Revealing the behindthescenes magic that occurs when you surf the Web

1
How the Web Is Spun

• Revealing the behind-the-scenes magic that occurs
  when you surf the Web
• Theron S. Welch

2
Presentation Goals

• Demystify the internals of the Web
• Clearly explain the basics of HTTP and TCP/IP and
  their differences
• Explain why this foundation is important

3
What is the World Wide Web?

• Everybody knows how to use a web browser!
• Butrealize that the Web is simply a convenient
  way to view and share documents using a
  technology known as HTTP.

4
More and more acronyms

• Essential Technologies HTTP, HTML, TCP/IP
• Newer technologies ASP, JSP, XML, SOAP, and
  much more

Source Phlman International Dec. 1, 1997
5
HTTP Hypertext Transfer Protocol

• Question HTTPs purpose?
• Answer Transferring hypertext documents

HTTP can transfer other types of files too
Source Landmark U.S. Market Survey OGJ, October
1997
6
Hypertext Files

• They are also known as HTML files Hypertext
  Markup Language files
• They are simple ASCII files. ASCII is about as
  platform independent as you can get!
• Markup means that the text is marked up with
  special tags
• HTML tags describe the formatting of a document

7
Sample HTML File
8
Sample HTML file in a browser
9
Common HTML Tags

• ltAgt - Anchor tag. Use this tag to hyperlink to
  another resource on the web. Example ltA
  HREFwww.geographix.comgtGGX Websitelt/Agt
• ltIMGgt - Image tag. Use the image tag to insert
  images (gifs or jpgs) into HTML documents.
  ExampleltIMG SRCmygraphic.jpggt
• Many other tags exist. You can also add scripts
  to web pages.

10
HTTP

• HTTP has various methods (functions) that clients
  can use to request information from servers
• A web browser uses HTTP to GET a web page from
  a web server.
• A web server is a machine running software that
  is listening for HTTP requests from clients
• HTTP is a text based protocol it sends ASCII
  requests and responses over the Internet

11
HTTP Method GET

• GET is by far the most commonly used method in
  HTTP
• GET is used to get resources from a server

12
A closer look at GET
Request line

• GET /index.htm HTTP/1.0
• From twelch_at_geographix.com
• ltblank linegt

Headers
Marks the end of a request
13
Some Basic Rules

• Requests are in upper case (as in GET)
• Headers are optional
• A blank line marks the end of a request

14
Typical Response from a server
Response Line

• HTTP/1.0 200 OK
• Server Microsoft-IIS/3.0
• Date Thu, 19 Apr 2001 152246 GMT
• Content-Type text/html
• Last-Modified Thu, 19 Apr 2001 000709 GMT
• Content-Length 7118
• ltblank linegt
• ltContentgt

Headers
End of servers response headers
Document that was requested follows
15
Connections (TCP)

• The client makes the connection to the server
• The client uses the connection to send a request
• The server sends the requested data on the
  connection
• The server closes the connection

16
Connection management

• HTTP 1.0 required one connection per web resource
  on the web
• One connection per resource is very inefficient

17
HTTP 1.1

• Can maintain connections to improve performance
  and reduce traffic
• Is an option selected by default in Microsoft
  Internet Explorer 5.0

18
Other Important HTTP Commands

• HEAD similar to GET but no content is sent
• POST used to send data to a server. Message
  boards, shopping cart, etc.

19
TCP/IP

• Where HTTP is simple, TCP/IP is very complex
• HTTP uses TCP/IP to connect and talk to a server
  TCP/IP functions on a lower layer
• TCP and IP are independent of each other but and
  typically work together
• The phrase TCP/IP refers to a large suite of
  protocols, the most common being TCP and IP

20
TCP Transmission Control Protocol

• Provides reliable transport service to HTTP (and
  other higher level protocols)
• TCP assigns a port for each connection

21
Ports (TCP)

• Q What the heck is a port? Is it a physical
  entry on a network card?
• A No. A port is just a 16-bit number that
  functions as a key
• 16-bit means 2 to the 16th power, or 65,536
  available numbers
• Used to identify which software applications a
  packet of data should go to

22
Well-known Ports
Many TCP/IP applications are so common they have
their own well-known port numbers. For example

• FTP 21
• HTTP 80
• Telnet 23
• SMTP 25
• Microsoft Active Directory 3268
• Sybase SQL Anywhere 1498

23
Port 80 HTTPs port

• Web Servers listen on port 80 for incoming
  connections. In other words, when a request to
  connect comes in on port 80, TCP ensures that it
  goes to the Web Server that is listening on
  that port.
• Web Servers dont have to listen on 80, but its
  the established convention. Hence, its a
  well-known port number.

24
Hackers

• Hackers typically break into networks by figuring
  out what ports are open.
• A receptive port that exposes a weakness can
  allow access.
• FYI Watch out for File and Printer sharing!

25
IP Internet Protocol

• Deals with identifying computers on the Internet
  (IP Addresses)
• Moves TCP data around the Internet

26
IP Addresses

• Example www.microsoft.com aka 207.46.131.199 is
  an IP Address
• This notation is called dotted-decimal notation
  and separates 4 bytes.

27
Typical Web Browser

• Internet Explorer

28
Not-so-typical Web Browser
WebSnatcher rendering the same site
29
The Ultimate Basic Browser Telnet
Instead of clicking on hyperlinks in a web
browser, why not just type in the HTTP commands
yourself in Telnet? -)
Type in a web address and a port number
30
Telnet
The screen will go blank. After typing GET /
HTTP/1.0 ltENTERgt ltENTERgtYoull see something
like this
31
Telnet
Heres what an error looks like as returned from
a web server.
32
What a typical browser has to do

• Formatting pages
• Rendering graphics, animation
• Supporting audio/video plug-ins
• Supporting Java, scripting languages
• And much more

33
What a typical server has to do

• Connection management
• Stress handling

34
New Technologies and Directions

• ASP - Active Server Pages
• JSP Java Server Pages
• XML Extensible Markup Language
• Web Services with SOAP

35
Constants

• New technologies come and go
• Only a few new technologies will make a
  significant impact
• HTTP has had explosive success
• TCP/IP has stood the test of time about 30
  years or so. Wow!
• HTTP and TCP/IP will remain with us

36
Excellent Resources on the Web

• www.w3c.org - The World Wide Web Consortium
• www.faqs.org - The Internet FAQ Consortium
• Dont forget the RFCs!

37
Bibliography

• TCP/IP Illustrated Volume 1, W. Richard Stevens,
  1994 Addison Wesley
• TCP/IP Illustrated Volume 3, W. Richard Stevens,
  1996 Addison Wesley
• Network Programming for Microsoft Windows,
  Anthony Jones and Jim Ohlund, 1999 Microsoft Press