i206: Lecture 21: Networking, WWW, and Internet Protocols

1
i206 Lecture 21Networking, WWW, and Internet
Protocols

• Marti Hearst
• Spring 2012

2
Network
Confidentiality Integrity Authentication
Distributed Systems
Security
Cryptography
Network
Standards Protocols
Inter-process Communication
Methodologies/ Tools
Principles
TCP/IP, RSA,
OperatingSystem
Application
Design
Formal models
Process
I/O
Finite automata regex
Context switch Process vs. Thread Locks and
deadlocks
Program
Algorithms
Analysis
Memory hierarchy
Memory
ALUs, Registers, Program Counter, Instruction
Register
Big-O
Compiler/ Interpreter
Assembly Instructions
Data Structures
Register, Cache Main Memory, Secondary Storage
Searching, sorting, Encryption, etc.
Machine Instructions
CPU
Op-code, operands Instruction set arch
Data storage
Stacks, queues, maps, trees, graphs,
Circuits
Lossless v. lossy Info entropy Huffman code
Decimal, Hexadecimal, Binary
Adders, decoders, Memory latches, ALUs, etc.
Gates
Data compression
Number Systems
Data
AND, OR, NOT, XOR, NAND, NOR, etc.
Boolean Logic
Numbers, text, audio, video, image,
Truth table Venn Diagram DeMorgans Law
Data Representation
Binary Numbers
Bits Bytes
3
Topics

• Network abstractions
• Network architecture
• How the WWW works, end to end
• Illustrated with the example of a web search
  engine

4
Network as Communication Channel
Source Coulouris, Dollimore and Kindberg
5
Network Cloud
Network
client server
6
Network Routers Links
Source Coulouris, Dollimore and Kindberg
7
Network More Details
Internet Service Providers
Customer Premises
Internet backbones
Telephone Network
Point of Presence
Backbone Provider 1
Router
Router
Tandem Switch
ISP
Local Exchange Carrier (LEC)

Exchange Point
DNS
Router
Local Egress Switch
Server
Local Ingress Switch
Content Provider
Packet Network
Local Loop
Headend
Backbone Provider 2
Cable ISP
Remote ISP
Analog Modem
xDSL Modem
Router
router
Cable Modem
Firewall
Client
Corporate LAN
Wireless ISP
Mobile Client
8
Network Utilities

• Run from Terminal in unix/mac
• Ping round trip time on an IP network from the
  originating host to the destination computer
• Traceroute displaying the route (path) and
  measuring transit delays of packets across an IP
  network
• ends a sequence of Internet Control Message
  Protocol(ICMP) echo request packets addressed to
  a destination host.
• ping www.ischool.berkeley.edu
• PING www.ischool.berkeley.edu (128.32.78.21) 56
  data bytes
• 64 bytes from 128.32.78.21 icmp_seq0 ttl61
  time0.846 ms
• 64 bytes from 128.32.78.21 icmp_seq1 ttl61
  time0.915 ms

