The web server is a relatively simple application that is designed to receive requests for files from web browsers and respond by sending the files to them. The server and browsers communicate using an application layer protocol called HTTP, the Hypertext Transfer Protocol. Hypertext is a term invented in the 1960s to describe a database that displays information and uses live links to jump to other areas of the database. In the context of the web, hypertext refers to the hyperlinks found on every web page that enable you to jump to other pages or other parts of the same page.

HTTP basically consists of request messages generated by the browser client and replies generated by the server. When you type a URL into your browser, it generates an HTTP request message containing that URL and transmits the message within a standard TCP/IP packet to the server.

The URL identifies both the server to which the message should be sent and the specific file being requested on that server. For example, in the URL http://www.mycorp.com/research/index.html, the >http:// specifies that the message will contain an HTTP request. Browsers also support other protocols like FTP, a URL for which would begin with ftp://. The www.mycorp.com part of the URL is a DNS host name that identifies the server hosting the web site. The server resides in a domain called mycorp.com on a computer with the host name >www. The /research portion of the URL identifies the directory on the web server where the requested file is located and index.html is the name of the file itself.

The only difference in this format on an intranet is that you can use the web server's NetBIOS name (that is, the computer name you supplied during the Windows 98 installation) instead of the domain name. Thus, if the NetBIOS name of the system running the server software is cz1, the URL would be http://cz1/research/index.html. Intranet servers can use domain names, but this is usually done only when a DNS server is running on the local network.

All web servers have a default file name that is assumed when no specific file name is included in the URL, usually either default.htm or index.html, depending on the web server being used. When you type a URL like http://www.mycorp.com, for example, that contains no file name, the default file on the root directory is assumed. It is the responsibility of the site's developer to use the server's default file name for the site's home page, so that users can access it easily.

The server, on receiving an HTTP request from a client, processes the URL, locates the appropriate file, packages it in an HTTP reply message, and transmits it back to the sender. The server is completely ignorant of the file’s contents and will remain so throughout the transaction. It is the responsibility of the client browser to process the file and display it properly.

In addition to simply transferring the files that make up web pages, a web server can also process application requests that either transfer an application file to the client or execute an application process on the server. Many web sites use scripts or other technologies to provide clients with more complex functions. You can use tools like Microsoft FrontPage to simplify the process of deploying more complex web page elements.

The tasks performed by web servers are quite simple; in most cases, it is the sheer volume of traffic that limits their capabilities. If you are running PWS on a Windows 98 system and you decide to connect it to the Internet using a 56 KB dedicated dial-up connection, you can support no more than two or three concurrent users comfortably, no matter how powerful a computer you use. With a cable modem connection, you can service from 10 to 20 users adequately. An intranet server running over a 10 Mbps Ethernet LAN, however, has more bandwidth available to it than all but the largest Internet sites. As a result, you will find that the web pages that you create for your intranet server will load much faster on your clients' browsers than standard Internet web pages do. Since the server transmits the pages faster, it can satisfy more requests. This means that an intranet server is capable of supporting more users than the same system could over a slower intranet connection.

The browser program that generates HTTP requests and processes the replies that it receives from web servers is more complex than the server in several ways. The lingua franca of the World Wide Web is HTML (Hypertext Markup Language). An HTML file is simply an ASCII file with an .html or .htm extension containing the text that is to be displayed, plus a series of tags that specify how the text is to be displayed and that identify the other files and elements that are part of the page. In most cases, the reply packet generated by a server in response to a browser's URL request contains HTML code.

Web servers transmit files to clients without being conscious of their contents in any way. Browsers, however, have to read the files in order to display them on the screen. For example, when a browser receives an HTML file from a server in response to an HTTP request, one of the first things it does is read the contents of the file and search for image tags, which appear like this: <IMG>. HTML files use image tags to specify the names of the graphics that are to be embedded as part of the web page. The image files are not part of the HTML file; they are separate .gif or .jpg files that the server must transmit to the browser separately. On finding an <IMG> tag, the browser generates a new request to the server, requesting the image file specified immediately following the tag. Thus, displaying a web page that contains a large number of graphics may require dozens of separate HTTP requests and replies traveling between the client and the server. In addition to images, HTML files can contain references to sound and other multimedia files, scripts, and applications, as well as code (like Javascript) to executed by the browser.

