The future of…

Click here for contact details, other articles and personal details

The future of the net

April 2002

According to the WWW Consortium, the 'Next Big Thing' will be the semantic web. This will allow logical relationships between pieces of information to be described. Information from many different sources can be linked and processed. This will allow a lot of new knowledge to be created by machines, and tidy up many administrative tasks. Our computers will be able to do far more for us automatically, and provide us with much more valuable information. However, in spite of its high ideals and theoretical technological potential, I think its impact will sadly be limited because people will not willingly surrender the value of their hard-won intellectual property and commercial contacts and relationships. Let me illustrate this with an example. I am trying to find a holiday cottage in Dartmoor for our summer break. A simple search brings up dozens of companies that act as agents for a handful of cottages each, as well as a lot of individual cottages. Many others presumably are not accessible on the net at all. There is no single institution that I can go to to get a list of all the available cottages in the area, even though hat is exactly what I want as a potential buyer. Checking out each individual site is a very tedious process, so I may give up and go elsewhere. These companies could all gain customers by co-operating, but this would mean each of them sacrificing their control, and risk losing commission. We see the same commercial struggles for market share recurring in every technology area, however much evidence there is that it puts customers off buying at all. It always ends up with companies preserving a share of a much reduced market. Technology changes but human nature doesn't! Ultimately, companies want to make money, and will only give customers what they want if it serves their own financial interest. If the semantic web allows machines and customers to shop around more easily, many companies will do their utmost to resist it.

I'm pretty sure the next 'Next Big Thing' will be 'Back To Basics'. We will have trillions of tiny sensors, processors, actuators and communicators infesting every area of our everyday environment, even our bodies. Most will be invisibly small, and have just enough capability to do their job. By keeping them simple, they can be low power and cheap, and the overall system capability will emerge from myriads of simple interactions between these devices. Ultra-simple, organically inspired approaches will revolutionise software, hardware and communications. The net itself will become more of a textureless mat, perhaps woven from quadrillions of ultra wide band connections between these tiny devices. The net will simply become the ether.

A range of technology trends are converging in this direction. First and probably most importantly, we have biomimetics, copying ideas from nature, which gives us a range of new tools. Nature makes very good use of emergent behaviours, allowing colonies of simple insects to run very sophisticated systems. Even the simplest bacteria are essentially mutually beneficial complex alliances of many simple processes. Nature has already given information technology the ideas of self organisation, evolution, emergence, immune systems and network management, just for starters. These can all be used to very good effect to simplify systems and make them more robust and reliable.

Secondly, many people are concerned that although chips today are very much faster than those of twenty years ago, subjective response time hasn't improved much. We have squandered most of the extra processing power by making software very obese. My first computer was an Apple 2, which used a 6502 processor. If this were built using the same technology as used for the Pentium 4, it would measure 0.17mm across, run 2000 times faster than the original, and cost a few pennies. But the PC using a P4 will take many times as long to boot up as the Apple 2, and the response time for many applications is little better, even though it has lightning fast processing under the lid. The fastest computer I have ever owned was in 1992, a Macintosh 2fx. Much of the extra functionality that we have since then is of debatable worth, and mostly just gets in the way or introduces security problems. If the computer just loaded what we need, as we need it, on the most basic operating systems, we could have almost instant response most of the time. We would need some major redesign, but it would probably be worth it.

The third trend of putting chips in everything is related to the miniaturisation and cost reduction. It is cheap to add intelligence to all kinds of everyday objects, so we are doing so. Chips are rapidly becoming ubiquitous, and as they get smaller, cheaper, and consume less power, this trend will doubtless accelerate. This applies not only to processing chips but also to sensors, communicators and actuators, especially micro-electro-mechanical systems (MEMS). The presence of so many devices around us will greatly encourage people to develop systems that make use of them in many different ways. This will be made much easier by keeping the devices simple, with a few more sophisticated devices dotted around to link their functions together and intermediate. We may see large numbers of devices communicating and co-operating amongst themselves for local functionality, with other devices helping them communicate and inter-operate with the wider world when necessary.

Also, the increasing problem of security will force us to accept more appropriate functionality, instead of expecting everything to be integrated with everything else and on tap all the time. I for one will be very happy if future office software isn't capable of executing macros that can spread viruses to my hard drive or access my address book, and if future browsers aren't capable of running harmful or irritating software hidden in a web page. All this extra functionality isn't worth a penny if it means I can't safely do business. I and many other people have already disabled scripts and Java on my browsers and I simply don't bother with sites that insist I switch them back on unless the task is absolutely essential.

The means of communication itself is also changing rapidly, with symbiotic networks likely to dominate in the future. Being a bit more daring, I would be surprised if ultra-wide-band communications doesn't also dominate in the long term, with devices communicating at extremely low power levels to devices very close by, trickling their date from device to device until it reaches the destination. Such communicating can be designed to be very effective, but needs to fixed network infrastructure, using highly distributed intelligence. Again, this methodology lends itself very well indeed to systems comprised of many simple devices rather than a few sophisticated ones.

Finally (unless you can think of other trends), the need to get rid of cables is pushing us to use energy much more efficiently in our devices, so that batteries last longer. With the myriads of tiny chips where the only power supply is by gradual energy harvesting, power management is all important, and that means they will try to avoid any function that is not essential to the task in hand.

So I believe that the future of the net will be a much more ethereal, organically inspired continuum of tiny, ultra-simple devices and communications links. Communication between the devices and interactions between logic-encoded information structures will create an almost magical infrastructure that will give us access with very high performance and reliability and fewer security problems. It will be held back only by human resistance to trust and co-operate, which will doubtless still fragment the world into competing empires. But we can't expect technology to fix everything.