9
TraceRoute

• traceroute www.ischool.berkeley.edu
• traceroute to www.ischool.berkeley.edu
  (128.32.78.21), 64 hops max, 52 byte packets
• 1 g2-11.inr-270-doecev.berkeley.edu
  (128.32.226.1) 0.681 ms 0.362 ms 0.495 ms
• 2 g3-3.inr-202-reccev.berkeley.edu
  (128.32.255.34) 0.437 ms 0.540 ms 0.476 ms
• 3 t5-5.inr-211-srb.berkeley.edu
  (128.32.255.127) 0.626 ms 0.648 ms 1.163 ms
• 4 www (128.32.78.21) 0.930 ms 1.220 ms 1.085
  ms
• traceroute www.google.com
• traceroute Warning www.google.com has multiple
  addresses using 74.125.224.83
• traceroute to www.l.google.com (74.125.224.83),
  64 hops max, 52 byte packets
• 1 g2-11.inr-270-doecev.berkeley.edu
  (128.32.226.1) 0.673 ms 0.431 ms 0.427 ms
• 2 g3-3.inr-201-sut.berkeley.edu (128.32.255.32)
  0.482 ms 0.505 ms 0.510 ms
• 3 xe-0-1-0.inr-001-sut.berkeley.edu
  (128.32.0.64) 0.597 ms 0.450 ms 0.355 ms
• 4 dc-svl-agg1--ucb-10ge.cenic.net
  (137.164.50.18) 10.662 ms 7.790 ms 6.443 ms
• 5 dc-svl-core1--svl-agg1-10ge.cenic.net
  (137.164.47.121) 3.623 ms 3.477 ms 3.133 ms
• 6 dc-svl-px1--svl-core1-10ge-2.cenic.net
  (137.164.46.13) 4.791 ms 3.045 ms 2.955 ms
• 7 137.164.131.61 (137.164.131.61) 3.582 ms
  3.415 ms 3.637 ms
• 8 137.164.130.94 (137.164.130.94) 8.095 ms
  58.649 ms 7.700 ms

10
Network Types
Range
Bandwidth (Mbps)
Latency (ms)
LAN
1-2 kms
10-1000
1-10
WAN
worldwide
0.010-600
100-500
MAN
2-50 kms
1-150
10
Wireless LAN
0.15-1.5 km
2-11
5-20
Wireless WAN
worldwide
0.010-2
100-500
Internet
worldwide
0.010-2
100-500
Source Coulouris, Dollimore and Kindberg

• An internet a set of interconnected networks
• The Internet the global internetwork based upon
  the Internet Protocol (IP)

11
Network Building Blocks

• Transmission media
• Copper (coax, twisted pair), optical fiber, free
  space (wireless)
• Signals
• Electrical currents, light, RF (radio-frequency),
  microwave
• Hardware devices
• End hosts, network interfaces
• Routers, switches, hubs, bridges, repeaters
• Software components
• Communication protocol stack

12
Network Architecture
13
Network Architecture

• Networking can be quite complex and requires a
  high degree of cooperation between the involved
  parties.
• Cooperation is achieved by forcing parties to
  adhere to a set of rules and conventions
  (protocol).
• The complexity of the communication task is
  reduced by using multiple protocol layers
• Each layer is implemented independently.
• Each layer is responsible for a specific subtask.
• Layers are grouped in a hierarchy.
• A structured set of protocols is called a network
  architecture, protocol architecture, or protocol
  suite.

14
TCP/IP Model
end-to-end
Appl
Appl
end-to-end
Trans port
Trans port
point-to-point
Net work
Net work
Net work
Net work
point-to-point
Link
Link
Link
Link
Host A
Host B
Router 1
Router 2
15
TCP/IP Model
server
client
(ping)
end-to-end
Appl
Appl
end-to-end
Trans port
Trans port
point-to-point
Net work
Net work
Net work
Net work
point-to-point
Link
Link
Link
Link
Host A
Host B
Router 1
Router 2
16
Message Flow
Appl
Appl
Trans port
Trans port
Net work
Net work
Net work
Net work
Link
Link
Link
Link
Host A
Host B
Router 1
Router 2
17
Encapsulation
Data
Appl
Appl
Data
Trans port
Trans port
Net work
Net work
Net work
Net work
Data
Data
Link
Link
Link
Link
Host A
Host B
Router 1
Router 2
18
Encapsulation Example Sending HTTP message
using TCP/IP over Ethernet
HTTP message
port
HTTP message
TCP header
TCP segment
TCP
IP header
IP datagram/packet
Ethernet header
IP
Ethernet frame
Adapted from Coulouris, Dollimore and Kindberg
19
ISO layer model