Aside from requesting all of the required files from the server, the browser also uses the HTML tags to determine how it should display the data. The tags can specify the size of the text, the font it should use, its color, and its location on the screen, along with many other elements that affect the final appearance of the web page. As the browser takes these tags into account, it uses the configuration of the computer's video display to lay out the page so that it fits the browser window. If you alter the size of the window, the layout of the web page dynamically adjusts to fit the available area.

Creating HTML files for your web site is not difficult; the tags used are relatively few in number and easy to learn. However, there are also a great many products available (such as FrontPage Express, included with Windows 98) that enable you to build web pages without working directly with the HTML code. If you use Microsoft Word, for example, you can save your documents in HTML format just as easily as in Word's own DOC format, and then simply copy them to your site.

In addition, when you're creating web pages for use on an intranet, you know more about the client environment than you do on the Internet. You would know, for example, that all of your network workstations have Word installed, so you can create a hyperlink to a Word document and be certain that all of the clients can load it. On the Internet, you have no way of knowing what applications are installed on the client systems, so you can only use file formats that you know are supported by all clients.

The Microsoft Personal Web Server (PWS) included with Windows 98 consists of the following components:

Microsoft Personal Web Server

Microsoft Transaction Server

Microsoft Data Access Components

Microsoft Front Page Server Extensions

Message Queue Server 1.0

Elements like the Transaction Server, Message Queue Server, and the Data Access Components are intended for relatively sophisticated web development projects. One of the intended functions of PWS is to be able to test web pages and applications developed for Internet and large intranet sites, so support for the more complicated development tools used to build these sites is provided. In addition to the basic HTTP server itself, the only additional component that is helpful in the development of a small intranet web site is the Front Page Server Extensions. These extensions are installed as part of the Typical PWS installation, while the other three additional modules are not. You'll learn more about using FrontPage Express to build web sites later in this chapter.

While the Personal Web Server is included in the Windows 98 release, it is not installed by default. In fact, even though the program is included in the Add/Remove Programs Control Panel, installing PWS through this interface does nothing but present you with instructions for launching the actual installation program, which you must access from the \add-ons\pws directory of the Windows 98 CD-ROM. When you run the Setup.exe program from this directory, you are prompted to select the Minimum, Typical, or Custom installation option. The Typical option requires 36.1 megabytes of disk space, and provides support for all of the features that the average beginning web site developer will need.

PWS is a scaled-down version of Microsoft’s main web server product, the Internet Information Server (IIS) package that runs on Windows NT. You can see the relationship between the two products on the installation screens displayed during the Custom installation process, which show grayed-out options for a FTP server that does not exist in the PWS release (see Figure 11-1). PWS has lower system requirements than IIS and is designed to support a relatively small number of users.

The Custom installation process enables you to select the PWS components that you want to install, and their locations. The directory that the server uses as the home directory for the WWW service will be the default location for your web site’s HTML and other files. When someone accesses the site without supplying a specific file name, it is the default HTML file in this directory (which is default.htm, by default) that the server will transmit. By default, the installation program creates a directory called C:\Inetpub\wwwroot and makes it the home directory. In most cases, beginning users will not need to run a Custom installation, A Typical installation will suffice, unless you want to install additional web server components. You can always run the setup program at a later time to add components.

The Personal Web Server runs as a service on the Windows 98, meaning that is runs in the background continuously. By default, the installation program creates an icon in the system tray that you can use to start, stop, and pause the service as needed.

As with any background process, PWS occupies some of the system resources on the computer where it is running. If a large number of clients are continually accessing the web pages hosted on the system, performance of other applications may suffer. However, on a small business or home network, there should not be a serious performance hit unless the computer is underpowered to begin with. If you plan to host your web site on a computer that also serves as a user's workstation, you should select a machine that has sufficient horsepower to do justice to both tasks. You should also make sure that the web server PC is always available. Don't run your web server on the computer that you also use for testing new hardware and software or that the user has to frequently reboot.

