1

The TCP/IP Protocol

2

Introduction To TCP/IP

• Transmission Control Protocol/Internet Protocol (TCP/IP)– Most commonly used network protocol suite today– Wide vendor support– Open protocol– Provides access to Internet services

• Windows Server 2003– Can use several protocols – Many of its main features require the use of TCP/IP

3

Internet History

•1961: Kleinrock - queueing theory shows ` effectiveness of packet-switching

•1964: Baran - packet-switching in military nets•1967: ARPAnet conceived by Advanced Research

Projects Agency•1969: first ARPAnet node operational•1972: ARPAnet demonstrated publicly, NCP (Network

Control Protocol) first host-host protocol, first e- mail program. ARPAnet has 15 nodes

1961-1972: Early packet-switching principles

4

Internet History

•1970: ALOHAnet satellite network in Hawaii•1973: Metcalfe’s PhD thesis proposes Ethernet•1974: Cerf and Kahn - architecture for

interconnecting networks•late70’s: proprietary architectures,DECnet, SNA,

XNA•1979: ARPAnet has 200 nodes

1972-1980: Internetworking, new and proprietary nets

5

Internet History

Cerf and Kahn’s internetworking principles:– minimalism, autonomy-no internal changes required to interconnect networks– best effort service model– stateless routers– decentralized control

define today’s Internet architecture

1972-1980: Internetworking, new and proprietary nets

6

Internet History

•1983: deployment of TCP/IP•1982: SMTP e-mail protocol defined •1983: DNS defined for name-to-IP-

address translation•1985: FTP protocol defined•1988: TCP congestion control

1980-1990: new protocols, a proliferation of networks

7

Internet History

•US networks: Csnet, BITnet, NSFnet, Minitel

•100,000 hosts connected to confederation of networks

1980-1990: new protocols, a proliferation of networks

8

Internet History

•Early 1990’s: ARPAnet decommissioned•1991: NSF lifts restrictions on commercial

use of NSFnet (decommissioned, 1995)

•early 1990s: Web–hypertext [Bush 1945, Nelson 1960’s]–HTML, HTTP: Berners-Lee–1994: Mosaic, later Netscape–late 1990’s: commercialization of the Web

1990, 2000’s: commercialization, the Web, new apps

9

Internet History

Late 1990’s – 2000’s:•more killer apps: instant messaging,peer-2-peer

file sharing (e.g., Naptser)•network security to forefront•est. 50 million host, 100 million+ users•backbone links running at Gbps•now: 10-40 Gbps (youtube, social

networking)

1990, 2000’s: commercialization, the Web, new apps

10

The (capital “I”) Internet The world-wide network of TCP/IP networks Different people or organisations own different

parts Different parts use different technologies Interconnections between the parts Interconnections require agreements

sale/purchase of service contracts “peering” agreements

No central control or management

11

The principle of “Internetworking” We have lots of little networks Many different owners/operators Many different types

Ethernet, dedicated leased lines, dialup, optical, broadband, wireless, ...

Each type has its own idea of low level addressing and protocols

We want to connect them all together and provide a unified view of the whole lot (treat the collection of networks as a single large internetwork)

12

What’s the Internet•millions of connected computing devices: hosts, end-systems

–PC’s workstations, servers–PDA’s phones,

•communication links–fiber, copper, radio, satellite

•routers: forward packets (chunks) of data through network

local ISP

companynetwork

regional ISP

router

workstationservermobile

13

TCP/IP Architecture Overview• The TCP/IP model can be broken down into

four layers:– Application– Transport– Internet– Physical Network Interface

• Application layer provides access to network resources. It defines rules, commands, and procedures for client to talk to a service running on a server

14

TCP/IP Architecture Overview (continued)

• Transport layer is responsible for preparing data ready to be transported across the network

• Internet layer is responsible for logical addressing and routing

• Physical Network Interface layer consists of the network card driver and the network card itself

