The Future of Tomorrow’s Internet
Thursday, November 11, 1999
By Matthew Symonds – The Economist
For a decade or more Microsoft’s monopoly of PC operating systems gave it a stranglehold on computing. The technologies now shaping the Internet mean that a single software company is unlikely again to attain such dominance.
WHEN Microsoft fought its way out of IBM’s shadow in the early 1980s, it was hailed as a monopoly-killer par excellence. The PC counterculture would liberate ordinary people from mainframe computers run by corporate IT departments — most of them made by Big Blue. As it turned out, this was only half right. The transition to the PC almost destroyed IBM as a single company; but Microsoft gained a share of the market for PC operating systems that even IBM never achieved for mainframes.
Now it is Microsoft that feels threatened by a technological shift. But this time the firm that dominates the old technology is unlikely to be pushed aside by a single newcomer. Instead the Internet promises to create a world in which no firm ever again has the power of a Microsoft or an IBM.
Judge Jackson described the threat to Microsoft in his findings. The Internet could “oust the PC operating system from its position as the primary platform for applications development and the main interface between users and their computers.” If so, Windows would become just one not very important piece of computing plumbing.
Yet, as the judge also made clear, this vision lies far in the future. Although the Internet is already affecting people’s lives, a further transformation is essential if the hype about what you can do online is not to fall flat. This is still early days for the Internet–it is about as developed as commercial flying was before the DC3 — and yet the transformation is already on its way. It will not come from a single dramatic technological breakthrough, like the jet engine, but from lots of connecting technologies coming of age more or less together.
These technologies are quite unlike the PC operating system. That is the intellectual property of Microsoft, but the Internet employs standards and protocols that are “open”, meaning they can be freely used by anyone. They have mostly been hammered out in public forums, and they are beyond the control of any single firm. The operating system is the control centre of the PC, but the Internet is managed at many levels, some within devices, some on the network, others at the abstract level of a “language”.
All this adds hugely to the complexity of the system, but it is the complexity that distinguishes a market economy from a centrally planned one. The PC has to evolve to the drumbeat of Microsoft’s programmers in Redmond, Washington. The Internet, by contrast, derives its adaptability precisely from its amorphous nature. If a firm creates and exploits a monopoly over one bit of technology, somebody, somewhere will think of a road around it.
The best way of looking at the Internet of the future is to ask what is wrong with today’s. Instead of one single, huge failing, the Internet has three middling, but infuriating, ones. First, it is still not free from Microsoft’s monopoly. To gain full access to the network, most people use a PC, a machine that is prone to crashing without warning and is awkward if you are on the move. Indeed, thanks to the PC, the Internet has a level of inconvenience that would be unacceptable in any other mass-market medium. Second, the Internet itself is hard to use. Consumers have trouble finding what they want, and companies intent on doing e-business struggle to integrate new technology with their existing systems. Third, the network’s security and reliability are both inadequate.
What are the technologies that will fix these problems? Much of the answer lies in new standards, or protocols, that will let computers, handheld devices and mobile phones talk to each other in new ways. A protocol called POP (post office protocol), for example, lays down the rules for delivering e-mail. Another, called HTML (Hypertext Markup Language), defines the layout of web pages. Rather as conventions in the game of bridge let partners use bids to convey information about their hands, protocols tell the designer of a piece of machinery or software how to format information coming from or going to the outside world.
Protocols are dull, unremarkable things when written down on paper — like recipes or magic spells. Yet they make the Internet work. Software wizards have conjured up a set of new protocols, each designed to make a particular Internet problem vanish in a puff of acronyms. It is largely these protocols, combined with some clever new software, that are destined to transform the Internet into a medium that is fast, convenient, and reliable.
Perhaps the Internet’s most dramatic shift will be to extend computing beyond the PC. The first to log on will be smart phones and wireless Palm Pilots, but other devices will follow. Engineers may take a few years to come up with transmitters, receivers and sensors cheap and tiny enough to put into the appliances of the networked home. But almost everything that can usefully transmit data will eventually be able to do so, including refrigerators that automatically reorder what has been taken out of them and car keys that tell the Internet (and you) where you left them.
Some devices may use an operating system from Microsoft, but the majority will not. Most of their intelligence will lie on the network, which means that operating systems will matter less than standards. These standards are mostly already defined. They tend to be open, precisely because millions of devices from thousands of firms will be connected to the Internet.