Once the PWS installation is complete, you have to reboot the system and the service will start. The web server includes a default home page, so you can connect immediately with a browser to test it. On an intranet, you are not likely to have a DNS server that will enable you to connect to the web server using a standard DNS name, like www.mycorp.com. However, you can use the computer's NetBIOS name instead, because Windows 98 will resolve the NetBIOS name into an IP address just as it does DNS names. Thus, if the web server PC's NetBIOS name is cz2, you can simply type cz2 into the browser's URL field or, to use the complete form: http://cz2. Later, as you build your site, you can construct more complex URLs using cz2 to identify the server, just as you would with a DNS name, as in the following:

http://cz2/documents/minutes1299.html

As with an Internet web server, you can also use URLs based on the server's IP address, as follows:

http://192.168.2.2/documents/minutes1299.html

It is also possible to use a standard DNS name for the server, even if you don't have a DNS server on your network. Every TCP/IP system supports the use of a HOSTS file that contains DNS names and their equivalent IP addresses. When you specify a DNS name, such as www.mycorp.com, in a TCP/IP client application, the system checks the HOSTS file for an entry containing the name before it generates a DNS request. If an entry is present, the system uses the IP address in the file and no DNS request is needed.

Windows 98 includes a sample HOSTS file, called Hosts.sam, in the C:\Windows directory. This file contains instructions for creating the file, which is simply a list of IP addresses and their equivalent DNS names, as follows:

Each address and name must be separated by a least one space and be on a separate line. Once you have added your entries, you must rename or copy the file to HOSTS, still in the C:\Windows directory. The only problem with using a HOSTS file for name resolution is that each computer on the network must have its own copy of the revised file. You can either edit the file on each machine or copy the revised file from one machine to all of the others.

The PWS includes an application called the Personal Web Manager that you can use to administer the server and publish documents. You launch the manager from the Start Menu or by clicking the secondary mouse button on the PWS tray icon and selecting Properties from the context menu. You then see the Main page shown in Figure 11-2. The Main page specifies the URL (containing the system's NetBIOS name) that your clients can use to access the server, as well as the home directory for your site. Clicking the Stop button disables the web server, preventing people from accessing it until you restart it. The Monitoring box displays statistics for the site, including the number of current connections and the connection history.

Figure 11-2. The Personal Web Manager's main screen

The default home page that is installed with PWS is just a placeholder that you can use to test that the server is running properly. In order to create your own home page on the site, you can use the Home Page Wizard incorporated into the Personal Web Manager. Using the answers to a series of questions and text that you supply, the wizard creates a rudimentary home page.

To launch the Home Page Wizard, you open the Publish page in the Personal Web Manager and click the Home Page Wizard button. The wizard prompts you to select a template for the page's appearance and specify whether you want to include features like a guest book and a drop box (that enable users to leave public and private comments). Then, you specify the text that you want to appear on the page and the wizard produces a default home page for your server, as shown in Figure 11-3.

The Home Page Wizard does not create an HTML file for your homepage. Instead, it uses active server pages (ASP). Files with an .asp extension are active server pages, which are HTML web pages that are dynamically-created by the web server. Using a development environment like Visual Basic, the site developer programs the server to use the information you have provided as part of the newly-created HTML code for the page. You can modify the page later by opening the Web Site page in the Personal Web Manager. You can also use any of the techniques described in Creating HTML Documents, later in this chapter, to create HTML-based pages of your own to use instead of the prefabricated one produced by the wizard.

On the Personal Web Manager's Advanced page (see Figure 11-4), you can see the server's settings for the default document name. Notice that the Default Document(s) field lists two file names: Default.htm and Default.asp. These are the standard default page names used by Microsoft web servers. By listing two files in this field, the server is configured to use the Default.asp file when a Default.htm file does not exist in the server's home directory.