15

TCP/IP Protocol

16

The TCP/IP Model

Network layer

PPP ATM Optics ADSL Satellite 3GEthernet

IP

UDPTCP

HTTP FTP Telnet DNSSMTP Audio Video

RTP

Physical and Data link layer

Application layer

Transport layer

17

Layer Interaction:TCP/IP Model

Host Router Host

Application

TCP or UDP

IP

Link

Physical

IP

Link Link

IP

Link Link

Application

TCP or UDP

IP

Link

PhysicalPhysical

Router

18

Layer Interaction:The Application Layer

Host Router Host

Application

TCP or UDP

IP

Link

Physical

IP

Link Link

IP

Link Link

Application

TCP or UDP

IP

Link

PhysicalPhysical

Router

Applications behave as if they can talk to each other, but in reality the application at each side talks to the

TCP or UDP service below it.

The application layer doesn't care about what happens at the lower layers, provided the transport layer

carries the application's data safely from end to end.

19

Layer Interaction:The Transport Layer

Host Router Host

Application

TCP or UDP

IP

Link

Physical

IP

Link Link

IP

Link Link

Application

TCP or UDP

IP

Link

PhysicalPhysical

Router

The transport layer instances at the two ends act as if they are talking to each other, but in reality they are each talking to the IP layer below it. The transport

layer doesn't care about what the application layer is doing above it.

The transport layer doesn't care what happens in the IP layer or below, as long as the IP layer can move

datagrams from one side to the other.

20

Layer Interaction:The Network Layer (IP)

Host Host

Application

TCP or UDP

IP

Link

Physical

IP

Link Link

IP

Link Link

Application

TCP or UDP

IP

Link

PhysicalPhysical

Router

The IP layer works forwards messages hop by hop from one side to the other side.

The IP layer has to know a lot about the topology of the network (which host is connected to which router,

which routers are connected to each other), but it doesn't care about what happens at the upper layers.

Router

21

Layer Interaction:Link and Physical Layers

Host Router Host

Application

TCP or UDP

IP

Link

Physical

IP

Link Link

IP

Link Link

Application

TCP or UDP

IP

Link

PhysicalPhysical

Router

The link layer doesn't care what happens above it, but it is very closely tied to the physical layer below it.

All links are independent of each other, and have no way of communicating with each other.

22

A Flow of Application messages across TCP/IP layers

Messages (UDP) or Streams (TCP)

Application

Transport

Internet

UDP or TCP segment

IP Packets

Network-specific frames

MessageLayers

Underlying network

Physical Network interface

23

Encapsulation of a message transmitted via TCP over an Ethernet

Application message

TCP header

IP header

Ethernet header

Ethernet frame

port

TCP

IP

24

Layering: physical communication

applicationtransportnetwork

linkphysical

applicationtransportnetwork

linkphysical application

transportnetwork

linkphysical

applicationtransportnetwork

linkphysical

networklink

physical

data

data

25

Application Layer Protocols• There are many Application layer protocols,

each of which is associated with a client application and service provided by a server (Client/Server Model)–HTTP–FTP–TELNET–SMTP–POP3– IMAP4

26

Application Model

27

HTTP• Hypertext Transfer Protocol (HTTP) is the most

common protocol used on the Internet today• HTTP defines the commands that Web browsers can

send and how Web servers are capable of respondingFTP

• File Transfer Protocol (FTP) is file-sharing protocol• FTP is implemented in stand-alone FTP clients as well

as in Web browsers • It is safe to say that most FTP users today are using

Web browsers

Application Layer Protocols

28

Application Layer Protocols

TELNET• Telnet is a terminal emulation protocol

that is primarily used to connect remotely to UNIX and Linux Systems

• The Telnet protocol specifies how a telnet server and telnet client communicate

29

SMTP• Simple Mail Transfer Protocol (SMTP) is used to send

and receive e-mail messages between e-mail servers that are communicating

• It is used by e-mail client software, such as Outlook Express, to send messages to the server

• SMTP is never used to retrieve e-mail from a server when you are reading it

• Other protocols control the reading of e-mail messages

Application Layer Protocols

30

POP3• Post Office Protocol version 3 (POP3) is the most

common protocol used for reading e-mail messages

• This protocol has commands to download messages and delete messages from the mail server

