|
Addendum: (continued)
How the Internet Works
"Cyberspace. A consensual hallucination
experienced daily
by billions of legitimate operators, in every nation, by children
being taught mathematical concepts. . . . A graphical representation
of data abstracted from the banks of every computer in the
human system.
Unthinkable complexity."
William Gibson
Neuromancer (1984)
We will begin our trip through
cyberspace with a high-level (i.e., not deeply technical)
view of the Internet, the wonderful communications medium
that is catapulting so many imaginations into a new realm
of pure thought. By the way, you don't necessarily need a
computer to journey into cyberspace because, as is explained
in this book's opening chapter, your mind is where cyberspace
actually exists.
As we travel from place to
place on this mental journey, I recommend that you picture
connections between the places we visit by stringing thin,
imaginary fibers of light between them. Just like the old
woodsmen who made crude diagrams of the trails they cut through
a forest, it is helpful to build a mental map so we don't
get lost. Imagining these fiber trails will make our trip
more manageable.
E-mail
Let's begin our journey through
cyberspace by traveling from your home to our first stop,
your cyber-neighborhood's electronic post office. Using your
mind's eye, connect one end of a clear plastic fiber, or fishing
line, to your house and then reel it out behind you as we
travel through cyberspace to what is technically referred
to as your e-mail
server. (In case you are wondering, in the physical
world your home computer is connected to your e-mail server
through telephone lines that connect you with your Internet
Service Provider, or ISP. Your e-mail server is often
located in the ISP office.) You can think of this e-mail server,
which is simply a special purpose computer, as your neighborhood
post office in cyberspace. Picture your e-mail server as an
actual post office building and attach the other end of your
plastic fiber to its side. Now imagine that fiber as being
filled with a pulsating light, which represents the "ones"
and "zeros" of computerspeak. The ever changing
patterns of this pulsating light are what carry the actual
information flowing into and out of your personal mailbox.
Simple, isn't it? It might
surprise you to learn how close this image is to the physical
reality of the Internet.
You may also be surprised
to see how alike cyberspace and our world of consensual reality
actually are. Things just happen faster in cyberspace, almost
as fast as the speed of light in many cases.
Now bring to mind all of the
people with whom you would like to be in contact-people like
your friends, co-workers, former neighbors, relatives, celebrities,
and so on. Then picture a thin, lighted fiber running from
your e-mail post office to each of their neighborhood e-mail
post offices. In your mind, connect a lighted fiber from each
of these post offices to the homes and places of business
where the people you want to contact will be reading their
electronic mail. Done? Congratulations, you have just begun
to construct your own working image of cyberspace.
What cyberspace looks like
in your mind's eye isn't important. What is important are
the lighted fibers connecting these imaginary edifices, for
it is these high speed information "pipes" that
really set cyberspace apart from physical spacelight speed
interconnections filled with pulsating information. After
all, information and connections are what the Internet is
really about.
Perhaps the most important
use of the Internet today is also its most commonly used application.
I am talking about electronic mail, e-mail, which wins the
prize as the killer
application of the formative years of the Internet.
Even after the advent of the World Wide Web and the introduction
of multimedia,
e-mail created the most traffic
on the Internet. As audio and video technologies become more
widely used on the Net, however, the raw number of bits
being transmitted for e-mail purposes will lose their dominance,
but we still may have a few more years before e-mail loses
its title as king of the Internet applications.
The
first few e-mail messages one sends are usually to friends
and family, but before long whole new vistas of like-minded
spirits come into sight. Then the real power of e-mail can
be appreciated. Two people who have never met in person can,
nonetheless, become intimate friends through electronic mail.
In fact, it may be the lack of face-to-face contact that is
responsible for some of these new relationships blossoming
into intimate friendships. Without having to process all of
the data that is involved in judging someone's personal appearance
and voice, we make a direct connection with the thoughts in
that other person's mind. Think about this for a minute; with
e-mail you can establish a direct channel of communication
with another person's mind. So many possibilities begin to
unfold.
There are some pitfalls with
electronic mail. For example, we have all read or seen reports
in the popular media about the romances of men or women who
get fooled by creative teenagers pretending to be someone
older. Cases like those remind me of the famous cartoon from
The New Yorker Magazine showing two dogs sitting in front
of a computer where one dog says, "No one knows you're
a dog on the Internet." Pause for a moment, however,
and think about what is actually taking place in these cases
of fictional e-mail identities. Forget for a moment about
the age and gender of the individuals involved and think about
the fact that what is happening is a sixty year old man or
woman is engaged in a deep and meaningful conversation with
a highly imaginative 15 year old boy or girl. There is a lot
of intellectual activity taking place here, for e-mail exchanges
like these create a temporary shared reality in cyberspace.
