Future Calling
A look ahead at how we'll communicate -- and
what it will take to make the future happen.
by Jerry E. Pournelle, Ph.D.Jerry Pournelle is the
principal columnist for Byte.com and Nikkei Byte; his is
the longest-running computer column in the industry. He
is a best-selling author of science fact and fiction works,
chairman of the Citizen's Advisory Council on National
Space Policy and a former White House adviser on space
and technology policy. He can be reached at
jerryp@jerrypournelle.com.
Sometime in the next dozen years, you are in a market, about
to splurge on a bottle of Dom Perignon for a celebration. Just
after the bottle of French nectar passes through the bar-code
reader, your pocket telephone rings.
It's your car, which has been checking with the house about
your bank balance. The two of them want to warn you that
this pricey purchase will far exceed your disposable-income
budget. "We're worried that this purchase will inhibit your
ability to continue making monthly payments for us," the car
says.
Fortunately, by the time this becomes a reality, pocket devices
such as your phone will still have an "off" button. But unless
you always pay cash--very unlikely by then--you won't be
able to turn off the system that records your purchases and
puts them in a general database, accessible by business entities
that want to track your purchasing habits. Meanwhile, your
purchases will also be automatically integrated into the store's
inventory database, where stock remaining will be compared
to what's desirable. If necessary, a purchase order will be sent
to the store's wholesaler, instantaneously triggering necessary
demand notifications in the grape-to-consumer supply
chain--all before you get to your car, let alone pop the cork.
When the telephone was introduced in the 19th century, one
British economic official said it wouldn't have much impact on
his nation's business because London had plenty of delivery
boys on bicycles. The communications revolution now under
way will have more impact on the way people communicate,
the way businesses interact, the way governments handle their
affairs, than the telephone ever did. And it will happen faster.
Signs of this fact are all around: Not only is the mobile phone
ubiquitous, but usage of handheld keyboard-and-screen
devices connected wirelessly to the Internet has begun its
inevitable upswing. This opening scenario may be fanciful, but
it's nearly certain that all of our devices will be smart--and able
to talk to one another--in the future. Your mobile phone will
know where you are, and your car will know where it is, and if
you're standing in the middle of a parking lot searching for the
car, you can call it and it will tell you.
(Actually, if you've left the car inside a parking structure, it
won't be able to see the global positioning system (GPS)
satellites, but it will remember where it was when it last had a
satellite fix, and a small gyroscope-based inertial platform
similar to a missile guidance system will know where the car
went from there.)
And if you forget where you are too often, your car and your
telephone will discuss with your house, which keeps records
of such things, whether you are developing a memory-loss
problem that might require medical attention. They'll keep
trying to help you out by reminding you of things you forgot to
do. They may also alert your doctor or your family.
Most of this communications capability is available today; it's
just not practically affordable. In the coming years, however,
this will change--a result of hardware evolution and a software
revolution.
All of the hardware required to build the communications
infrastructure of the future is subject to Moore's Law, whereby
everything gets twice as fast for half the cost every 18 months.
That "law" is a purely empirical observation, but it has
described the microcomputer world for the past 20 years. An
accelerated version is now being applied to optical and
electrical networking systems, as well as to various other
gadgets and gear.
Precision machining is also cheap and getting cheaper.
Enormous-capacity disk drives cost very little now. The same
technology will be used to make the inertial devices. The
hardware for communications is getting smaller, more powerful
and cheaper at exponential rates.
The software part of the future communications infrastructure
is trickier. True artificial intelligence is highly debatable, but the
kind of intelligence described in the above scenarios is merely
a complex rules-based expert system. Plain-language
communication to and from computers requires many rules,
but researchers have reduced what is required to a much
smaller set of classes and objects, plus a dictionary. A dozen
companies are engaged in a mad race to find ways to make
computers easier to use, and they will all succeed.
Speech is the natural human interface tool, and we can expect
to see our computers use it a lot more. It's already possible to
talk to computers. And while computer language recognition is
still a bit crude, companies are now demonstrating speech
recognition software that is far in advance of anything that was
even remotely available just a few years ago.
As the hardware gets better, complex programs become
easier to write. Neural-net systems, which can learn much the
way we do, were all the rage in the early days of small
computers, but the excitement died out when it hit hardware
limits. Now the hardware has leapt far ahead of the software's
capability to use it.
Each iteration of the hardware-software leapfrogging cycle
seems to take about five years, or roughly three cycles of
Moore's Law. Therefore, look for dramatic improvements in
computer communications and learning systems in about 2005,
when neural-net systems that learn from observation will come
into their own.
A feeling of connectedness
Many businesspeople and professionals, as well as an
increasing number of other consumers, are loaded with
equipment, including, but not limited to: a cell phone, a pager,
a handheld personal digital assistant (PDA) with wireless
connectivity and a laptop computer, as well as a digital speech
recording device and a digital camera in some cases.
Fortunately, consolidation of such devices is well under way.
The laptop is becoming a handheld system that contains a
video camera and digital speech recorder with transcription
capability. The mobile phone and pager have long been
combined, and the move to fully Web-enable the resulting
gadget is yielding results. Finding a practical way to combine
handheld computers and a telephone has been tricky. One that
might work for some is a tiny device that looks like a hearing
aid, contains a microphone and connects wirelessly to a
combination computer-phone-video camera-sound recorder
device. Add the GPS and inertial systems, as well as a
200-gigabyte storage device, and you've got a picture of the
future communications platform. Everybody will have one or
something like it.