One of the most important, called Bluetooth, is the work of more than 500 firms, led by Ericsson, Nokia, Toshiba, Intel and IBM. Bluetooth defines how devices should transmit data to each other.The idea is to replace the cables that tie devices to one another with a single short-range radio link. Within two years, about 80% of mobile phones will carry a $5 Bluetooth chip that can connect them to similarly equipped notebook computers, printers and, potentially, any other digital device within about ten metres. As well as defining how these devices find and talk to each other, Bluetooth also ties into existing data networks, including the Internet. In future, you might tap out an e-mail on your Palm Pilot or Psion, tell it to make the Internet connection through your mobile phone, print a copy on the printer upstairs and store another on your PC.
Standards are also open because customers and suppliers have learnt to be wary of being locked in to someone else’s technology. Several firms have defined their own standards to adapt the Internet for the small, monochromatic screens of handheld computers and smart phones. Such gadgets are not suited to the lively, colour graphics on most web pages — let alone bandwidth-devouring multimedia and animation. Worse, they have neither keyboard nor mouse, and lack the computing power and memory they need to run modern browsers.
But proprietary efforts have not got far. Wireless Internet access will take off only if there is a single, open standard for delivering suitable content to wireless devices. That is the job of something called WAP (Wireless Application Protocol). This allows mobile devices to gain access to the Internet using a “microbrowser”, which displays web pages specially formatted for tiny screens.
Nokia’s first WAP telephone will be available soon; Ericsson has just launched a WAP-based wireless notebook computer. Handelsbanken, a Swedish bank, has extended its Internet banking service to mobile devices. Using WAP, the bank’s customers will soon be able to look up stockmarket prices, buy and sell shares, consult their accounts, transfer money and pay bills — any time, anywhere. Within a few months, Motorola will be selling WAP telephones in America, where firms such as AirTouch and Sprint PCS have announced new services. As wireless bandwidth increases, the only limit on what they can do will be the size of the screens.
Like Bluetooth, WAP is supported by the largest firms in the business. It will work with any of the mobile networks across the world, and eventually with proprietary operating systems likely to be in smart-phone handsets, such as the Symbian joint venture’s Epoc (from Psion) or Microsoft’s own Windows CE. In effect, WAP lays down the rules for the wireless incarnation of the web. Like the web, it is not owned by anyone, yet it is the standard everyone will follow.
A kinder, gentler Internet
By escaping from the desktop, the Internet will both escape Microsoft and become a bigger part of everyday life. But this will not be enough to deliver its promise. For that, the Internet must also become easier to use.
John Patrick, a founding member of the World Wide Web Consortium who leads IBM’s efforts to create web technologies, calls what comes next “Natural Net”. Today’s Internet is a powerful way to communicate, given e-mail, instant messaging and chatrooms. Even so, people cannot really work together over the public Internet. Again, each of many different technologies has a part to play in making collaboration feasible. And again, none of them holds out the prospect of domination in the way that the operating system dominates the personal computer.
Take the infrastructure of the Internet. Videophones have failed to take off, and videoconferencing is still a minority sport, mostly because the Internet is still so clunky. The Internet needs more capacity, or bandwidth, both in its main backbone and in its connections to schools, houses and offices. The network must also be reliable, so that connections will not only be fast, but also available whenever needed.
Broadband technology can exploit four channels needed for online discussion — text, voice, video and graphics. Once broadband is common, people can work together, wherever they are, on anything from the design of a new car to a patient’s MRI scan. Forrester Research, a consultancy, reckons that, by 2003, live collaboration, now used only in one-off projects between companies, will have become as much part of the Internet as e-mail is now.
But who will own the broadband connections? On the Internet backbone, where there is no risk of a natural monopoly, collaboration is already the pattern. The backbone is struggling to keep pace with demand that at present appears to double every six months. Relief may come from the “Abilene Project”, or I2 (Internet2), as it is known. Some 150 universities and firms such as Cisco Systems, Nortel, Qwest and IBM are together designing a backbone advanced enough to solve today’s Internet gridlock.
In local connections, where the risk of monopoly is higher, fast connections are arriving thanks to both cable modems and DSL (Digital Subscriber Line), a technology that turns traditional copper wires into fat data pipes. Each approach has strengths, and each offers “always-on” Internet connections as much as 30 times faster than the speediest dial-up modems. This will hugely increase both the range of Internet applications — with high-quality streamed video to the fore — and the convenience of gaining access to Internet-based services.
The key thing is that DSL and cable represent the claims of two separate and hostile industries to be the standard-bearers of the data revolution. There is thus the promise both of rapid investment (after many delays) and genuine price competition.
Not all areas are free from the threat of domination. This is partly because high-tech industries are prone to “increasing returns to scale” and high barriers to entry. One example is speech-recognition and machine-translation technology. This could, in Mr Patrick’s words, turn the Internet into a “multilingual real-time intercom”. It will, he suggests, soon be possible to hold a conversation in English with a friend in Tokyo who is reading your words in Japanese. As
speech-recognition algorithms improve, the spoken word could become the main interface to the network. This will be especially valuable with mobile devices, because it overcomes the limitations of a fiddly keypad and a screen-tapping stylus.