In the case of online flirting, is that boy or girl lying
or just being creative? As long as no one is physically, financially,
or emotionally harmed, has any serious damage been done? While
I am not promoting this type of behavior, it is nonetheless
obvious that minds have been stretched, never again to return
to their original dimensions. Emotions have been triggered.
Human to human communications have taken place that otherwise
never would have transpired. This can't be all bad. I believe
it is important to keep in mind that there are many sides
to these situations, and that not all unexpected results are
negative.
The next question is, "How
does your electronic letter get delivered to the proper post
office after you click
the SEND button?"
Obviously, there is a lot of complex equipment involved, the
most important being the router.
Without routers, there is no Internet.
Routers and Packets
Whenever I think about what
actually takes place when someone sends an e-mail message
to a friend on the Internet, I am amazed that this technology
even works once in a while, let alone millions of times a
second. My highly technical friends may sneer at the following
explanation of the inner-workings of the Internet because
they are comfortable talking about TCP/IP
stacks,
and "improving performance by moving routing
functions down to layer
3," and other seemingly arcane topics. But most
people simply are not interested in that level of detail.
One way to think of a router
is to compare it with a heart. Both hearts and routers "pump"
stuff, but the "stuff" routers pump is information,
and that information is contained in packets.
The concept of a packet is actually quite simple.
The best analogy for a packet
is that of a letter and an envelope. The letter contains the
information, and the envelope, with the letter in it, is analogous
to a packet. So a packet may be thought of as an electronic
envelope, complete with an address, a return address, and
some information inside. The thing about packets that may
surprise you, however, is that your e-mail letter to a friend
is broken into more than one packet. "Why," you
might ask, "do we cut a letter into little pieces, send
them one at a time, each in their own 'envelope,' and then
force the recipient to tape them back together once they are
received?" Good question.
Let us continue with the metaphor
of a postal letter to better understand how packets work.
Assume for a moment that the governments of the world agreed
to a common price for all postage. Assume further that, in
their infinite wisdom, they also agreed that for a letter
with only one sheet of paper in the envelope they will charge
fifty U.S. cents to mail it anywhere in the world, but if
you put two sheets of paper in the envelope they will charge
you two U.S. dollars. (Don't laugh, governments have passed
regulations far more moronic than that.)
Now, how would you send a letter
under those conditions if it contained two pages? The most
economical way would be to place each page in its own envelope,
purchase two fifty-cent stamps, and save a dollar in the process.
Economics is also the reason your e-mail message is broken
into smaller sections, or packets. It is just more efficient
to send information in smaller, and more uniform-sized, packets.
When you click the SEND button
on your e-mail program, the first thing your computer does
is to "cut" your message into smaller pieces. It
then puts each piece in its own packet, adds the recipient's
address, adds your return address, and adds some information
that tells your friend's computer where to place each piece
in relation to the other pieces that are coming in separate
packets. These packets are then sent from your computer to
the e-mail server in your office. Next, the packets proceed
from your ISP router (which we will discuss in a moment) and
on through a series of other routers to your friend's e-mail
server. Eventually, they reach your friend's computer, which
then magically stitches all the packets back together into
a single message. Why is it done this way? Because it is more
efficient. It costs less in terms of router and computer processing
time. If you would like more detailed information on this
topic, just visit your favorite bookseller and browse through
their Internet section, where you will find dozens of texts
on the subject.
I should add one additional
comment about e-mail, however. As you begin exchanging messages
with more and more people, you may notice that sometimes the
messages you receive are not formatted very well. You may
see paragraphs with strange line spacing, perhaps with just
one or two words on a line followed by a long line of text
which is in turn followed by blank lines, etc. Don't panic.
What has happened is that your e-mail program, technically
your e-mail client, and the e-mail client
of the sender are slightly incompatible. So, your e-mail program
makes an educated guess as to how the letter should be reconstructed
after it is received. Often they guess incorrectly about page
margins, etc. Don't let it put you off, for some of the e-mail
you send to others may also have this problem. These incompatibilities
can be corrected, of course, but it is easier to live with
a little sloppy formatting rather than spend time correcting
software bugs, which often are due to the pressures of a marketplace
that forces software manufacturers into quality compromises.
Now that you understand the
concept of packets, let's return to routers for just a moment.
