How to install the Web page on a server

[fumble]

This guide explain how to install the web page on a server that is accessible to all. The server will be able to send multiple html pages to several customers, it will be able to take into account the appealing CGI procedures, it still will be able to filter and trace all the connections connected to a file. These options depend on the material used and the server software that you choose.

It is obvious that a server running on a Macintosh or Windows will be less powerful than its counterpart running on a Unix system. But it is sometimes useful to be able to run an HTTP server on a personal computer to test the web pages, CGI procedures etc.


The various software servers

Servers here are..

• Apache
• Microsoft Internet Information Server
• Netscape Enterprise Server
• Netscape FastTrack
• OmniHTTP
• Oracle Web Server
• Sioux
• WebSite from O'Reilly and Associates

[fumble]

Apache

Apache is not named after the Native American community, but actually the phonetic implementation of "a patch" which is an additional software that is initially on the NCSA server. Today the Apache HTTP server is the most used in the Internet world and this success is partly due to its robustness, and to the current popularity of free software led to Internet. In January 1998 Apache was used by more than 50% of the Web server on the Internet. Apache was originally intended only for UNIX, later version 1.3 has made it compatible with Windows NT and Windows 95. If you have not configured module through the web, the graphics installation modules have emerged until the end of 1997, where we find an impressive number of modules that adjoins to the Apache server and all free.

Apache on Windows NT runs as a service or an application, and is compatible with Microsoft ISAPI. Apache was the first servers to offer a PERL module installed in memory and allows execution of CGI PERL at high speed.

Note also that the log files are quite standard and can be used by all software for analysis and editing.
The CGI may have their own log file, which is essential since it allows to test the problems of your programs.

Apache supports SSI, the inclusion of HTML code, the HTTP/1.1 protocol, passwords on pages and, finally, the SSL 2 and 3 are supported.

On the Apache Web, an interesting topic to be visited, that of the Apache projects around which there are:

• a set of Perl modules for Apache
• SSL support
• protages of the Amiga, Macintosh, OS / 2
• Java Apache integration
• adding GUI
• a scripting language to internal HTML pages (PHP / FI) Note that this language allows the manipulation of SQL queries

Microsoft Information Server

Microsoft offers a server that is very fast and robust in Windows NT environments. IIS is used on 20% of Internet sites and is probably more used in intranet environments. IIS is currently available in version 4.0.

The strengths of Microsoft server reside in its ergonomics, since the server is managed by a graphical wizard and also via the Web. In current versions, managing logs is quite similar to other servers. (HTTP 1.1, SSL2, SSL3)

The originality of the IIS server also lies in its scripting language Active Server Page (which is recognized by the extension ASP pages). As against this language prohibits the owner carrying on a server other than the brand. Thus caution is up for users of Microsoft servers in Intranet Information Server that could be asked to put their pages on the Internet via a provider who does not have this type of server.

Note the support for SNMP agents allow the administration including the administration tools.

In addition, IIS comes with Microsoft Index Server which is a search engine to index and retrieve keyword by pages archived on the server. IIS is also closely coupled with Front Page Express SiteServer and development tools from Microsoft.

[fumble]

Interprise Server and Netscape Commerce Server

Netscape is the third of the HTTP servers used on the Internet since it was counted under Internet Information Server in terms of number of users. NFTS that can run on most operating systems on UNIX and all Windows NT architectures (but not Windows 95), fails to make the most use. Since Apache runs on Windows 95 and NT, the advantage of a multi-platform server is no longer the single argument of the firm's green lizard.

The Netscape server has its own API, called NSAPI, which are of course different from those proposed by Microsoft as ISAPI. Today peripherals of http servers products such as database connections are usually ISAPI and NSAPI but a statistical study shows that ISAPI is monitoring more than NSAPI. The latest versions of Netscape include a Java Applet API interface which allows the server to run Java programs faster, bearing in mind the Java interpreter.

Maintenance interfaces via the Web or through specific programs are as good quality as those proposed by Microsoft. So version 3 of Netscape Inrterprise server is equivalent to the version 4 of Internet Information Server.

It should be noted that JavaScript compiler on the server is to dynamically generate HTML code. Netscape joined some interesting technologies such as LDAP and CORBA. Finally Interprise Netscape Server is available in desktop products with a personal development FastTrack provided with a version of Communicator with its HTML editor.