• POP3 does not support sending messages• POP3 supports only a single inbox and does not

support multiple folders for storage on the server

Application Layer Protocols

31

IMAP4• Internet Message Access Protocol version 4

(IMAP4) is another common protocol used to read e-mail messages

• IMAP4 can download message headers only and allow you to choose which messages to download

• IMAP4 allows for multiple folders on the server side to store messages

Application Layer Protocols

32

Transport Layer Protocols• Transport layer protocols (TCP & UDP) are

responsible for getting data ready to move across the network

• The most common task performed by Transport layer protocols is breaking entire messages down into segments suitable to form packets

• Transport layer protocols use port numbers• When a segment is addressed to a particular

port, the Transport layer protocol knows to which service to deliver the packet

33

TCP• Transmission Control Protocol (TCP) is the most

commonly used Transport layer protocol for most Internet services

• TCP is connection-oriented and reliable• Connection-oriented means that TCP creates and

verifies a connection with a remote host before sending information

• Verifies that the remote host exists and is willing to communicate before starting the conversation

• Provides flow control, segmentation, and error control

34

TCP

• Connection-oriented– Establishes a connection before transmitting data– Three-way handshake

SYN

SYN/ACK

ACK

35

TCP• Error control & Flow control

– Require acknowledgements from receiver to ensure data was received correctly

– Checksum• Unique character string allowing receiving node to

determine if arriving data unit exactly matches data unit sent by source

• Ensures data integrity

Send data, wait for ACK

ACK

Send more data, wait for ACK

36

• Segmentation– Breaking large data units received from Session layer into

multiple smaller units called segments– Increases data transmission efficiency– MTU (maximum transmission unit): Largest data unit network

will carry (Ethernet default: 1500 bytes)

• Sequencing– Method of identifying segments belonging to the same

group of subdivided data

• Reassembly– Process of reconstructing segmented data units

TCP

37

Transport Layer (cont’d.)

Figure 2-2 Segmentation and reassembly

38

1 2 3 4 5 6 7 8 9 10 11 User Data

1 Source ID or port 16 bits2 Destination ID or port 16 bits3 Sequence number 32 bits4 ACK number 32 bits5 Header length 4 bits6 Unused 6 bits7 Flags 6 bits8 Flow control 16 bits9 CRC 16 16 bits10 Urgent pointer 16 bits11 Options 16 bits

TCP Segment

39

UDP• User Datagram Protocol (UDP)

– Not as commonly used as TCP– Used for different services– Connectionless and unreliable

• UDP is the appropriate if– Unconcerned about missing packets– Want to implement reliability in a special way

• Streaming audio and video are in this category

40

UDP – Segment

1 2 3 4 User Data

1 Source ID or port

2 Destination ID or port

3 Length

4 Checksum

41

TCP versus UDP

• TCP is connection-oriented and reliable–Like registered mail

• UDP is connectionless and unreliable–Like sending a message split on several

postcards and assuming that the receiver will be able to put the message together

42

Internet Layer Protocols

• Internet layer protocols are responsible for all tasks related to logical addressing

• An IP address is a logical address• Any protocol that is aware of other networks

exists at this layer• Each Internet layer protocol is very specialized• They include: IP, RIP and OSPF, ICMP, IGMP,

and ARP

43

IP• Internet Protocol (IP) is responsible for the logical

addressing of each packet created by the Transport layer to produce a complete IP Packet

• As each packet is built, IP adds the source and destination IP address to the IP packet

ICMP• Internet Control Messaging Protocol (ICMP) is used

to send IP error and control messages between routers and hosts

• The most common use of ICMP is the ping utility

Internet Layer Protocols

44

IP Packet version 4

1 Version number 4 bits2 Header length 4 bits3 Type of Service 8 bits4 Total length 16 bits5 Identifiers 16 bits6 Flags 3 bits7 Packet offset 13 bits8 Hop limit 8 bits

9 Protocol 8 bits10 CRC 16 16 bits11 Source address 32 bits12 Destination Address 32 bits13 Options varies14 User data varies

