Internet Technology

1
Internet Technology Network Programming

• Dan Wang

2
Outline

• Data Communication Model
• Network Protocol Model
• OSI Reference Model
• TCP/IP Protocol Suite
• Internet Addressing
• IP Address
• Port Address
• Socket Programming

3
Data Communication Model

• Source Generates data to be transmitted
• Transmitter Converts data into transmittable
  signals
• Transmission System Carries data
• Receiver Converts received signal into data
• Destination Takes incoming data

4
Network Protocol Model

• Two entities (source and destination) need
  network protocols to communicate with each other
  over the communication network.
• The network protocols are organized as
  multi-layer hierarchies
• OSI Reference Model
• TCP/IP Protocol Suite

5
OSI Reference Model

• ISO established the OSI model as a standard for
  data communication.
• Comprehensive model 7 layers
• Each layer performs a subset of the required
  communication functions
• Each layer relies on the next lower layer to
  perform more primitive functions
• Each layer provides services to the next higher
  layer
• Changes in one layer should not require changes
  in other layers, they are hided by interface
  between two layers
• Peer layers sit at both sides
• Organized but too complicated and not easy to
  implement or deploy
• Today, OSI model is used as a tool for academic
  teaching and conceptual model to organize
  networking functionality

ISO International Standard Organization OSI
Open System Interconnection
6
OSI Reference Model

• Physical Layer
• Representation of bits as electrical signals
• Physical interface to media
• Data Link Layer
• Framing of data into units of frame.
• Control data transmission from point-to-point or
  broadcast, including flow and error control
• Network Layer
• Delivery of packets across network from source
  to destination
• Logical addressing to provide unified addressing
  scheme
• Routing
• Transport Layer
• First end-to-end protocol that deals with
  end-to-end issues.
• Provide services such as error, congestion, flow
  controls.

7
OSI Reference Model

• Session Layer
• Establishes dialog and synchronizes sessions
  between two end systems.
• E.g. checkpoint session.
• Presentation Layer
• Concern with the syntax and semantic of
  information exchanged between two systems
• Translate data representation to a common format,
  e.g. from source-dependant format to a common
  format at the source and from a common format to
  a destination dependant format
• Service encryption and compression
• Application Layer
• End user software that accesses the network
• E.g. Email, FTP, web browser

8
OSI Reference Model
9
TCP/IP Protocol Suite

• Internet was developed prior to the OSI model, it
  adapt network protocol suite that was developed
  by the US DARPA for its ARPANET
• Dominant commercial protocol used by the global
  Internet
• Made up of 5 layers
• Physical Layer
• Physical interface between data transmission
  device (e.g. computer) and transmission medium or
  network
• Characteristics of transmission medium, signal
  rate and encoding, etc

10
TCP/IP Protocol Suite

• Make up of 5 layers (contd)
• Data Link Layer
• Logical interface between end system and network
• Exchange of data between end system and network
• Internet layer (IP)
• Best effort transfer of data, no delivery
  guarantee
• Provide uniform logical interface for different
  networks
• Routing functions across multiple networks
• Implemented in end systems and routers
• Transport layer (TCP/UDP/SCTP)
• End to end transfer of data
• Provide reliability mechanism (TCP) or not (UDP)
• Hides detail of underlying network
• Application layer
• Support for user applications, e.g. HTTP, FTP,
  SMPT

11
TCP/IP Protocol Suite
12
Web on Internet Architecture
13
Layered Web Architecture

• Web uses the Internet as the underlying network
  infrastructure. The complete architecture of the
  Web is best illustrated using a layered approach.
  
• Datalink Layer Implements the link interface to
  the physical medium and provides point-to-point
  services to move frames around the network.
• Network Layer Overlays a layer of network
  software to provide uniform network layer
  services such as routing across different
  datalink layers.
• Transport Layer Provides end-to-end delivery of
  packets that resides only in end host.
  Specifically, it adds services to the network
  layer to provide reliable transmission.
• WWW LayerProvides the concept of hyperlinking
  documents across the network. The service uses
  technologies such as HTML, HTTP, and XML
• Server-Side Computing Concern with extension of
  server-side functionality that is beyond
  delivering of static documents.
• Client-Side Computing Concern with client-side
  extension to basic web.