OmniHTTP

Recent version for Windows version 2 is not without interest. Version 1 is still freeware.

[fumble]

Writing CGI procedure on the server

Writing CGI application procedures to know the programming, this activity is rather reserved for expert people. Development is in C, Pascal, Basic or any other language. There is a language which is very rich and suitable for handling files, it is the PERL.

The Perl language includes the power of C, Shell Script, awk, sed and grep. This does not say anything to the uninitiated but for connoisseurs of the UNIX system, this refers to the languages that are most powerful and most clever.


SSI Option

SSI (Server Side Includes) Procedures are used to display dynamic information to customers. For example, you can get the number of customers who visited the page.

This option is not present on all servers including freeware servers.

This can be easily ascertained by the following in your html page:

The date is:

<!--exec cmd='date' -->

If your httpd server accepts SSI you can commit to a fast and flexible programming.

[fumble]

Server configuration

Security

The safety of a HTTP server is to limit the access of the server to clients according to criteria such as:

• Limitations on the directories and files.
• Protections on the files with a password.
• ACL (Access Control List) which contain information on individuals, groups and IP addresses that have right of access to files or directories.
• Password files that contain passwords and names of users who have access to the server.
• The group files, which contain groups of users who have right to access the server.

One can change from one server to another but generally there are similarities in the methods of reporting.

Some servers do not have the option of access control. It must refer to a case by case documentation of http servers.


The reference pages on another http server

You may decide to change your HTML from http server. But this can be a problem if your page is found in many sites, it is indeed impossible to have an accurate list of referencing pages.

The solution is provided on most http servers. For example, in the NCSA server, this is done in the configuration file to Topic srm.conf redirection, making the new URL, as shown in the following example:
# ========================
# Aliasing and Redirection
# ========================
#
# Redirect allows you to tell clients about documents which used to exist in
# Your server's namespace, but do not anymore. This allows you to tell the
# Clients where to look for the relocated document.
#
# Format: Redirect fakename url
#


Aliases

Imagine any of your home page is called http://www.techarena.in//~gmaire/toc.htm and it does not seem very easy to remember for your readers. You can ask your site manager to change the address for a more user-friendly alias.

This is done in the configuration file srm.conf, under Aliasing. As shown in the following sequence:
# Aliases: Add here as many aliases as you need, up to 20. One useful
# Alias to have is one for the path to the icons used for the server -
# Generated directory indexes. The paths given below in the AddIcon
# Statements are relative.
#
# Format: Alias fakename realname
#
Alias /ime/ /~gmaire/manuel.htm

[fumble]

Error messages

Error messages are standardized http servers as they relate to issues including the best known is undoubtedly 404, which means that the syntax is correct, but when requested, that requested file is not on the server. In general, it is of course a page that no longer exists on a server.

List of error messages:

301 - The document has moved permanently
302 - The document has been moved temporarily
304 - The document has not been modified, it is possible to use the cached copy of Netscape or a proxy server
400 - The address of the document contains a syntax error
401 - You are not authorized to access the document
402 - Access to document is subject to a payment
403 - You are forbidden to access the server
404 - The requested URL is valid but is not on the server
405 - The method of application form is not allowed
406 - The request is not accepted by the server
407 - A proxy authorization is required
408 - The waiting time to access the requested page has expired
500 - An internal server error has occurred
501 - A request made to the server is not supported by it
502 - Poor Access Gateway
503 - Service Unavailable
504 - Access time to the Gateway access expired

ImageMap Option

When clickable image is used, the ImageMap Option allows to determine the actions carried out in the designated area.

Generally the option is contained in the file imagemap.conf on the http server with all configuration files. This file contains lines like:

Map:/directory/.../.../carte.map.

This will be called in the html page in the sequence:

<A Href="/imagemap/carte">
<img src=gif/arte.gif ismap>
</A>

Where there is a map of image file to click. You can have as many rows as cards click. For each card found in the map file. Corresponding map, a set of lines like:

# default program called prog1 (there is a space between default and /demo)
default /demo/prog1.htm
# The circles are defined by the center and the radius, then we call prog2:
circle /demo/prog2.htm 50.50 50.10
# The rectangles are defined by the opposite corners:
rect /demo/prog3.htm 130.10 170.90
# The polynomials are defined by a set of points
poly /demo/prog4.htm 250.10 210.90 290.90

Thus, a menu including a map will be defined with several rectangles:

default /ISMAP/cit/no_image.html
rect /ismap/prog1.html 382.262 390.271
rect /ismap/prog2.html 373.394 383.406
rect /ismap/prog3.html 472.324 483.336

To find the coordinates of the rectangles or polynomials, simply take the mouse on the map in a Web reader. You will see the address field of the URL to the information in the form /cgi/procedure?X,y and you just have to note these values.

