Webbery Index

What is the Internet For?

What you get is what you see

It seems obvious; look around you, and you'll see what the Internet is for - anything you want it to be for. Anything that can be communicated can be "internetted". More prosaically, anything that can be expressed in digital form can be transmitted across the Internet.

All of that is true; the Internet does not discriminate. If it's digital, it goes. But it should be equally obvious that some things are going to more appropriate than others. Anything which "goes with the flow" of the Internet has a better chance of working. Anything which goes against the "grain " of the Internet may struggle.

The Founding Fathers

The men and women who created the Internet did not have a business plan. They didn't even have business in mind. Their brief came from strategic thinkers within academe and the military, who became aware that existing means of communication were strategically vulnerable, and that a society as complex as the USA (and its allies) could not afford such vulnerability.

These "designers" were themselves a mixture of academics and engineers; they weren't businessmen.

Their brief was to produce a system of communication which did not rely on any centralised distribution hub, would be transparent to content and wouldn't require replacing all the existing hardware. It went without saying that it needed to connect computers to one another - since that was where most of the complexity of modern society was expressed (even then).

American cultural influences added another unspoken condition; that the new system would not compete directly with any existing, commercial communication system - although the telephone companies of the day were far from happy about the Internet's creation; they wanted to offer net-like services on proprietary channels.

In the 30 years since the first foundations of the Internet were laid, there have been a great many changes - in the technology, in the organisation and in the funding - but those original features still matter.

It's not a new thing

In concept, the Internet is over 30 years old. As a technical reality, it's 25 or so years old. As a useful means of communication - 20 years. As a serious element of global communication - 15 years. As a serious, almost indispensable part of the commercial fabric - 10 years.

It's only in the last seven or eight years that the Internet has come to the notice of the general public (including the businessmen of the world), and many of these assume that the Internet is only a little older than their awareness of it.

The most potent force in popularising the Internet has been the World-Wide Web - and that is comparatively new. But, even so, it was invented in 1989, not yesterday. And the Web is only a secondary artifact - "floating" along on top of Old Father Internet itself. So it can't break the rules of the 'Net; it can only add a few more of its own.

The consequence of all this history is that it's too late to tell the Internet to ‘grow up and be sensible’. It's already pretty mature and fairly sensible. The problems encountered during the dot-com bubble-and-burst can be ascribed to the failure of business and finance to be grown-up and sensible.

Let's face it, not only did the bozos who fuelled that brush-fire not understand the Internet, they didn't understand business and finance either.

It's not built for speed

Probably the most common complaint of ordinary Internet users is - how slow it all is. They're accustomed to a TV channel-changer working almost instantly, a radio sounding out the moment it is switched on, a telephone ringing the moment the last digit is dialled (and a conversation beginning as soon as it is answered).

None of that applies to the Internet. A "busy" web page may take a minute to appear, an email may take hours to be available, newsgroup messages can take days to propagate.

For those who know a bit more about how it all works, it's a wonder that it works so fast. Indeed, that it works at all.

That's because the Internet wasn't designed to be fast; it was to be reliable - to make sure that a message would get to its destination, eventually.

It's also worth remembering that those messages, in the early days, were mostly small chunks of text - nothing like the huge binaries being bandied about today.

Of course, the "pipes" down which this material passes are a lot fatter today, but there are still plenty of bottlenecks.

Any enterprise which depends on raw speed is pushing its luck on the Internet.

It's not built to broadcast

The Internet was not designed to compete with television or radio or cinema. Indeed, it would be difficult to imagine a more hostile environment for continuously-transmitted material.

The Internet works by taking digital information and chopping it up into smaller "packets". These packets are then sent off to their destination, individually - each with their own address label. Each of these packets will find its own way to its destination - often using several different routes, at several different speeds. When they arrive, the receiving machine then attempts to put them back together again in the right order.

That isn't a healthy road for a song or a movie to travel. It's entirely possible for your heavy metal anthem to arrive backwards (and we all know what happens if you play them backwards, don't we children?).

The Internet Protocol (IP) we've used for the last decade or so is version 4. Over the last few years, a lot of very clever people have been working on a new version - version 6 (version 5 was the one where each packet was wrapped in pink ribbon and delivered by pigeon; we don't talk about version 5).

One of the features of IPv6 is the ability to label each packet with its place in the larger object - which should make streaming a bit more reliable and a bit less jerky.

Unfortunately, the use of IPv6 is not compulsory, so any streamed object is unlikely to find an IPv6-aware path to its destination.

That said, a great deal of clever work has been done by the wizards of streaming media. These days, it almost works OK.

The problem is - success is the enemy. If more people try to receive a "broadcast", then the overall speed reduces, for everyone. There are shortcuts available ("edge networks"), to deliver streams closer to the audience and handle greater numbers. But these cost money. With a TV or radio broadcast, a larger audience lowers the unit cost of the transmission. On the Internet, the cost rises with the size of the audience.

At the time of writing, a lot of smart money is going into delivering audio and video streams, banking on the spread of broadband connections and other technology to make this a profitable pursuit, but this is still going against the grain of the Internet. Consequently, any enterprise which attempts to emulate broadcasting across the Internet is unlikely to make a fortune.

It's not built for people

I'm sure a lot of people have understood this instinctively, without necessarily knowing why it should be so. The reality is that the Internet is a network of networks, and a network is a matrix of computers. If a human being happens to sit in front of one of those computers and tap in a few commands, that might be very interesting - but the Internet isn't impressed.

