Going mobile: the next step to pervasive computing

Nick Langley measures the gap between hype and reality as mobile computing takes us into the pervasive era

Nick Langley measures the gap between hype and reality as mobile computing takes us into the pervasive era

One day, hopefully in the near future, digital information will pervade our lives. The fridge, the pool table in the pub, the car, your wristwatch and the fabric of buildings could be "clients" in the same way mobile phones, personal digital assistants (PDAs)and PCs are today.

But, for now, you still need three separate power-guzzling plastic boxes if you want to type, talk and keep your appointments handy.

The suppliers' visions of a world pervaded by invisible Web links at every level - business, social and domestic - co-exist with product announcements that amount to little more than providing limited e-business and Internet information to wireless application protocol (Wap)-enabled mobile phones.

This gap between hype and reality did not matter when business computing was just about PCs and local area networks. But mega-billions will be spent worldwide in the next decade acquiring the ownership of wireless and fibre-optic networks. Pervasive computing is what will deliver return on investment to the network carriers and backbone owners, so IT strategists need to start asking where their business benefits will come from.

IBM, for example, tells us: "E-business content can now be delivered effectively, efficiently, and economically to anywhere, and to any device... [pervasive computing] provides convenient access to relevant information from disparate, unrelated sources, stored on one seamlessly integrated system." Well, not yet it can't.

We will access this information through laptop, notebook and handheld computers, pagers, "smart" phones which incorporate the functionality of PDAs, and wearable devices, such as smart badges or tags and wristwatch computers.

The reality at present for most users is expensive, unreliable, low performance mobile networks, geographically limited, and delivering cut-down versions of Web pages to phone handsets with unfeasibly small displays.

The services offered are equally limited, and often available more conveniently and cheaply elsewhere. You can get share prices and test match scores over conventional telephones, from newspapers or the radio.

If you want to know where you are, and how to get to where you're going, you can ask a policeman.

"It's debatable whether ordinary consumers are actually demanding mobile e-commerce services right now," says Ovum analyst Duncan Brown. "Business users, rather than the mass market, will be the first serious adopters, but even they won't pay a premium for services which are easier and cheaper to access using a phone or PC. If suppliers are to survive and prosper, their early offerings will have to be very targeted, and very compelling."

Behind the hype, IBM has been getting down to defining strategies which overcome the barriers to pervasive computing. The low performance of wireless networks means mobile users operate mostly in disconnected mode. But even when not connected, they must be able to browse Web pages, read and respond to e-mail, and access calendar information, groupware and productivity applications. When mobile users do connect, they must be able to access enterprise applications and synchronise their local data with the server.

What they get depends on "dynamic content adaptation", or transcoding, which addresses device characteristics like screen size, format and graphical capability, speed of network connection, network charges, and user preferences for what is displayed and how.

Transcoding software which converts and reformats data will be packaged as plug-ins for IBM's On-Demand Server.

All this may do the business for IBM users, but how do you make your architecture so ubiquitous that anyone can use it? "The good thing about the Web is that it's already ubiquitous," says Hewlett-Packard Labs' business development manager, Gene Becker.

HP says the Web supports mobility in two senses. First, resources on the Web can be accessed from any device that supports the standard HTTP protocol. It's a simple matter to put HTTP into devices nomadic users encounter, like printers and projectors. Second, the Web allows mobile users transparent access to resources outside their current environment.

At a local level, HP aims to deliver Web services to mobile users without requiring a global wireless connection. This has the advantage of minimising how much of the infrastructure needs to be up and running for users to interact with local services. That may be the future as HP sees it. The more prosaic present is a bundle of e-services for Wap, such as HP Openmail, which enables users to send and receive e-mail from Wap-enabled cell phones.

Wap service providers could be riding for a fall, the victims of their own hype. "2G+ [technology midway between second and third generation mobile] is so limited in bandwidth that excessive Web graphics can be a nuisance," says Ovum analyst Iain Stevenson. "In its initial release, the practical link speed for GPRS will not exceed 14.4 kilobits per second - the speed of an analogue modem in 1994."

