Web Hosting
Michael Tsai
2015/05/25
Web Hosting Basics
•
A daemon (server) listens for connection on TCP port 80
•
Accept request for documents
•
Transmits them to the requesting user’s browser
URL
•
URI = {URL, URN}
•
URI: Unified Resource Identifier URL: Unified Resource Locator URN: Unified Resource Name (e.g., urn:isbn:0-13-020601-6)
•
URL: {protocol/app., hostname, [port, directory,
filename]}
URL Examples
Proto What it does Example
file accesses a local file file:///etc/syslog.conf ftp accesses a remote file
via FTP
ftp://ftp.admin.com/
adduser.tar.gz http accesses a remote file
via HTTP http://admin.com/index.html https accesses a remote file
via HTTP/SSL https://admin.com/order.shtml
ldap accesses LDAP
directory services
ldap://ldap.bigfoot.com:389/
cn=Herb mailto sends email to a
designated address mailto:[email protected]
How HTTP works
•
Stateless client/server protocol
•
A client
asks the server for the contents of a specific URL!
•
The server responds with the data (or err. msg.)
•
Try it: telnet to port 80
人肉Browser
•
telnet to www.csie.ntu.edu.tw port 80 (http default TCP port)
•
Type the following:
GET / HTTP/1.1
Host: www.csie.ntu.edu.tw (hit <enter> twice)
•
What do you get?
•
Try a nonexistent URL. What do you get?
Content Generation
• Dynamic content is better
1. CGI (Common Gateway Interface):
Allow external program to interact with the web server
2. FastCGI: Allow external program to continue running to server multiple requests
3. Embedded interpreters:
(e.g., Perl, PHP, Python, Ruby on Rails)
Executing external script within the server (.php, .pl)
e.g., LAMP: linux + apache + mysql + php/perl/python 4. Application servers:
Entire, full-fledge, platform for web
(e.g., Tomcat, WebSphere, WebLogic, Jetty)
Security!
•
Bottom line:
you allow the entire world to execute a script on your server (access to files, networks, and
more!)
•
Need to make sure that the script is secure
(as much as other network-accessible program)
Load Balancing
• Many factors affecting the maximum load a server can handle:
• Hardware architecture
• Operating system
• System tuning
• Sites being served
(static vs dynamic (database))
• And, in addition, network bandwidth
• Stress testing
• Implications in security
Create Scalability
1. Round robin DNS (we’ve talked about it) Note that the order in the DNS record is irrelevant.
(Think about its disadvantage)
2. Hardware solution (e.g., Big-IP from F5)
Takes response time of individual servers into account
3. Software solution (e.g., Linux Virtual Server, proxy
load balancing in Apache)
Apache
•
Web server with the largest market share (53.8% of top M busiest sites, 2014/04)
•
Runner-ups:
Microsoft & nginx take 17.8% and 12.4%
•
Began in 1995
•
First web server software to server more than 100M sites (in 2009)
•
Versatile
source: http://news.netcraft.com/archives/2014/04/02/
april-2014-web-server-survey.html
Hello World HTML
<!DOCTYPE html PUBLIC "-//IETF//DTD HTML 2.0//EN">!
<HTML>!
<HEAD>!
<TITLE>!
A Small Hello ! </TITLE>!
</HEAD>!
<BODY>!
<H1>Hi</H1>!
<P>This is very minimal "hello world" HTML document.</
P> !
</BODY>!
</HTML>!
!
You can copy the file to your user dir. for testing
In-class Homework
• Install Apache in your VM
• Instead of serving via port 80, serve via port 8080
• Turn on UserDir module, and allow the users to publish their web pages at
http://<hostname>/~<user>/web/
(note that the default is public_html, not web, so you need to change the configuration)
• <Bonus> Install JMeter to stress test your web server
Ref: https://www.digitalocean.com/community/tutorials/how-to-use- apache-jmeter-to-perform-load-testing-on-a-web-server
<<<DO NOT STRESS TEST ANY PRODUCTION SERVER!!!!>>>
Some Useful Tips
•
Apache configuration files are in /etc/apache2 (Ubuntu default installation)
•
Create a symbolic link in mods-enable/ to linked to files in mods-available/ to enable a certain
module. (And if needed, modify the config files in mods-available)
•
