Computer Networks

1
Computer Networks

• Definition and goals of networks
• Classification of networks
• Transmission media
• Types of network
• Packet switching
• Circuit switching
• Virtual-Packet circuit switching
• LANs and WANs
• The Internet
• Protocol stacks

2
The problem

• Problem Given a set of devices that want to
  exchange information.
• (Device telephone, computer, terminals, etc.)
• Simple Solution Connect each pair of devices by
  a dedicated point-to-point link
• The simple solution is sufficient if the number
  of devices is small.

3
The problem (cont)

• With a large number of devices it is not
  practical to connect each pair of devices

4
Solution ??

• A communication network provides a general
  solution to the problem of connecting many
  devices
• Connect each device to a network node
• Network nodes exchange information and carry the
  information from a source device to a destination
  device
• Note Network nodes do not generate information

5
What is a computer network ?
An interconnection of autonomous computers (as
opposed to communication between separate but
interdependent parts of a single computer)
6
Goals of computer networks

• Access to remote resources
• Human communication
• Mobile computing
• Computing power through parallelism
• Optimising resources - load balancing
• Incremental growth of computer systems (reduced
  cost and risk)
• Increased robustness through graceful degradation

7
Uses of computer networks

• Email, World Wide Web, Video Conferencing, File
  Transfer, Collaborative Virtual Environments,
  Remote control of robots and machines, Dial up
  databases, Web-casting, Distributed Programs,
  Hacking, Banking, Internet telephone

8
Classification of networks

• By size
• Local area networks (LANs)
• Metropolitan Area Networks (MANs)
• Wide area networks (WANs)
• By media type
• Point to point networks vs. broadcast networks
• By mobility
• Mobile networks vs fixed networks
• The Internet?

9
Issues for local and wide area

• Scale - geographic and number of computers
• Speed - bandwidth and latency
• Management
• Security
• Billing
• Heterogeneity (and standards)

10
Transmission media
The transmission medium is the physical path
between transmitter and receiver in a data
transmission Transmission medium can be Simplex
- Transmission in one direction only Half-duplex
- Transmission in both directions but not at the
same time Full duplex (duplex) - Simultaneous
transmission in both directions
11
Transmission media

• Copper Wire
• Glass Fibre
• Microwave
• Infrared

12
Comparing transmission media

• Various transmission media with different
  properties in terms of
• range
• capacity to carry information
• cost
• directionality
• reliability (noise and blocking)

13
Types of network

• Packet switching store and forward messages
  (Internet)
• Circuit switching Carry bit streams (old
  telephone network)

14
Circuit switching

• In a circuit-switched network, a dedicated
  communication path is established between two
  stations through the nodes of the network
• The dedicated path is called a circuit-switched
  connection or circuit

15
Circuit switching (cont)

• A circuit occupies a fixed capacity of each link
  for the entire lifetime of the connection.
  Capacity unused by the circuit cannot be used by
  other circuits
• Data is not delayed at the switches

16
Circuit switching (cont)

• Circuit-switched communication involves three
  phases
• 1. Circuit Establishment
• 2. Data Transfer
• 3. Circuit Termination
• Busy Signal if capacity for a circuit not
  available.
• Most important circuit-switching networks
• Telephone networks
• ISDN (Integrated Services Digital Networks)

17
Circuit switching (cont)
18
Circuit switching (cont)
19
Packet switching

• Most networks transmits data in small blocks
  called packets
• helps in detect transmission errors
• gives fair access for a shared connection between
  many computers
• These are packet networks or packet switching
  networks

20
Packet switching

• Data are sent as formatted bit-sequences,
  so-called packets.
• Packets have the following structure
• Header and Trailer carry control information

Header
Data
Trailer
21
Packet switching (cont)

• Each packet is passed through the network from
  node to node along some path (Routing)
• At each node the entire packet is received,
  stored briefly, and then forwarded to the next
  node (Store-and-Forward Networks)
• No capacity is allocated for packets

22
Packet switching (cont)
23
Packet switching (cont)
24
Virtual-Circuit Packet switching

• Virtual-circuit packet switching is a hybrid of
  circuit switching and packet switching
• All data is transmitted as packets
• All packets from one packet stream are sent along
  a pre-established path (virtual circuit)
• Guarantees in-sequence delivery of packets