1 2 3 4 5 6 7 8 9 10 11

IP4

12 13 14

45

IGMP• Internet Group Management Protocol (IGMP) is used

for the management of multicast groups• Hosts use IGMP to inform routers of their

membership in multicast groups• Routers use IGMP to announce that their networks

have members in particular multicast groups• The use of IGMP allows multicast packets to be

distributed only to routers that have interested hosts connected

Internet Layer Protocols

46

ARP• Address Resolution Protocol (ARP) is used to

convert logical IP addresses to physical MAC addresses

• This is an essential part of the packet delivery process

Internet Layer Protocols

47

Network Interface Layer Protocols

• Most of the common Network Interface layer protocols are defined by the Institute of Electrical and Electronics Engineers (IEEE)

48

• Internet Protocol (IP):– a protocol used in the internet layer.– IP makes use of the existing networks to deliver

information, where these networks may use a variety of protocols.

• Each computer has two addresses:– hardware address: used by the underlying network

protocol for deliver data frame; – IP address: used by the internetworking protocols

for deliver IP Packet.• Hardware address is also known as physical address.

IP Addresses

49

Types of addresses used on hosts

Address Example Software Example Address

Application Layer Web browser www.cba.uga.eduNetwork Layer TCP/IP 128.192.98.5:80Data Link Layer Ethernet 00-0C-00-F5-03-5A

50

IP Addressing Scheme• Each computer / router is assigned a unique IP address

having 32 bits.• Each IP address has two parts:

– The prefix (network ID or NetID) specifies the network to which the computer is attached.

– The suffix (HostID) specifies a particular computer on a network.

• Problem– Given only 32 bits, how many bits should be allocated to

the prefix and the suffix?• around 4 billion addresses.

IP Addresses

51

IP Addressing Scheme• Considerations

– If the prefix has many bits (large prefix, small suffix), there are many networks you can built but each network can only have a few computers.

– If the prefix has a few bits (small prefix, large suffix), there are only few networks you can built but each network can have many computers.

IP Addresses

52

Subnet Masks

• A subnet mask defines which part of its IP address is the network ID and which part is the host ID

• Subnet masks are composed of four octets just like an IP address

• Wherever there is a 255 in the subnet mask, that octet is part of the network ID

• Wherever there is a 0 in the subnet mask, that octet is part of the host ID

53

Subnet Masks (continued)• A computer uses its subnet mask to determine

– Which network it is on– Whether other computers are on the same

network or a different network• If two computers on the same network are

communicating, then they can deliver packets directly to each other

• If two computers are on different networks, they must use a router to communicate

54

Subnet Masks (continued)

55

• The IP addressing scheme defines three primary classes (A,B,C), where each class has a distinct prefix/suffix size, and two reserved classes (D&E).

• The internet can accommodate large networks, medium networks, and small networks.

• Classes A, B, C are the primary classes. The IP addresses of computers and routers belong to these classes.

• Class D is used for multicasting. When a packet is sent to an IP multicast address, all the computers sharing this address will receive this packet.

• Class E addresses are considered experimental and are not used

IP Address Classes

56

The Classful Addressing Scheme

57

• The first decimal value defines the class of the IP address as follows:

58

IP Address Classes & Default Subnet Masks

59

• In each primary class, the number of networks and the number of computers per network are as follows:

• Each packet sent across the internet contains:– the IP address of the source, and– the IP address of the destination.

60

• Dotted Decimal Notation– Commonly we use the dotted decimal notation to

represent the 32-bit IP address.• more convenient for human to manipulate

– Each octet (8-bit) is expressed as a decimal value, and adjacent decimal values are separated by a dot.

– Example:

61

• Loopback address– 127.x.x.x– intended for use in testing TCP/IP and for inter-process

communication on the local computer• Other special value of primary classes:

62

Assigning IP Addresses• Assigning Prefix Address

– Each network must have a unique prefix address throughout an internet.

– To connect a network to the global internet, an organization obtains a unique prefix address from the Internet Service Provider (ISP).

– In turn, the ISP coordinates with a central organization (the Internet Assigned Number Authority (IANA, on or before 1998); the Internet Corporation for Assigned Names and Numbers (ICANN, after 1998)) to ensure the uniqueness of the prefix.

– To connect a network to a private internet (Intranet), the organization can determine the prefix while ensuring its uniqueness.