• Application (layer 7) specific to application
  need
• Presentation (layer 6) conversion of data
  representation
• Session (layer 5) access mgt, synchronization
• Transport (layer 4) end-to-end delivery,
  congestion and flow control
• Network (layer 3) addressing, routing
• Data Link (layer 2) framing, error detection
• Physical (layer 1) bits (0/1), voltages,
  frequencies, wires, pins,

20
Layered Protocol Architecture
21
The IP Hourglass
HTTP, FTP, SSH, SMTP,
Application Layer
Your python program, ...
Transport Layer
TCP, UDP
A single protocol
Network Layer
IP
Ethernet,
Data Link Layer
WiFi, SONET
coax, twisted pair, fiber,
Physical Layer
wireless, pigeons, ...
22
Ensuring Reliability

• Layering
• Hourglass many different applications and
  underlying network technologies, but Internet
  Protocol establishes universal addressing scheme
• Envelope/Encapsulation layer-specific
  functionalities isolation between layers
• Reliable communication over unreliable network
• IP provides best-effort packet delivery service
• TCP supports retransmission of lost packets

23
Internet vs. WWW

• Internet and Web are not synonymous
• Internet is a global communication network
  connecting millions of computers.
• World Wide Web (WWW) is one component of the
  Internet, along with e-mail, chat, etc.
• Now well talk about both.

24
How Does the WWW Work?

• Lets say Oski received email with the address
  for the i206 web page, or saw it on a flyer.
• He goes to a networked computer, and launches a
  web browser.
• He then types the address, known as a URL, into
  the address bar of the browser.
• What happens next?

(URL stands for Uniform Resource Locator)
25
How Does the WWW Work?

• Say Marti has written some web pages for her
  class on her PC.

• She copied the pages to a directory on a computer
  on her local network at the ischool. The
  computers name is herald.

• This computer is connected to the Internet and
  runs a program called Apache. This allows herald
  to act as a web server.

26
How Does the WWW Work?

• How does the computer at Oskis desk figure out
  where the i206 web pages are?
• In order for him to use the WWW, Oskis computer
  must be connected to another machine acting as a
  web server (via his ISP).
• This machine is in turn connected to other
  computers, some of which are routers.

• Routers figure out how to move information from
  one part of the network to another.
• There are many different possible routes.

27
How Does the WWW Work?

• How do Oskis server and the routers know how to
  find the right server?
• First, the url has to be translated into a number
  known as an IP address.
• Oskis server connects to a Domain Names Server
  (DNS) that knows how to do the translation.

28
Domain Name Syntax

• Domain names are read right to left, from general
  to more specific locations
• For example, www.xyz.com can be interpreted as
  follows
• com commercial site top-level domain
• xyz registered company domain name
• www host name (it is a convention to name web
  server hosts www which stands for world wide
  web)

29
Typical Domain Name
www.xyz.com
Server (host) name
Registered company domain name
Domain category (top-level domain)
Domain names are part of URLs, used in web pages.
30
Top-Level Domains

• com, biz, cc commercial or company sites
• edu educational institutions, typically
  universities
• org organizations originally meant for clubs,
  associations and nonprofit groups
• mil U.S. military
• gov U.S. civilian government
• net network sites, including ISPs
• int international organizations (rarely used)
• Many other top level domains are available

31
Converting Domain Names

• Domain names are for humans to read.
• The Internet actually uses numbers called IP
  addresses to describe network addresses.
• The Domain Name System (DNS) resolves IP
  addresses into easily recognizable names
• For example
• 12.42.192.73 www.xyz.com
• A domain name and its IP address refer to the
  same Web server.

32
Internet Addresses

• The internet is a network on which each computer
  must have a unique address.
• The Internet uses IP addresses for example,
  heralds IP address is 128.32.226.90
• Internet Protocol version 4 (IPv4) supports
  32-bit dotted quad IP address format
• Four sets of numbers, each set ranging from 0 to
  255
• UC Berkeleys LAN addresses range from
  128.32.0.0 to 128.32.255.255
• Other addresses in the iSchool LAN include
  128.32.226.49
• Using this setup, there are approximately 4
  billion possible unique IP addresses
• Router software knows how to use the IP addresses
  to find the target computer.