Stevenson thinks suppliers are hyping the cellular market to a worrying extent. "This is creating false expectations among users, and possibly leading to inadvisable investment decisions." he says.

"The cellular technology that will become available in the next three years is no revolution. It is limited in bandwidth, and data services are likely to be priced at premium rates. There is certainly a market for information services and low value e-commerce to cellular phones, but the technology will not support the multimedia applications that are widely touted. You will be able to buy theatre tickets and place bets over the Internet from a 2G+ cellular phone, but a videoconference with your Aunt Mabel in Sydney is simply impossible.

"In short, the industry is trying to promote a rather weak 2G service with a vision of 3G services. It is not until 3G is available that radically new applications will become possible - with new terminal equipment," Stevenson adds.

And while the bandwidth issues are being addressed, the problem of small screen size will persist.

Ovum predicts that to succeed, pervasive computing demands new, and often uneasy, alliances. "The industry seriously needs to rethink its priorities," says Brown. "Top of the list must be the industry agreements and co-operation needed before any wireless transactions can take place."

Brown says the new industry will consist of a web of open partnerships between players from completely different backgrounds, providing a range of competing and complementary services. They will include device manufacturers, network operators, financial service providers, content providers and aggregators, systems integrators and infrastructure providers. This will produce short-term disputes as they all try to "own the customer", or stray onto each others' territories.

Back in the world according to the visionaries, two projects offer a glimpse of the wired world of the future. AT&T Cambridge Labs' Spirit project is the latest incarnation of the "intelligent building". The network knows the exact location of every person, and every piece of equipment in a building. There are large screens in public areas, showing cutaway views of the building and its occupants going about their business. If you need somebody, you can see who they are with and what they are doing.

When you sit at a workstation, the network brings up your desktop in front of you. When you walk down the corridor to talk to a colleague, your desktop follows you.

Each floor of the building is treated as a big graphical user interface. A mobile desktop application registers a space around each display, and around each person. When a person space becomes contained by a display space, the person's desktop comes up on the screen.

In HP's Cooltown, people, places and things are all "citizens of the Web". They are connected to the Web, have Web facing representations, and can offer and participate in services on the Web, providing systematic linkages between the Web and real world entities.

As you enter a Cooltown conference room, you collect the room's URL on your PDA or cellphone, enabling you to control printers or digital whiteboards. You don't need to access a global network to use such local services, although the room's "Place Manager" acts as a portal to the wider web if you need it.

You leave Cooltown on a Web Bus, equipped with both intra-bus and Internetwireless connectivity. Passengers pick up its URL and use its services via a Web browser. The bus services can be "location-aware": the computer contains a global position sensor. A would-be passenger waiting at a bus stop can pick up the URL of the bus, and the bus can ring them up and tell them where it is, and how long it's going to be.

That Web bus could stand as a metaphor for pervasive computing. From the industry's point of view, it represents the opportunity it mustn't miss.

But from the customer's point of view, a closer analogy would be with our deregulated bus service, with vehicles of all shapes and sizes, in dozens of confusing liveries, taking risks with their passengers' safety to overtake or force rivals off the road, charging different fares to get to the same destinations by different routes - and all coming along at once.


  • Filtering reduces connection time and costs by discarding data devices are incapable of using, or users unwilling to wait or pay for, and converting formats to make images smaller

  • Data and applications can be cached, "pre-fetching" Web pages which can be refreshed while connected, then browsed when disconnected

  • Server-based synchronisation can eliminate the need for handheld devices to run a complex desktop, because the network maintains the user's data

  • Disconnected operations can be controlled by proxies, which intercept requests and act on them by, for example, batching requests and adapting content

  • Alternatively, intelligent agents can act on the network on the user's behalf, finding, filtering, personalising and queuing information, and negotiating for services

    The market

    According to the McKenna Group, the pervasive computing market could be worth $120bn in the next five years. IDC predicts 18.9 million handheld PC companions will be shipped worldwide in 2003.

    Ovum expects the number of mobile devices to exceed the magic one billion mark by 2003, with a large proportion technically capable of mobile e-commerce. If the market fulfils its promise, Ovum predicts that end-user spend on services will rise to more than $200bn in 2005.

  • Read more on Wireless networking