63

Assigning IP Addresses• Assigning Suffix Address

– Each computer must have a unique suffix address in the same network; while two computers in two different networks can have identical suffix address or HostID.

– If the suffix is 00…0 or 11…1, the corresponding IP addresses have special meaning. Do not assign these suffixes.• An IP address with suffix equal to 00…0 is used to refer to

the network itself.• An IP address with suffix equal to 11…1 is a directed

broadcast address, i.e., it refers to all hosts on the network.

64

• Example– An organization wants to form a private TCP/IP

internet with four networks, where one network is large (with many computers), two are medium, and one is small.

– Firstly, assign a unique prefix to each network:• Assign a class A prefix for the large network (say, 10).• Assign a class B prefix for each of the two medium networks

(say, 128.10 and 128.11).• Assign a class C prefix for the small network (say, 192.5.48).

– Secondly, assign a unique suffix to each computer within each network:

65

66

Private IP Addresses• You can use these addresses on any

private LAN. • You CANNOT use them on the internet. • Internet routers will block them.

67

Default Gateway• Default gateway is another term for router• If a computer does not know how to deliver a

packet, it gives the packet to the default gateway to deliver

• Routers can distinguish multiple networks and how to move packets between them

• Routers can also figure out the best path to use to move a packet between different networks

68

Classful IP Address A classful network had a “natural” or “implied”

prefix length or netmask: Class A: prefix length /8 (netmask 255.0.0.0) Class B: prefix length /16 (netmask 255.255.0.0) Class C: prefix length /24 (netmask 255.255.255.0)

Modern (classless) routing systems have explicit prefix lengths or netmasks You can't just look at an IP address to tell what the prefix

length or netmask should be. Protocols and configurations need explicit netmask or prefix length.

69

Classless addressing

Internet routing and address management today is classless

CIDR = Classless Inter-Domain Routing routing does not assume that class A, B, C

implies prefix length /8, /16, /24

An ISP gets a large block of addresses e.g., a /16 prefix, or 65536 separate addresses

70

Classless addressing

• Allocate smaller blocks to customers– e.g., a /26 prefix (64 addresses) to 4 customers

for their medium public networks, a /28 prefix (16 addresses) to 32 customers for their medium public networks, and a /29 prefix (8 addresses) to another 64 customers for their small public networks (and some space left over for other customers)

71

Binary presentation of Classless IP

137.158.128.0/17 (netmask 255.255.128.0)

198.134.0.0/16 (netmask 255.255.0.0)

205.37.193.128/26 (netmask 255.255.255.192)

1000 1001 1001 1110 1 000 0000 0000 0000 1111 1111 1111 1111 1 000 0000 0000 0000

1100 0110 1000 0110 0000 0000 0000 0000 1111 1111 1111 1111 0000 0000 0000 0000

1100 1101 0010 0101 1100 0001 10 00 0000 1111 1111 1111 1111 1111 1111 11 00 0000

72

Classless addressing exercise• Consider the address block 133.27.162.0/28

and 133.27.163.48/29. What are the IP addresses range can you

obtain from each block? in prefix length notation netmasks in decimal IP address ranges

What blocks are still available (not yet allocated)?

73

Sockets and Ports

74

Sockets and Ports• Processes assigned unique port numbers• Process’s socket

– Port number plus host machine’s IP address

• Port numbers– Simplify TCP/IP communications – Ensures data transmitted correctly to the specific

application among multiple applications running on same host

• Example– Telnet port number: 23– IPv4 host address: 10.43.3.87– Socket address: 10.43.3.87:23

75

Sockets and Ports (cont’d.)

Figure 4-12 A virtual connection for the Telnet service

76

Sockets and Ports (cont’d.)• Port number range: 0 to 65535• Three types

– Well Known Ports• Range: 0 to 1023• Operating system or administrator use

– Registered Ports• Range: 1024 to 49151• Network users, processes with no special privileges

– Dynamic and/or Private Ports• Range: 49152 through 65535• No restrictions

77

Sockets and Ports (cont’d.)

Table 4-3 Commonly used TCP/IP port numbers