At the time of the Internet's genesis, most of the networks it connected consisted of one or two stonking great computers, serving a few dozen fairly dumb terminals.

An Internet address only needed to identify the boss computer. That machine would then identify one of those terminals to complete the last leg of of an IP packet's journey.

You've seen numerical IP addresses, in the format A really big network (like IBM, for instance) would be allocated a range of numbers - all starting with the same triplet of digits (123.etc.etc.etc). A slightly smaller organisation would be allocated a smaller range - all starting with the same two triplets (123.124.etc.etc).

And so on - until you reach a single computer, linked to a larger network, which would receive one, definitive number (

When this system was devised, it was thought that it would last almost indefinitely. But for some time now, these numbers have been running out. So, most individual computers no longer have a fixed IP address. Instead, these are allocated dynamically by the ISP.

This shortage of IP addresses is one of things IPv6 (see above) is intended to cure. However, it's going to be a major task to establish this new system.

You may be more familiar with name addresses, like dc@strum.co.uk or www.strum.co.uk. Don't be fooled - they're not the real thing. They're just a sop for us "organics" (that's what computers would call us, if they had any sense of humour).

What, then are the consequences of this? That any enterprise which thinks the Internet can provide reliable identification of their customers is kidding itself.

The best the protocols can do is identify a computer - sometimes not even that. Identification will always require something else. The problem is - any additional protocol can be copied/forged, if it's worth someone's time to figure out how.

Possibly even more worrying, an enterprise's customers can't be absolutely certain of identifying it.

I don't want to overstate this. It isn't a technical issue, because every company worth its salt will have a static IP address. However, ordinary Internet users have come to sense the forgeability of Internet communications, and are often dubious about the bona fides of messages they receive.

For this reason, I believe it is very important for a company to take care to establish and maintain a "house style" - not just things like logos and colours, but also language and behaviour. A regular customer should be able to recognise who he's communicating with - even in a text-only email.

It's not built for security

For a system created, largely, on military money, it's a little surprising that the Internet seems about as secure as a paper piggy bank.

There are two basic reasons;
  1. when the system was designed, the few computers that existed tended to have their own access security
  2. these computers weren't using consumer-level software (like Microsoft's) that anyone could practice on at home
  3. in those days, the military had encryption technology pretty well tied up. They could send encrypted material via the Internet with some confidence that no-one else could decrypt it

Things are different today; sometimes it seems that every teenager knows something about cracking codes, and there are close to half a billion computers connected to the Internet (and with the Internet connected to them).

Viruses, trojans, site defacements, unauthorised accesses - all make the Internet seem a very hostile environment to do business in.

But it's not nearly as bad as it seems. Most of the security problems experienced by businesses on the Internet have come about because of sloppy, lazy, ignorant system administrators. (If you are a sloppy, lazy, ignorant system administrator, I'm sure you're quite nice really).

Those super-hackers in the movies - the ones who can break into a top-secret government database while the computer store assistant has his back turned - that's fiction.

There are usually much easier ways to compromise security - open doors, logged-in PCs, unobservant staff. Or sloppy, lazy, ignorant administrators who haven't got round to installing the latest patch to the server software, despite a well-publicised security alert.

So, despite the fact that the Internet was not set up to be secure, any enterprise which finds its security has been breached, probably deserved it.

it is a point-to-point network

In the beginning, the purpose of the Internet was to get a message from one point to another. Still is.

The reality may be that the technology makes it a point-to-point-to-point-etc journey, but the idea is to reach one place, a particular computer. With luck, that can be refined to make it reach a particular human.

Above, I have mentioned the technical problems in trying to use the Internet as a broadcasting medium. But there's more to it than just IP packets and bandwidth limits.

Every Internet connection is a single one-to-one conversation. So, one-size-fits-all doesn't fit the Internet. There's no technical barrier preventing a company establishing a website which says what they want to say - to anyone who wishes to visit. But there is a cultural barrier, which will favour a company whose website says what each visitor wants to know - especially if they also listen to what the visitor has to say.

The Cluetrain Manifesto is an interesting expansion on this phenomenon. Its assertion that ‘Marketing is a conversation, not a soliloquy’ seems to be a case of the "real world" learning from the realities of the Internet.

IBM developerWorks also has some useful advice about this.

There's another side to this point-to-pointness; as far as the Internet is concerned, each end of the connection is as important as the other. (In reality, the "client" initiating the connection is more important than the "server" delivering content.) Too many Internet sites behave as if visitors are supplicants, begging favour from some mediæval court.

But this "democracy of the Internet" goes further still. The conversation doesn't end with the connection between vastco.com's server farm and serf@aol.com's PC. The Internet can also connect your customers to each other.

Recent years have seen a rash of "VastcoSucks.com" sites, where disgruntled customers of an unresponsive company simmer their anger into a potent stew. Many corporations have responded to these sites by trying to close them down - using expensive lawyers and supine arbitrators. Dumb. These same companies spend fortunes on market research, to find out what their customers think of them - then spend fortunes to stop themselves hearing the answer.

It doesn't have to be so negative. Some customer sites contain more and better information than the company site. Some companies have learned to participate in customer sites, to respond to customer problems, to find out what the customer really wants (and deliver it). That's Internet thinking. The Internet is on their side.

Webbery Index
© 2001 David Craig Send me a message