25
Virtual-Circuit Packet switching (cont)

• Communication with virtual circuits (VC) takes
  place in three phases
• 1. VC Establishment
• 2. Data Transfer
• 3. VC Disconnect
• Note Packet headers dont need to contain the
  full destination address of the packet

26
Virtual-Circuit Packet switching (cont)
27
Switching Summary
28
Summary (cont)

• Telephone network
• The largest world-wide computer network,
  specialised for voice.
• Circuit switching
• Internet
• A new global and public information structure
• Datagram packet switching
• ATM
• Was intended to replace telephone networks and
  data networks but lost momentum due to the
  success of the Internet
• Virtual-Circuit packet switching

29
Local Area Networks

• Shared LANs invented in the 1960s
• Rely on computers sharing a single medium
• Computers co-ordinate their access
• Low cost
• But not suitable for wide area due to
  communication delays inhibiting co-ordination

30
Local Area Networks (cont)

• LANs may be categorised according to topology
• Star
• Ring
• Bus

31
Local Area Networks (cont)
Computers connected to a network
Hub
Star - ATM
32
Local Area Networks (cont)
Connection from one computer to another
Ring -IBM Token Ring
33
Local Area Networks (cont)
Cable connection (Ethernet)
Bus - Ethernet
34
Wide Area Networks

• Key goal is scale-ability
• Composed of many packet switches
• Store and forward approach instead of shared
  medium
• Each packet switch does next hop routing by using
  a local routing table

35
Example WAN technologies

• Phone network
• ISDN
• ATM

36
Wide Area Networks

• Key goal is scale-ability
• Composed of many packet switches
• Store and forward approach instead of shared
  medium
• Each packet switch does next hop routing by using
  a local routing table

37
What can go wrong ??

• Bit-level errors (electrical interference)
• Packet-level errors (congestion)
• Link and node failures
• Messages are delayed
• Messages are deliver out-of-order
• Third parties eavesdrop.

38
The Internet

• Global inter-connection of different networks
• The Internet is a datagram packet-switching
  network
• Station or Endsystems are called Hosts
• Nodes are called Routers

39
HTTP- The World Wide Web Protocol

• HTTP
• Web Content
• Caching

40
Web Architecture

• The Web is a client-server architecture
• Servers and browsers use the HTTP protocol

41
Server and Client Examples

• Server software
• Apache
• Microsoft IIS
• Netscape Web Server
• Client software
• Netscape Navigator
• Internet Explorer

42
HTTP

• 80 of Internet flows are HTTP connections
• Early protocol is HTTP 0.9
• read, execute (only).
• Today we mostly use HTTP 1.0
• read, input, delete, execute, ...
• New version HTTP 1.1
• performance optimizations

43
HTTP Overview

• Client (browser) sends HTTP request to server
  (usually port 80)
• Request specifies URL and operation
• Server performs operation on URL
• Server sends response
• Client parses response possibly follows link to
  (request) another URL.

44
HTTP Operations

• GET retrieves URL (most widely used)
• HEAD retrieves only response header
• POST posts data to server
• PUT puts page on server
• DELETE deletes page from server

45
HTTP Headers

• HTTP headers are in pure text
• Request header
• operation, requested URL, protocol version,
• blank line
• Response header
• protocol version, operation status, content
  length, content type,
• blank line

46
Simple HTTP Request and Reply

• Request
• GET http//www.server.com/page.html HTTP/1.0
• ltblank linegt
• Response
• HTTP-Version HTTP/1.0 200 OK
• Content-Length 3012
• Content-Type text/html
• ltblank linegt
• ltbodygt

47
Example

• Establish a connection to web server port 80
  using
• telnet www.scarlet.cs.nott.ac.uk 80
• When the connection is established try
• GET /index.html HTTP/1.0
• Hit Enter twice to create a blank line after
  header

48
HTTP 1.0

• Client opens a separate TCP connection for each
  requested object
• Object is served and connection is closed
• Advantages
• maximum concurrency
• Limitations
• TCP connection setup/tear-down overhead
• TCP slow start every time

