Internet - The World Wide Web
Web Servers

Web Servers

So the Web is really just a giant network of computers all hooked together. It works like any other computer network, that is, by client computers accessing information from server, or host, computers. All these computers are connected together by a variety of different methods, both wired and wireless, but primarily over good old copper telephone wires.

There is no such thing as cyberspace. Web sites do not just float around and wait for someone to access them from thin air. All Web sites reside on a Web server, somewhere. A Web server can be any type of computer that is running server software and connected to the Internet. Technically, a plain old PC that has a full-time connection to the Internet is a Web server, but most Web servers are specialized computers that run on either Windows NT or the Unix operating system.

Once you have made your decision and your Web server is set up and configured properly, it waits to be contacted by the right client software (Web browser software like Netscape Communicator or Microsoft’s Internet Explorer are common examples of client software). The Web browser and the Web server speak to each other through a well-defined language called a protocol. The World Wide Web is that protocol.

These two software applications connect with each other by sending messages back and forth. Each message or request that the browser sends to the server contains a header, with information about who it is and where it is coming from. More importantly, it also contains a set of instructions requesting a specific response from the server, which could include HTML pages, graphic files, and any other information or files that are used by the browser to build a Web page that the viewer sees.

A key element of the request sent by the browser to the server is the browser’s Internet Protocol Address (IP). The IP address system is critical to the operation of the Internet. It is the only way that browsers and servers know how to reach each other and know where to send files and information.

Every computer that is connected to the Internet, whether it is a browser or a server, has an IP address. Computers that are connected to the Internet all the time, like Web servers, usually have what’s known as a static IP address. This IP address never changes. People with dial-up connections to the Internet usually have dynamic IP addresses—their IP addresses are randomly assigned by their Internet service provider from a large block of IP addresses every time they connect to the Internet.

Every machine that is connected to the Internet has a unique IP number. No two computers can ever have the same IP address at the same time. The IP address is different from the domain name. Several domain names can refer to the same IP address. Domain names are just a lot easier to remember than an IP address, which consists of four parts, separated by dots. For example, 143.938.3.32 is an IP address, but Amazon.com is a domain name. (We made up that IP address by the way.)

Switched any Packets Lately?
Actually, you have. If you’ve checked your e-mail or browsed the Web, you’ve been sending and receiving—switching—packets. And what exactly is a packet? A packet is a piece of information sent over the Internet. Think of it this way. Take a document like that new business plan, chop it up into little pieces, and write each on a postcard. Then mail each card. Each card has on it the same information contained in an online packet: the payload, or the information, and the to and from, otherwise known as the source and destination addresses.

Packet switching begins once the cards are delivered to the post office and begin their journey to the destination. Each one of those postcards may take a different route to get to the final address. All kinds of things can dictate how the individual postcards, or packets, get routed to the destination. When they finally arrive, they may not do so in order, but someone will sit down and reassemble them.

If you were to send that same business plan online, your computer would break it into the little pieces and send it out over the Net. Each packet may take a different route, but on arrival they are all reassembled courtesy of the TCP protocol. Rather than sitting down and putting them in order, the computer does it automatically.

The route a packet takes depends on how congested it is. Packets get sent along the least congested route at the time the packet leaves the first computer for the second. Because of this, each and every packet can go on a different route.

Web Application Servers
A Web server does the primary job of responding to Web browsers and sending Web pages across the Internet to the user. What about all of the databases, audio, video, and e-commerce that are also part of the Internet? How does all of the other cool content get down to my Web browser? More importantly, how do all of those Web sites like My Yahoo! maintain my personalized information for financial portfolios, bookmarks, and any number of the services that fit me alone? Are these handled by the Web server also? Well, sometimes they are.

Technically, the term Web server describes the Web server software, although the entire computer box where the Web site resides is often referred to as a whole as the Web server. The Web server can only deliver HTML files and graphic files. Most other information that is delivered from the Web site is sent by a special server that is dedicated to a particular task—a Web application server. Some examples of these kind of servers are database servers and video and audio servers. These application servers are important because they help balance out the load from browser requests among several different computers. On high-traffic Web sites, hundreds of people making hundreds of separate requests would overload many regular servers, causing big delays and even crashes. By having application servers that handle discrete parts of a Web site, traffic is more easily balanced, creating less jams and keeping the site moving.

Cross-server Communication
So the Web server has to talk to the application server. How does it do that? The original method for cross-server communication used something called the Common Gateway Interface (CGI).

CGI applications are referenced within HTML pages and then run on the server to produce the desired result. They are usually run when information from the user is being delivered to a database on the server. Also, CGIs often instruct the Web or application server to deliver a customized HTML Web page back down to the user.

A CGI program can be written in any programming language, but most are written in Perl for the Web. Other popular programming languages for CGI applications are C, C++, and VB script. Perl CGIs run very fast and are easy to develop (if you know how to program in Perl!), but they are not great on heavy-traffic Web sites. They tend to bog down the site.

Microsoft Active Server Pages (ASP) work in conjunction with Microsoft Internet Information Server to deliver a plethora of interactive functionality to a Web site. ASP files are like CGI scripts in that they are little programs that run on a browser. An ASP file can contain HTML, scripts written in any language, and ASP objects or commands. When the ASP file is requested by the browser, the server reads through the code and executes any of the scripts or ASP commands it contains. ASP commands (called objects) instruct the server to do one of the following:

Request Object—Get information from the user
Response Object—Send information to the user
Server Object—Control various aspects of the server
Session Object—Store and edit information about the user’s current session
Application Object—Control other applications running on the server

Scalability
Scalability is the key issue in distributing server load among several different computers. While you are interested in how many potential customers visit your Web site in a week or month, that is of little issue to the people who administer your server. For server administrators, the people who build and maintain Web servers, the key issue is gauging how many simultaneous users there will be at any one time—how many people will you have using your Web site at once. On big, high-traffic sites such as Amazon.com you can guarantee hundreds of thousands of simultaneous users. How many simultaneous users your Web site has is referred to as the peak load or as peak concurrent users. Scalability is important because when a Web server or an application server is overloaded, it can slow down to a crawl or even crash completely.

For this reason, many Web server administrators prefer using computers running the Unix operating system. Unix is much more robust and reliable than Windows NT, although Microsoft and many Windows NT evangelists will dispute that claim. However, Unix systems are much more difficult to set up, configure, and maintain than Windows NT. For that reason, most small to medium-size business Web sites will have an easier time running their Web site on the Windows NT platform.

* Source - Everything Online Business Book
              Get Your Business On the Web!
              Use the Internet to build your business