33
How the Internet Works

• Network Protocols
• Protocol an agreed-upon format for transmitting
  data between two devices
• Like a secret handshake
• The Internet protocol is TCP/IP
• The WWW protocol is HTTP

• Network Packets
• Typically a message is broken up into smaller
  pieces and re-assembled at the receiving end.
• These pieces of information, surrounded by
  address information are called packets

34
IP Packet Format (v4)
Field length in bits
Bit 0
Bit 31
Total Length in bytes (16)
Version (4)
Hdr Len (4)
TOS (8)
Identification (16 bits)
Flags (3)
Fragment Offset (13)
Time to Live (8)
Header Checksum (16)
Protocol (8)

• Header

Source IP Address (32)
Destination IP Address (32)
Options (if any)
Data (variable length)
Data
35
How Does the WWW Work?

• What happens now that the request for information
  from Oskis browser has been received by the web
  server herald at www.ischool.berkeley.edu?
• The web server processes the url to figure out
  which page on the server is requested.
• It then sends all the information from that page
  back to the requesting address.

36
Reading a URL

• http//courses.ischool.berkeley.edu/i206/s12/index
  .html
• http// HyperText Transfer Protocol
• courses service name (often is www)
• .ischool host name
• .berkeley primary domain name
• .edu/ top level domain
• i206/ directory name
• s12/ directory name
• index.html file name of web page

37
Web Pages and HTML

• So what do we see at http//courses.ischool.berkel
  ey.edu/i206/s12/index.html ?

38
Web Pages and HTML

• What does HTML look like?

39
HTML

• HyperText Markup Language
• Uses lttagsgt which mark up the text and tell the
  browser how to display the content.
• A backslash tag means the end of the command but
  is sometimes optional
• Examples
• This is ltbgt boldface text lt/bgt.
• ltpgt indicates a paragraph break
• lth1gt This is a large heading lt/h1gt
• lth3gt This is a smaller heading lt/h3gt

40
HTML Hyperlinks

• Hyperlink is the most important
• lta hrefhttp//www.berkeley.edu/map/maps/BC23.html
  gt 100 Genetics Plant Biology Bldg lt/agt
• The green part is called anchor text
• Its the text you see on the link
• The pink part is the url that the link will take
  you to if you click on it. The http// at the
  front indicates the http (Web) protocol.
• The lta href gt lt/agt is the command that
  indicates the enclosed information is a
  hyperlink, and the that text between the tags is
  the anchor text.
• A hyperlink can be clicked on by a person OR
  followed by a computer program.

41
HTTP

• HTTP is the protocol used by the WWW
• When a user clicks on a hyperlink in their web
  browser, this sends an HTTP command to the Web
  server named in the URL
• This command usually is to GET the contents of
  the web page and return them to the users
  browser.
• It is a very simple protocol
• It relies on the TCP/IP functionality

42
HTTP Request Example
This information is received by the web server
at www.ischool.berkeley.edu
GET i141/s07/index.html HTTP/1.1ltCRLFgt
Request line
Host courses.ischool.berkeley.edu ltCRLFgt
Request header
ltCRLFgt
Blank line
Because HTTP is built on TCP/IP, the web
server knows which IP address to send the
contents of the web page back to.
43
How Does the WWW Work?

• When Oski typed in the url for the i206 home
  page, this was turned into an HTTP request and
  routed to the web server in Berkeley.
• The web server then decomposed the url and
  figured out which web page in its directories was
  being asked for.
• The server then sends the HTML contents of the
  page back to Oskis IP address.

• Oskis browser receives these HTML contents and
  renders the page in graphical form.
• If he clicks on a hyperlink in that page, a
  similar sequence of events occurs.