14
Network Layer

• IP (Internet Protocol)
• Define the basic unit of data transfer used in
  TCP/IP internet
• Best effort and connectionless delivery
• Make an attempt for delivery, but does not
  guarantee delivery
• Provide a connectionless unreliable packet
  delivery service
• In the Internet, as routers become overrun or
  change routes, datagram can be lost, duplicated
  or out of order.
• Provide routing functions across multiple
  networks
• Other protocols at Network Layer
• ARP for address translation from logical to
  physical address
• RARP for address translation of physical to
  logical address
• ICMP allows gateways and hosts to send
  notification of problems back to sender
• IGMP facilitates transmission of multicast
  messages

15
Transport Layer

• Provide end-to-end data transfer for applications
• Transport Layer Protocols
• UDP (User Datagram Protocol)
• TCP (Transport Control Protocol)
• SCTP (Stream Control Transmission Protocol)

16
TCP

• Provide a connection-oriented, ordered,
  full-duplex, reliable service.
• Connection-oriented service
• An end-to-end virtual circuit connection is
  established before data exchange can happen.
• Connection is terminated when finished.
• Use 3-way handshake for connection establishment
  and termination
• Similar to the phone system
• Ordered service
• Data bytes are delivered in sequence.
• Full-duplex connection
• Can send and receive data at the same time
• Reliable service
• Delivery of data is guaranteed
• Flow control, congestion control and error
  control
• TCP datagram is encapsulated in IP datagram for
  delivery.
• IP identifies the destination computer (How?)
• TCP identifies the application on the destination
  computer (How?)
• Used by many of the popular applications in the
  Internet, such as Telnet, FTP, SMTP, HTTP.

17
UDP

• Provide connectionless, best-effort, not reliable
  datagram service.
• Connectionless service
• Provide the ability to send and receive messages.
• No connection is required before data transfer.
• Similar to the postal system.
• Best-effort service
• Not reliable
• Message can be delayed, lost, or duplicated, or
  arrive out of order
• Application accepts full responsibility for
  errors.
• Used by multimedia applications, DNS

18
Internet Addressing

• Challenges
• How to send packets from source to destination
  across the Internet?
• How to send packets to a specific application?
• Three different levels of addresses
• Physical address
• IP address
• Port address

19
Internet Addressing
20
IP Address

• Uniquely identify a host on the Internet.
• There are two popular ways
• Using IP address
• Using names registered in the domain name system
  (DNS).
• IP Address
• Each host on the Internet is assigned a 32-bit
  unique address (4 bytes).
• An IP address is assigned to a single host only.
• A host may have more than one IP address
  (multi-homed host).
• Dotted representation
• Internet addresses are represented in the form of
  four integers separated by decimal points known
  as dotted representation.
• Examples 131.91.81.40, 131.91.131.67
• Has netid and hostid
• Five classes Class A, B, C, D, E
• Loopback address (127.0.0.1) is used by a host to
  test its internal software

21
Domain Name System (DNS)

• Translation of hostnames to and from IP addresses
• E.g., www.polyu.edu.hk, rocket.comp.polyu.edu.hk
• Why hostname, not only IP address?
• Easy to remember
• Re-locatable
• Same web site available in multiple hosts
• DNS resolver performs two functions
• Coverts name to IP address
• Converts IP address to name
• A database of mapping is cached

22
DNS

• How does a client locate DNS?
• Configured with a list of DNS addresses (How to
  check?)
• Discover thru DHCP.
• DNS is a distributed database of hierarchy
  servers.
• The DNS naming convention is hierarchical.
• Written in the local-most level first and the
  top-most level last fashion.
• Each organization connected to Internet is
  responsible to maintain its DNS server.