Yet speech recognition depends on dictionaries of words and sounds that are hugely expensive to compile. Because most costs are upfront, the more software a speech-recognition firm can sell, the lower the average cost for each copy. A firm that thrives in this area — Lernout & Hauspie, in Belgium, stands a good chance of doing well — will make a lot of money.
However, few of the core technologies of the Internet are easily owned. Take the example of what might be called dumb data. A mass of information resides on millions of websites, but how do you turn it into something useful? Most of it is unstructured and stored in a non-machine-readable form (eg, English), which means that it can be read only by humans. Because data are dumb, web search-engines often fail to find a sensible answer to queries — a page containing the word “Apollo” might be about mythology or about moon landings. Similarly, although companies can display online catalogues or details of their stock, this information means nothing to other computers.
At the root of this problem is HTML, which defines web pages. This instructs a web-browser how to lay out the contents of a page, but does not tell the computer what the page is describing. A solution is to hand, in the form of an extension to HTML, called XML (eXtensible Markup Language). This adds invisible labels (called “metatags”) describing the objects contained in the web page, so that computers know what they are dealing with. The specifications, price and availability of each product in an online store could, for example, be labelled with metatags. A computer would then be able to compare prices across several stores and make a recommendation, no matter how much various websites differ from each other in appearance.
Once information “knows what it is”, Internet searches will become faster and more accurate. Metatags will make it easier to extract information from a page to suit a particular use, or the limitations of a particular device. In addition, XML will help firms to strengthen their online business relations with suppliers and customers, by knitting together business processes and software applications both within and across companies.
It is inconceivable that a lingua franca such as XML, which allows previously incompatible computing platforms to understand each other, could be owned by anyone. For a start, the egalitarian ethos of the Internet resists an authority “imposing” its will — cyber groups have, for example, used peer pressure to quell the bureaucracy that manages the Internet’s domain names. Since no single organisation controls the Internet, XML could not be imposed in any case. Instead, the Internet muddles through: XML has been on the drawing-board for several years, and it has won almost unanimous support from Internet standards bodies, such as the World Wide Web Consortium and industrial firms. Lastly, in all its ramifications, XML is astonishingly complicated. It will not be complete until each industry has the electronic schemas that describe common processes for each industry — and the world is still waiting for its various Dr Johnsons to compile the necessary dictionaries.
But can you trust it?
And when there is no protocol to hand? The Internet grinds to a halt, until someone comes up with an answer that users feel comfortable with. That, at least, is what has happened with Internet security.
This is essential if firms are to throw open their electronic doors to the world. Consumers are, if anything, even more wary over the lack of security on the Internet than are firms. According to Mr Patrick of IBM, e-commerce and business collaboration will take off only if several conditions are met: authentication (checking that someone is who he claims to be, or a website what it seems to be); authorisation (controlling access in a sophisticated, “granular” way ); confidentiality
(keeping private information private); integrity (ensuring that information has not been tampered with); and non-repudiation (being sure the terms of a transaction are binding and legitimate).
In theory, the solution is obvious. Encryption so powerful that it cannot be broken without mammoth computing power or the theft of a digital “key” from one of the parties to a transaction would pass all Mr Patrick’s tests. Reality is more complicated, mainly because no single open standard prevails on the wider Internet. Various proprietary systems are competing for attention, but no single one has taken off, because they are complicated and users fear that they will be lumbered with a particular technology. What is needed is another open protocol.
Coming to the rescue is the PKI (Public Key Infrastructure). By presenting “digital certificates”, the online equivalent of a handwritten signature, those who register with a “certification authority” will be able to use the public Internet to transmit data securely, make payments and prove that they are who they say they are. This will make it possible to check, for example, that an e-mail really does come from who it says it comes from. Certification authorities manipulate digital “keys” that authenticate the identity of each party to a networked transaction. These services are already on offer from firms such as telephone operators and banks.
With a lab-full of new technologies and protocols, the Internet is likely to overcome its current drawbacks and continue its march into every corner of modern life. No single technology on its own can solve the Internet’s failings. Moreover, since no single organisation controls the Internet, every new protocol will have to prove its worth if it is to be widely adopted. Such an apparently haphazard way of doing things might seem like the Internet’s fatal weakness. In fact, it is its greatest strength.
Monopolies still occur in computing: they do in many industries. But, thanks to the nature of the Internet, they are likely in future to be modest, and limited in scope. When Bill Gates concluded that Netscape had to be stopped, by fair means or foul, he could be accused of ruthless bullying — but not of paranoia.