prefix length and N is the number of addresses

in the block.

The first address (network address) in the block can be found

by ANDing the address with the network mask:

The last address in the block can be found by either adding the first address with the number of addresses or, directly, by ORing the address with the complement (NOTing) of the network mask:

Example One of the addresses in a block is 167.199.170.82/27. Find the number of addresses in the network, the first address, and the last address. Solution The value of n is 27. The network mask has twenty-seven 1s and five 0s. It is 255.255.255.240.

a. The number of addresses in the network is 232 − n = 232 − n = 25 = 32. b. We use the AND operation to find the first address (network address).The first address is 167.199.170.64/27.

c. To find the last address, we first find the complement of the network mask and then OR it with the given address: The last address is 167.199.170.95/27.

One of the addresses in a block is 7.63.110.114/24 Find the number of addresses, the first address, and the last address in the block.

The network mask is 255.255.255.0. a. The number of addresses in the network is

232 − 24 = 256. b. The first address is 17.63.110.0/24. c. The last address is 17.63.110.255/24.

The Internet is running out of addresses Allow division between prefix and suffix to appear at

an arbitrary boundary Consider network with only 9 hosts

Only need 4 bits for host suffix

Class C (smallest) address uses 8 bits for host suffix

Sol: Can subdivide a class C address into 16 addresses with a 28 bit prefix and 4 bit suffix

Extend dotted decimal notation 193.68.138.0/28, 193.68.138.16/28, …,193.68.138.240/28

Routers are assigned two or more IP addresses

So are multi-homed computers

An Internet packet passes through a series of routers each hop takes it over a particular network,

either to a specific computer on that network or to the next router

in either case, the sending router has to map between the protocol (IP) address and a hardware address

this is called Address Resolution

Table lookup Closed-form computation Message exchange

send message to specific server computers

broadcast message, only the required computer responds

TCP/IP defines the Address Resolution Protocol (ARP) which defines the format of resolution requests and responses

Special DHCP server that assigns IP addresses to hosts

Newly booted machine broadcasts a DHCP discover packet

DHCP server sends back an IP address Permanent IP addresses

▪ Manually assigned by administrator

Automatic IP address from a pool of addresses to be allocated on demand ▪ Leased for a finite period of time

DHCP server does not need to be on the same

network as the host

Uniform addressing Address classes Dotted decimal notation Classless addressing Special IP addresses

Service paradigm, IP datagrams, routing, encapsulation, fragmentation and reassembly

TCP/IP supports both connectionless and connection-oriented services

fundamental delivery service is connectionless at the Internet layer

optional reliable connection-oriented service is layered on top of this at the transport layer

Packets of data are sent across multiple physical networks via routers

Internet protocols define a universal virtual packet - the IP datagram

The amount of data carried in a datagram is not fixed and is determined by an application

• Each router forwards a virtual packet by using a local routing table

• Each entry is: – destination address

– mask

– next hop • IP address of a router or

• Deliver direct

• Then does address resolution

IP attempts best effort delivery and does not guarantee to deal with:

datagram duplication

delayed or out of order delivery

corruption of data

datagram loss

These issues are dealt with other protocol layers

When an IP datagram is sent across a physical network it is placed in the data area of a frame and the frame type is set to IP

Maximum transmission unit - max of data that a frame can carry on a given network

A packet may have to cope with different MTU sizes as is passes over an internet

A datagram that is larger than MTU is fragmented into smaller datagrams

Is done at the final host

routers require less state information

fragments can take different routes

Header fields indicate when the data is a fragment and also where it belongs

Whole datagram is lost if any fragment is lost