23
DNS

• By late 90s, assigning of IP domain names are
  controlled and managed by ICANN.
• Top level domains are handled by collection of
  root servers, which assist query to recursively
  resolve domain name.
• Can you think of an example where an application
  would want to perform reverse resolution of IP
  address to name?
• Logging of clients hostname by Server
• Restrict access to certain resources based on
  domain
• Targeted advertisements

24
DNS

• Busy website may have replication on multiple
  machines with the same contents
• Located in different geographical locations to
  spread loading, e.g. www.cnn.com may correspond
  to 5 IP addresses located across 5 continents
• DNS have mapping of 1 hostname to multiple IP
  addresses
• Local DNS server can select one out of the list.
  (How?)

25
Port Address

• A (protocol) port is an abstraction used by
  TCP/UDP to distinguish applications on a given
  host.
• A port is identified by a 16-bit integer, known
  as a port address or a port number.
• Three ranges of port numbers
• Well-known ports
• Registered ports
• Dynamic ports

26
Port Address

• Well-known Ports
• Port numbers ranging from 0 to 1,023.
• A set of pre-assigned port numbers for specific
  uses.
• Are managed by ICANN.
• Registered Ports
• Port numbers ranging from 1,024 to 49,151.
• Not assigned or controlled by ICANN however
  their uses need to be registered via an
  ICANN-accredited registrar to prevent
  duplications.
• Dynamic Ports
• Port numbers ranging from 49,152 to 65,535.
• Neither assigned or registered. They can be used
  by anyone.
• These are ephemeral ports.
• Also known as private ports.

27
Well-Known Ports
28
Socket Programming

• What is a socket?
• Socket is a programming abstraction that enables
  application to interface directly to the
  transport layer and communicate over the network.
• Socket provides a network programming interface.
• Used for inter-process communication over the
  network.
• Used by a process to communicate with a remote
  system via a transport protocol.
• A socket provides a full-duplex communication
  mechanism.
• Application uses socket to communicate with the
  network.
• Data come from and go to the transport layer via
  a port number.
• Socket address combines IP address and a port
  number.

29
Socket Programming

• Sockets were first introduced in Berkeley UNIX.
• An extension of the UNIX abstraction of file I/O
  concepts.
• Now are commonly supported in almost all modern
  operating systems for intersystem communications.
• Sockets are popularly used in client/server
  computing.

30
Socket Programming

• In a client/server model as shown, we have the
  server listening to a socket port continuously
  for a service request.
• The client wanting to communicate with the server
  is required to know servers socket address (IP
  address and the port number).
• Upon receiving a request from the well-known
  port, the server will generate a new socket port
  and uses it to connect to the client.
• The new socket port will be bound to the port at
  the client-side to begin communications.TCP uses
  socket address as connection identification
  (IPclient, portclient), (IPserver, portserver)
• The server will return to listening at the
  well-known port for other requests.

31
Socket Programming Client-side

• In Java, socket programming is supported by the
  java.net package that provides a Socket class.
• Also includes the ServerSocket class, which
  servers can use to listen and accept connections
  to clients.
• The basic network programming structure at the
  client-side
• Open a socket Socket clientSocket new
  Socket(hostname, portNumber) (What is this
  socket, TCP socket or UDP socket?)
• Open an input and output streams to the socket.
• out new PrintWriter(clientSocket.getOutputStream
  (), true)
• in new BufferedReader(new InputStreamReader(clie
  ntSocket.getInputStream()))
• Read and write to the corresponding stream
  handlers to communicate. E.g. out.writeBytes()
  in.readLine()
• Close the streams out.close() in.close()
• Close the socket clientSocket.close()
• Ex1 echoClient.java

32
Socket Programming Server-side

• The basic network programming structure at the
  server-side
• Open a serverSocket, with its well-known port
  number
• ServerSocket listenSocket new
  ServerSocket(portNumber)
• If server successfully connects to a client, the
  Socket object is returned
• Socket clientSocket listenSocket.accept()
• Open an input and output streams to the socket.
• out new PrintWriter(clientSocket.getOutputStream
  (), true)
• in new BufferedReader(newInputStreamReader(clien
  tSocket.getInputStream()))
• Read and write to the corresponding stream
  handlers to communicate. E.g. out.writeBytes()
  in.readLine()
• Close the streams out.close() in.close()
• Close the socket clientSocket.close()
• Ex2 echoServer.java