You can modify the default document names, to suit your tastes or in order to publish HTML files that already exist using another name. Microsoft web servers use Default.htm, but most other web server products use the standard UNIX home page file name: index.html. Be aware, however, that the file names specified here are active throughout your web site. If a user connects to your site by specifying a URL that contains a directory name but no file name, then the server will use the default document in that directory. If you clear the Enable Default Document checkbox, the server will not deliver any document file to a client unless a file name is specified in the URL.

The Allow Directory Browsing checkbox on the Advanced page specifies whether or not the server should display a list of the files in a directory, as shown in Figure 11-5, when there is no default document to load. This files in the list are hyperlinked, enabling users to open any file in the directory. If you clear the Enable Default Document checkbox and enable Allow Directory Browsing, the server will always display the file list for the directory specified by the client's URL (including the home directory) and will not load any .html or .asp file unless the user selects one from the list. If you clear both checkboxes, the server will not display anything but an error message unless the URL contains a specific file name.

The Save Web Site Activity Log checkbox specifies whether or not the server should maintain a log of all the clients that have visited the site and the files that they have accessed. This information is less useful on an intranet server than an Internet one, but the only harm in collecting it is the disk space occupied by the log file. However, a web server on a small intranet is not likely to generate a huge amount of traffic.

The Virtual Directories box on the Personal Web Manager's Advanced page is where you control the directory structure of your web site. When the PWS installation program creates the server's home directory, it creates the root directory of the web site. Any subdirectories you create under the home directory become valid subdirectories on the site as well.

When building your web site, you can create subdirectories beneath the home directory and store your documents there, but it is also possible to publish documents located outside of the home directory structure using virtual directories. A virtual directory is a directory located elsewhere on a local drive that appears to web users as a subdirectory of the site's home directory. To create a virtual directory, you select the directory that you want to publish on your site and assign it an alias, that is, the name under which it will appear in the site's directory structure.

For example, if your site's home directory is C:\Inetpub\wwwroot (the default), you can create a subdirectory under \wwwroot called \sales, and it will be accessible to users by specifying the subdirectory name as part of the URL, as in http://cz2/sales. You can also create a virtual directory by selecting a directory on the D: drive called D:\reference\marketing\docs and assigning it the alias marketing. This directory, located on a different hard drive, will then be accessible as a subdirectory called \marketing on the web site, using the URL http://cz2/marketing.

There are several advantages to using virtual directories to build your site. First, the actual locations of the files remain hidden from the site's users, which server as an additional security mechanism. Second, you can leave your documents in the original location where people normally access them. For example, you can have your company's bookkeeper continue to maintain the firm's financial spreadsheets in the usual way and publish them for the rest of the company in a virtual directory at the same time. Since web site access is read-only, there is no danger of the files being modified or damaged, nor can web users determine where the files are actually located.

To create a virtual directory, you click the Add button in the Virtual Directories box, to display the Add Directory dialog box shown in Figure 11-6. Here you can browse to the directory that you want to publish, specify an alias for the directory, and indicate what permissions you want to grant to web users accessing the directory. The available permissions are as follows:

By default, newly created virtual directories are granted the Read and Scripts permissions, which enable client browsers to process standard HTML pages and image files, as well as active server pages and other types of scripts. Once you create the virtual directory, it appears in the Virtual Directories list on the manager’s Advanced page. However, while the directory is accessible to web users via hyperlinks on your other pages or through a URL, it does not appear in the directory browse list.

In addition to creating virtual directories, you can use this same interface to modify the location of the server's home directory. Rather than move documents into the home directory, you can move the home directory to where the documents are by highlighting the <Home> object at the root of the directory tree in the Virtual Directories box and selecting Edit Properties. As with virtual directories, the home directory can be located on any drive installed in the local machine.

As a result, you can conceivably use your intranet web site to publish documents without employing any HTML code at all. By enabling the Allow Directory Browsing feature and deleting the default home page (Default.asp), you can provide your users with a browsable directory structure that provides access to files of any type.