As you now know, when you click the SEND button on your e-mail
application, the first thing that happens, right inside your
own computer, is that your message is broken down into small
sections, and each section is placed in its own "envelope,"
or packet. After your computer has all of these packets ready,
it sends them along to your e-mail server, which, in turn,
forwards them to the router that connects your ISP to the
Internet. When that router sees the TO address on your first
packet it asks the question: "Is the recipient's e-mail
server connected to me?" If the answer is "Yes,"
the packets are sent to your friend's e-mail server. If the
answer is "No," your router asks the question, "Am
I connected to the router that the recipient's e-mail server
is connected to?" If the answer is "Yes," your
packets are sent to that router. If the answer is "No,"
it asks itself the question, "Which router am I connected
to that will get this packet closer to the router connected
to the recipient's e-mail server?" After looking up the
address of that next closest router, it sends your packets
on their way to that router. The router in your ISP office
continues this process until all of your packets are "off
its desk."
All of the routers between
your ISP office and your friend's e-mail server, in turn,
ask themselves the same questions until the first of the above
questions is answered in the affirmative, and the packets
are delivered to your friend's e-mail server. From there your
friend will pick up her or his e-mail and reassemble the packets
into the letter you wrote. It is all quite simple, actually.
At least it appears simple until one considers the fact that
the latest generation of routers can move millions of these
little packets on to the proper router every second! (Within
five years of the time this book is first published people
will laugh at how slow today's routers are.)
One question you may ask is,
"How do these routers know what e-mail servers, computers,
and other routers are connected to them?" The answer,
while seemingly simple, hides one of the most complex, and
rapidly changing, parts of the Internet. The key to traffic
flow on the Internet is in its routing
tables. Essentially, a routing table is a map that
shows the best route to every destination on the Net.
How is your mental image of
the Internet coming along? Do you remember the picture of
cyberspace you began by connecting your home to your local
electronic post office with a light-filled fiber? To make
that picture a little more accurate, place a router in front
of every post office in cyberspace. Now envision your computer
connected to your ISP router, which is in turn connected to
your electronic post office. It is important to add the router
to the picture because it is also your gateway to other routers,
which, in turn, will take you to web sites, chat
rooms, Inhabited Virtual Worlds, IVW, and all of the
other interesting places to be found in cyberspace.
By now you can visualize the
computer in your home connected by a light-filled fiber optic
cable to your Internet service provider's router. Visualize
that router connected to other routers which, in turn, are
connected to yet more routers, and on to yet others, until
you at last connect to the computer in your friend's house
without even a post office between you. In fact, the post
offices just seem to hang off these routers. So, when you
check your e-mail what you are actually doing is going through
your local router, which acts like a traffic cop, and then
stopping in at your local post office to check your mail box.
As you now see, there really isn't true home delivery of mail
on the Internet. You have to "travel" to your local
electronic post office to pick it up!
Now let's try a mental experiment.
Imagine the series of routers between you and your friend
as each being connected to several other routers that aren't
directly in the path to your friend's house. Can you see the
web-like structure that is unfolding? If you and a friend
are connected to the Internet and want to communicate directly
with each other in a chat room, under normal conditions your
messages would be passed from Router A to Router B to Router
C and on to your friend's computer. Now assume that the connection
from Router B to Router C has been broken. What do you think
happens? Router B simply sends its packets along a different
path to Router C. This new path may be less direct, but a
few broken links won't keep your packets from getting through.
It is not only a broken link
that causes a router to send some of your packets along a
different path. Perhaps the next router along the shortest
path is busy, and the packets are stacking up on its "desk."
(The technical term for this "desk" is buffer,
which is the source of the expression, "My buffer just
overflowed," that programmers sometimes use when they
can't absorb any more information at the moment.) An overly
busy router simply tells its neighboring routers that it is
running behind. The neighboring routers then take mercy on
it and send their packets along an alternate path. In theory,
but seldom in actual practice, each of your packets could
take a unique path to your friend's house. While you might
think that all of this routing and packetizing
is terribly inefficient, and that it may cause big delays,
actually it works extremely well. In fact, it works so well
that tens of millions of e-mail messages are sent and delivered
every day, and this number continues to rise exponentially!
Here was the real surprise
for me when I first learned how routers worked: The largest
amount of time consumed in this entire process is usually
at the sending and receiving machines. Your computer and your
friend's computer are quite often the worst bottlenecks in
transmitting information over the Internet, not the routers
that connect them. The disassembling and assembling of these
messages into and out of packets usually takes the most time
in this process. By the way, web pages are sent in exactly
the same manner as e-mail messages, they are broken into packets
and sent through routers. One way to speed up your Web browsing
experience, therefore, is to get a faster computer, rather
than change to a different Internet service provider.