Anybody can now get the answer to any question that actually
has an answer. The connectivity that this requires has
far-reaching implications. All business competition, for
example, is global. If you simply need a light bulb and some
toothpaste, you will run down to the corner store and get it;
however, if you need a dozen light bulbs and aren't in a hurry,
you may order them online, which implies a
goods-and-services location independence that will change
business strategy at a fundamental level.
Intelligent agents can go out and look for bargains on the
Web, for example, and they are improving rapidly. Rules are
supplied, such as: "Buy me one at the best price under $50,
and only deal with people I've bought from before. If there's a
really great bargain from a stranger, tell me before you order."
Project that capability out a few years, so that anybody can
have an intelligent software assistant for shopping simply by
talking to the combination phone-computer-camera that all will
carry, and you'll see a new type of competition that junks
many previous business theories.
The magazine calls itself Wired, but the future is wireless. On
the local end, the Bluetooth specification--a very high-speed
wireless networking technology under development by a
consortium of L.M. Ericsson AB (Stockholm), IBM, Intel
Corp., Nokia OY (Espoo, Finland) and Toshiba Corp.
(Tokyo)--promises to let devices of all types, regardless of
size, shape or function, communicate with one another. How
Bluetooth works, or even its short-term feasibility, is not as
important as that it, or something like it, is inevitable.
Everything else will be wireless, too. At the moment there is a
shortage of global bandwidth, despite tremendous fiber optic
network build-outs by service providers in the economically
powerful regions.
Satellites make it simple
It is 2020, 100 miles west of Umtata in the Transkei region of
the Republic of South Africa. A teenage Xhosa cattle herder
notices strange behavior in his herd. He uses a satellite-linked
wireless handheld phone to message the Health Office in Port
Elizabeth. The health officer is Zulu, not Xhosa, and while he
speaks English as well as Zulu, the herder speaks only Xhosa.
No matter. The herder's written description of the symptoms
are automatically translated into English and directly fed into an
expert veterinarian program. The program diagnosis is
encephalitis. The database indicates that all cattle in that region
were inoculated recently. The health officer realizes that a new
mutated strain of tsetse-fly-vectored encephalitis is emerging.
Fortunately, it has been reported early enough to do something
about it.
Satellites will hold the key to solving the bandwidth problem
on a more global basis. Ten years ago, one expert spoke of
the complexity inversion: As satellites become more complex,
the equipment needed by the customer on the ground can be
simpler. The implied communications benefit is obvious and
powerful for those who are unable or disinclined to spend
resources on expensive high-tech apparatus. We have known
how to build such complex satellites for years. Satellite
communications companies are going broke, however,
because of the high cost of getting their hardware into orbit.
Orbital-launch costs must be made more reasonable so that
service suppliers can develop the capability to do on-orbit
assembly. Moore's Law applies here, too: Even if launch costs
per pound don't fall drastically, the smarts per pound in the
satellite will rise exponentially. That should help.
However, bandwidth also requires power, and power requires
large satellites and big solar arrays. Big solar arrays need to be
assembled in orbit. While the development of on-orbit
assembly has taken far longer than is desirable, breakthroughs
in orbital launch costs are on the horizon. In addition, the
communications revolution will create such an enormous
market that if the current leaders in launch technology, such as
the United States, cannot make launches cheaper, others will
quickly fill the void.
New protocols will also be required to build the future
communications infrastructure. At the moment, the Internet is
built on a suite of protocols that is adequate but has a fatal
flaw: It cannot determine the quality of service (QoS)
provided, so service providers cannot bill or manage traffic
accordingly.
As a result, transmissions cannot be prioritized, with customers
paying more to get critical messages sent quickly. While we
will always use more bandwidth than we have--demand rises
to exceed supply--we are not efficiently using what already
exists.
Other data networking protocols, most specifically
asynchronous transfer mode (ATM), solve this problem. The
changeover to a new way of building the future network
infrastructure will be awkward and expensive. The Internet
actually consists of hundreds of basements and warehouses full
of routers sending Internet protocol (IP) packets to each
other. Those routers will have to be replaced with new gear
that understands ATM or some other protocol set and still
makes room for IP.
At the moment, ATM has its own fatal flaw: Its chip sets are
relatively high-priced, and the virtual circuits that it sets up are
incredibly complex and expensive to manage at the very high
speeds today's fiber optic networks can support, let alone
those that will be possible in the future. While many
service-provider networks make liberal use of ATM to
prioritize network traffic, it is the devices in the supermarket
and your office and your telephone that demand IP. It is
inevitable, therefore, that a new protocol--supporting IP but
managing traffic like ATM--will emerge. The only question is
when. The result will be much better Internet access for
everybody, with customers only paying for the QoS they need.
Come the revolution
It is now nearly 2000, and already anyone can get the answer
in about 10 minutes to any question that has an answer. This is
what the information revolution has produced. Next up, the
communications revolution will have equally profound effects
on our lives, some of them unimaginable today. The important
points to remember are:
* Everyone will be connected.
* The future is wireless.
* The future is digital.
* All of our devices will be connected, and they will all be
smart.
* All competition will potentially be global. Successful
competitors must be world class.
* The communications revolution will be as profound as the
computer revolution, with both dangers and opportunities.
And that's only the beginning. |
