|Internet Basics Review 03 - How the Internet Works|
from the May 2003 Newsletter
by Rob Zorn
I've tackled this subject a couple of times before with the following articles:
What Happens When I Click Connect? (May 2000);
How Does Your Browser Know Where to Go? (September 2001).
I'd like to think those articles did a reasonable job of taking what can be quite complicated and rendering it down into manageable pieces for the newbie or novice. You can check them out by clicking the links above. This time, however, I want to have another go, and make it as simple as possible. I'll try to use images and analogies and I'll gloss over a lot of the specifics. A lot of people, especially those completely new to the Internet, have very little understanding of what it's all about. This is a shame in a way. A basic understanding of how it all works could be quite helpful when it comes to comprehending what is happening and why, and what is possibly going wrong. In my time on the help desk, I dealt with a lot of people in states of panic and confusion that were unnecessary, but were the result of just not understanding, even in a basic way, how things work online. If this article is too basic for you, try the ones above. Names and routes have changed here at Actrix, but the meat of the articles is still good to eat.
What is the Internet?
The Internet is just a big network of computers all around the world that are all interconnected. It's also referred to as the World Wide Web. No one owns it or has total control over it, and it is constantly changing. Imagine a large hall full of people. Each one of those people could represent a computer on the Internet. From one minute to the next, that collection of people is always different. Some people are going in and out (like dialup users connecting and disconnecting), some people are always in there (like the important computers at your ISP that always have to be on and connected), some people know more than others about the hall, and some people know more of, or about, the other people in the hall than others do.
Now, imagine you're in that hall. Even if every person in the hall had to stand in one place, you could still get a message to anyone else in the hall, very simply, by passing it to the person closest to you, who could pass it to the next person until it reached the person you wanted to communicate with. That person could then send a message back to you using the same means. This is pretty much what happens when you send an e-mail or click on a link on a web page. Your computer sends a message (e.g. "I'm Bill, and I want www.google.com.") to the Actrix computer. The Actrix computer looks at the message and says, "Ah yes, I know that Mr Google is on the north end of the hall," so he passes the request to someone to the north, who looks at the request, sees it's for Mr Google, and passes it northwards again. This happens until finally it gets passed to Mr Google who is a special type of computer called a server. Mr Google then gives a copy of the requested page to the one who gave him the request, who passes it on down the line, eventually back to you.
Say you wanted to send an e-mail to someone such as email@example.com. Your e-mail gets sent to the Actrix computer that knows where aol.com is, and passes it to someone in that direction, who passes it on until it reaches a computer who looks after the post office at aol.com. That computer then looks at the first part of the e-mail address and sees that the intended recipient is fredlmnop. He puts the e-mail into an electronic mailbox called fredlmnop, and once it's in there, no one can touch it except for someone who knows the password to that box. When fredlmnop comes into the hall (comes online), one of the first things he will usually do is walk up to the post office man at aol.com and ask whether there is any mail awaiting him. This is what is occurring when you start up your e-mail program and it does a "send and receive." The post office man will only release the e-mail to fredlmnop if fredlmnop is able to communicate the correct password.
Now, what if you know the name of the person you want to communicate with in the hall, but you have no idea whereabouts she is, or what she looks like? You could just send your message off and people could just keep passing it around at random until it reached your intended recipient. This would eventually work, but it would be rather inefficient and time-consuming. It would be really helpful if there were a few individuals interspersed around the hall who specialised in knowing roughly where everybody else was. In the same way, your ISP has what are called name servers. These servers keep enormous lists (and can read them very quickly) of what direction certain people are at in the hall. What's great is that the name servers from all the ISPs around the world talk to each other frequently and co-operate by sharing information to keep everybody's lists updated.
So, if you want to send an e-mail to someone at some weird little ISP in Bolivia, or if you want to visit a web site that is located on a computer in a tiny little backstreet in Uzbekistan, and the Actrix name server doesn't know where to go to get the right information for you, it will ask another name server which way it should pass your request or message. If that computer doesn't know, it will ask another, and so on until the right answer is found. Then the Actrix name server will remember the pathway for next time, and will share that pathway with any other name servers it talks to during the course of the day.
So that's how computers on the Internet are able to communicate and pass web pages and e-mails around. Of course they do it a lot faster than passing messages around a hall full of people. If you type www.google.com into your web browser and click Go, your request for that page may have passed through ten or more computers on the way there and ten or more on the way back, yet the page will usually download for you in mere seconds.
So what else can we learn from this?
Firstly, it's not really correct to say that "you visit a web page." What really happens is that the web page is downloaded onto your computer. When you click a link (or type one into your browser) all you are sending is a request for the page to the server somewhere in the world that actually has that content on it. When that computer receives your request, it "serves" the page back to you. That means that all the stuff you've looked at online (at least lately) is stored somewhere on your computer.
Secondly, your software and settings are your own business. Your ISP doesn't tend to supply you with any programs. Most people use Outlook Express for checking e-mail, and Internet Explorer for surfing the web, but these come with your computer, not with your Internet connection. There are all sorts of other programs that do the same job. Eudora and Pegasus are both very good e-mail programs. Opera and Netscape are both very good browsers if you'd like to explore alternatives to Microsoft products. I've come across customers who assumed that all the folders and functions of their Outlook Express program were somehow stored at Actrix. Unless you're using IMAP (and if you don't know what that means you probably aren't using it), none of your settings and stuff are here. We just hold your mail in one place for you. When you download it, it's your program that sorts it into folders, displays it for you, allows you to edit and reply to it, and so forth.
This also means that e-mail settings aren't transferable from one computer to another instantly. I remember one lady, one afternoon, who was quite concerned. She'd used a friend's computer, but had used her own user name and password to connect to the Internet. When she opened Outlook Express on the computer she was quite dismayed to find that all her friend's e-mail was there, and none of hers came down when she clicked Send and Receive. This was , of course, because she was using a friend's computer, and her settings weren't in her friend's Outlook Express program. I guess this lady assumed that if she connected to the Internet with her username and password, that Actrix must somehow be able to serve up her version of Outlook Express. Imagine if that were the case! Would you really want all your e-mail stored permanently at your ISP?
Thirdly, ordinary e-mail isn't very secure between the time that it leaves your program and the time it arrives at the recipient's ISP's mail server . In fact, e-mail is pretty much like a postcard and anyone with control over any of the machines it passes though could make a copy of it for themselves. It's not too much of a worry because the servers that pass on e-mail usually have hundreds or thousands of e-mails passing through every second. It's unlikely that anyone is going to read your e-mail unless they are specifically targeting you, and most people don't need to worry about that. Nevertheless, an unencrypted e-mail is not the best way to communicate to someone else where you hide your treasures and credit card details.
Lastly (I could go on, but this will do for now), because there are so many computers involved in moving your requested web pages and your e-mails around the Internet, lots of things can go wrong that are hard to identify and correct. I don't mean that you should fear losing your e-mail. E-mail rarely gets lost because one computer won't delete it from itself until the next computer has confirmed that it has been fully received including all the destination information needed for its delivery. There can be delays, however. One or more of the computers somewhere along the line may be busy, or only working part of the time, or may not be there even though the name servers think it is. This is why on some days, web pages seem to load faster than others, or why sometimes pages don't seem to be available, when a few hours later they suddenly are again.
I hope that was clear and helpful. If it sets off further questions for you, please feel free to ask away.