During a conversation I once
had with a man who spent his entire career designing routers
and other black
boxes that make up the Internet's infrastructure,
I was taken aback when this highly respected scientist answered
my question about the workings of a new multi-gigabit
router by saying, "it's magic." In a sense,
this was the only answer possible.
The deeper one digs into the
essence of these black boxes and learns the details of moving
packets of information from Des Moines to Moscow, the more
the process resembles magic. The point at which understanding
bleeds into magic is different for each of us. For the Internet
novice, just clicking on a hypertext link for the first time
and watching a web page appear on the screen is often a magical
experience. For the Internet engineer who spends her days
working with complex routing algorithms, the magic may begin
to shine through at the chip level; deep within the routers
of which she appears to be the master. As Arthur C. Clarke's
(1)
third law of technology says, "Any sufficiently advanced
technology is indistinguishable from magic."
Of course, magic carries different
meanings for each of us depending upon our cultural backgrounds.
Personally, I don't like television programs that expose famous
magical illusions. For me, a little magic makes life more
interesting. So, you might wish to be careful about trying
to gain too complete an understanding about the technical
nature of the Internet if all you want to do is surf the Web
and communicate with friends. In the final analysis, all of
the technical descriptions in the world will not explain the
phenomena we call "the Internet." I see nothing
wrong with believing that one aspect of the spirit of the
Internet is magical.
What is the Internet
Now that you have a basic understanding
of routers and packets, it is time to tackle the issue of
just what is this thing we call the Internet. Given its importance
and the amount of media coverage the Internet receives, it
is sometimes hard to believe that the Internet is nothing
more than a network of networks that are connected by routers
and other such devices. Of course, the phrase "nothing
more than" belies the fact that it is also the largest
and most complex technological artifact ever created by humankind!
What then, you ask, is a network?
In the context of the Internet, the word "network"
refers to a chain of interconnected computers. For example,
many businesses today connect their computers to a Local
Area Network, a LAN, to save the expense of buying
individual printers
for every employee. By use of a LAN, several computers can
share common services such as file storage and printing services,
which often require expensive equipment. All you really need
to know about networks for the purposes of this book is that
they are somewhat complicated combinations of wire, software,
and black boxes that move information, usually in the form
of packets, from one computer or peripheral
device to another.
In addition to Local Area Networks,
which are usually confined to a single building or to a small
group of geographically close buildings, there are also Metropolitan
Area Networks, MANs, which cover relatively small geographical
areas. Another common type of network is the Wide Area Network,
or WAN, which generally covers a much larger area, such as
a state or country.
All of these types of networks,
LAN, MAN, and WAN, operate independent of each other. Not
all use the same protocols,
and each has its own administrator and rules of operation.
For a computer on one network to exchange information with
a computer on a different network, a router must connect the
two networks. (Actually, there are other connecting devices
such as gateways
and bridges
involved in interconnecting networks, but a discussion of
these devices is too technical to be included here and would
add nothing to the discussion that follows.) Two networks
connected to each other are called an "internet."
(Note the lowercase "i.") While there are many of
these smaller internets, often called "Intranets,"
the interconnection of all of them into a single, global network
is the communications medium we call the Internet.
Why, you may ask, would someone
own a network and not connect it to the Internet? There could
be many reasons for such a choice, but the most important
ones are security and economics. Although it is difficult,
and in some cases extremely difficult, for unauthorized persons
to gain access to information they aren't entitled to see
on a computer network, whenever computers are interconnected
the possibility of unauthorized access exists. A good computer
security expert will tell you that the only way to ensure
that your data remains secure is to disconnect the computer
from all networks, unplug it, and bury it in the back yard.
What they are saying is that given enough time and computing
power, any computer network can be hacked.
(2)
That, in addition to the fact that it costs money to connect
a LAN to the Internet, is the primary reason some computer
networks are not "on the Net."
There you have it, at least
at a basic level. Using your telephone line, you connect your
home computer to the network that is operated by your Internet
Service Provider, your ISP. This ISP network is then connected
to other ISP networks and to the networks of information providers
through routers. (See Figure A-4 on page 192 of the PDF or
paperback edition.)
This globally interconnected
network of networks is what we call the Internet. It sounds
simple, and from this elementary point of view it is. Only
when one begins to dig deeper into the underlying technology
do you begin to understand the "unthinkable complexity"
of this global communications medium we are building.
Fortunately, in order to use
the Internet you don't need to know any more about its inner
workings than you do about the ignition system on your automobile.
All you really need to know is how to get it started and which
roads to take. If you want to, you can leave everything else
up to the technicians, scientists, and engineers who are building
this infrastructure. The only semi-technical details you may
want to check out are the "rules of the road" and
who makes these rules.
|
|