[fumble]

The famous cookies

Introduction

Two consecutive connections from one client are independent given the HTTP protocol, ie they are seen as two totally independent connections, this is not practical, since the CGI program on the server, can not keep the particular context in more complex applications than the simple database consultation. Note that there are several means to circumvent the problem:
use hidden fields (hidden) in the forms (FORM), but this is cumbersome to manage unless you have a meta language like that used for the development. Use a file on the server storing the states, but these files can not be kept indefinitely, and if reconnection is via a bookmark, it is impossible to honor the request; so use cookies which we speak now and have been much discussed for the reason. The HTTP server sending the cookies in the HTTP header to the client. These cookies contain the accompanying descriptive URLs visited, cookies are stored on your hard disk to be operated at a future connection.

On your equipment, you can read the cookie file in the Netscape folder, program, and it is in the file cookies.txt and is read by a text editor. This is the first thing to know, because it becomes easy to analyze what are the URLs you visit (if the URLs contain cookies and your browser is enabled to receive them).

Let the cookie file be of structure by analysis of a line in the cookie file:

.techarena.in TRUE / FALSE 942195540 LANG_scarabee

We already know that the browser has an address containing interviewed. techarena.com probably www.techarena.com and then we show the information to be stored on this site by the browser.


Syntax

The Cookies are stored on the client in a cookies.txt file, these cookies can be used by a CGI, or JavaScript using one of the properties of the document.

A CGI program uses cookies by adding information in the HTTP header in the following syntax:

Set-Cookie: NAME= value [;EXPIRES= date] [;DOMAIN= domainname] [;PATH= path] [;SECURE]


The different parameters are used as follows:

• NAME = value should not contain; of, nature and white, this value is a URL that may be accompanied by arguments in the form% XX. NAME is a variable value and its value. Example valid: STATUS=CLIENT
  
  
• EXPIRES = date date of validity of the cookie. After this date the cookie will be stored on the browser cookie file. If no date is given, the cookie expires at the end of the session. The date format is jds, dd-mm-yy HH: MM: SS IST. JDS is the day of the week, dd is the day of the month, mm is the month of the year, yy is the year on two characters.
  
  
• DOMAIN= domainname is the domain of a cookie. No value is the name of the server that is used.
  
  
• Path = PATH set path for a cookie. Without value, the browser uses the path of the document that created the cookie.
  
  
• SECURE defines whether the cookie is transmitted securely to the server.

[fumble]

Remarks

When the cookie is sent by the server, the browser consults its list of cookies and compare the list of domains and the domain of the URL accessed.

The correspondence is made solely on the end of the URL, ie including the domain extension (.Com, .In, .Edu) and the previous name. Thus www.google.com will be saved with the identifier DOMAIN google.com, but the cookie www.google.tech.in site will be registered with the identifier DOMAIN tech.in while www.mail.google.com will registered with the identifier DOMAIN google.com

The PATH is variable as it specifies the correct path for the cookie.

Thus, when viewing a URL corresponding to the domain name and path, the browser will send the list of variables stored on your hard drive.

When cookies are deleted from the server, it send a variable with an expiry date already expired.

Browsers can save 300 cookies in a day, knowing that they can hold 4 KB of information. There is a limit of 20 cookies per domain. If these limits are reached, the browser cookies destroys with latest expiry date.


Example

A browser sends a request to an HTTP server and receives from the HTTP header of the response document:

Set-Cookie: VIsitor=fumble; path=/; expires=Wednesday, 09-Nov-99 23:12:40 GMT

This cookie is saved on the file cookie.txt and when the customer makes a request on the same server with the path / server sends the following

message: Cookie: VISITOR=fumble

Then he may receive another cookie from this server,

for example:
Set-Cookie: CATEGORIE=Prospect; path=/

At this point, information on the cookie is URL

Cookie: VISITOR=fumble; CATEGORY=Prospect

Then in the same URL, though always with the same browser, it decides to buy a product after the cookie is sent:

Set-Cookie: CATEGORY=Client; path=/