49
Authentication

• Server may request client authentication by
  issuing code 401 in response header, e.g.
• HTTP-Version HTTP/1.0 401 Unauthorized
• WWW-Authenticate Basic RealmTopSecretDoc
• Client must then send valid ID and password in
  request header
• Authentication tokens are commonly named cookies

50
Sessions

• HTTP 1.0 is stateless (no concept of client
  sessions)
• To maintain session state (e.g., a shopping cart)
  server must perform authentication on every
  request
• State can be encoded in cookies
• Back-end applications invoked by server to map
  authenticated cookies to session state

51
HTTP 1.1

• To avoid a connection per object model, HTTP 1.1
  supports persistent connections
• Client opens TCP connection to server
• All requests use same connection
• Problems
• Less concurrency
• Server does not know when to close idle
  connections

52
Static Content and HTML

• Most static web content is written in HTML
• HTML allows
• Text formatting commands
• Embedded objects
• Links to other objects (Hypertext)
• Server need not understand or interpret HTML

53
Fetching Multiple Objects

• Most web-pages contain embedded objects (e.g.,
  images, backgrounds, etc)
• Browser requests HTML page
• Server sends HTML file
• Browser parses file and requests embedded objects
• Server sends requested objects

54
Dynamic Content

• Web pages can be created as requests arrive
• Advantages
• Personalization (e.g., my.yahoo.com),
• interaction with client input
• interaction with back-end applications
• Disadvantages
• Performance penalty
• Generating dynamic content (CGIs, ASPs)

55
Common Gateway Interface (CGI) Scripts

• URLs with a .cgi extension denote CGIs
• The script is a program (e.g., C, JAVA, )
• When the URL is requested, server invokes the
  named script, passing to it client info
• Script outputs HTML page to standard output
  (redirected to server)
• Server sends page to client

56
Fast CGI

• CGI scripts are invoked on each request
• Invocation overhead is large (e.g., UNIX fork())
• Fast CGIs are invoked a priori and blocked on
  server I/O
• Client requests unblock script
• Script outputs page and blocks again

57
Active Server Pages (ASPs)

• Active server pages are HTML documents with
  extensions for embedded program execution
• When request arrives, server fetches and parses
  the HTML document
• Server executes embedded executable code and
  plugs output into page
• Expanded page is sent to client

58
Web Caching

• Caching improves perceived performance

59
Types of Web Caching

• The cache retains most popular URLs
• Client caching
• browser caches URLs for future access
• Proxy caching
• proxy server caches most requested URLs
• Server-side caching
• server-side cache reduces load on server

60
Caching Policies

• A URL is cached when it is requested
• Cache must replace content when full
• Cache replacement policies
• LRU (replace least recently used page)
• LFU (replace least frequently used page)
• Other considerations
• page size, client importance, cache distance from
  server

61
LFU

• Replaces least frequently used page
• Optimal replacement policy if
• all pages have same size
• page popularity does not change
• In practice has disadvantages
• slow to react to popularity changes
• needs to keep statistics (counter) for every page
• does not consider page size

62
LRU

• Replaces least recently used page
• maintains linked list of pages
• each newly fetched page is put on head of list
• tail page is deleted when storage is exceeded
• Exploits temporal locality
• Performs better than LFU in practice
• No per-page counters
• Used in todays caches (e.g., Squid)

63
Cache Consistency Management

• Problem with caching
• Cached pages may get modified on server
• Clients accessing cache will get stale copy
• Possible solutions
• Finite time-to-live
• Invalidation
• Invalidation contracts

64
Time-to-live Conditional GET

• Server specifies a time-to-live for requested
  page
• Proxy caches page and serves subsequent clients
  until time-to-live expires
• Proxy sends if-modified-since query to server on
  subsequent request(s)
• Server sends page if it was modified
• (presently used method)

65
Cache Invalidation

• Server remembers where its pages are cached
• When a page is modified, server sends
  invalidation messages to caches
• Caches mark page as stale
• Subsequent client requests will fetch fresh copy
  from server

66
Invalidation Contracts

• Invalidation traffic can get very large
• To limit invalidations server gives invalidation
  contracts to caches
• Contracts have expiration time
• When content changes server notifies only those
  caches whose contract has not expired