Karen Dearne | November 11, 2008
COMPANIES are facing massive shocks from information and communication technologies that will radically change existing business models.
Output from a 3D printer
The ripples that began with the internet will have second-round impacts we can only begin to imagine, according to Digital Disruptions, a report from CSC's Leading Edge Forum.
Bill Koff, chief technology officer of CSC's innovation office, says the coming shocks - on par with the invention of the telephone and the motor vehicle - will transform the market and society so completely "it will take decades for their full impact to be realised".
"These disruptions are powering the networked information economy and may be the biggest challenge to the status quo the world has ever seen," Koff says.
CSC has identified seven digital upstarts already affecting today's business models: new media, augmented reality, social power, information transparency, digital spectrum, new platforms and a smarter world.
In the never-ending quest to make computing faster and cheaper, the platform is undergoing an extreme makeover.
"Overall, makeover is about optimising performance: getting the right power to the right applications, not wasting hardware or energy, not running software you don't need," CSC says.
"The next generation of new materials, which will take us well beyond Moore's Law, is on its way."
Some key points from the Leading Edge Forum:
This is not your parents' computer platform. In 10 or so years, today's silicon chips will be replaced by new materials.
As Moore's Law comes to an end, after some 45 years of silicon-based processing, far smaller, lighter materials will be at work: atoms, DNA, electron spins and light.
While this may sound futuristic, CSC says it's not science fiction.
Nanotechnology, for example, allows things to be built atom by atom.
"Although atomically precise manufacturing and self-assembly are quite a way off, the use of nano-sized materials could pave the way for revolutionary applications in medicine, smart materials and energy," the forum says.
"This includes precisely targeted agents for cancer therapy, artificial blood and equipment that self-repairs."
Task: 3D printing.Consumers will be using 3D printers at home to print design specifications from the web and produce anything from toys to spare parts to landscape and furniture designs.
Reality check. Existing 3D printers yield a solid object that you can hold in your hands. Prototype 3D printers that produce an object by adding layer on layer first appeared 20 years ago, but lower cost machines in the 1990s using technology such as fused-deposition modelling and powder binding are available to researchers for between $US30,000 ($44,600) and $US50,000.
"We are on the verge of new systems that will cost less than $US10,000, require very little training and will be operated in a typical university office or lab," the forum says.
"At least one vendor plans to produce 3D printers for less than $US1000 within five years."
DNA computing is based on attempts to discover the computational power of the millions of molecules serving as natural supercomputers in the human body.
Researchers are trying to design computers based on this new model of computation, which offers extremely dense information storage, enormous parallel processing capacity and extraordinary energy efficiency.
DNA, or molecular, computing is still in its infancy, but the field is particularly suited to pharmaceutical and biomedical applications, including implants that release drugs according to conditions in the body.
"DNA is a plentiful, cheap source of processing power, and it can be produced cleanly," the forum says.
"The key advantage is being able to make computers far smaller that are able to hold much more data." Four hundred and fifty grams of DNA has the capacity to store more information than all the electronic computers ever built.
Task: Molecular computing. In 2003, Israeli scientists demonstrated a limited but functioning machine that uses DNA and enzymes, instead of silicon chips, for both data storage and processing.
The computer performed 330 trillion operations a second, more than 100,000 times the speed of the fastest PC - all from something that looked like a drop of water.
Reality check: The focus is on medical applications that require autonomous computation in a biomedical environment, rather than sheer computational power.
In 2006, Israeli scientists developed a molecular computer that uses enzymes to perform calculations from within the human body and monitor the release of drugs.
"In the future, there may be hybrid machines that use silicon for normal processing tasks and DNA co-processors for tasks they are more suited for," the forum says.
Qantum computing makes direct use of natural mechanical phenomena such as the spins of electrons or currents in superconducting circuits to store information as quantum bits (qubits) and perform computations.
"What makes quantum computing so powerful is that qubits can store much more information than binary bits, and qubits can be used together to perform calculations on that information in a parallel, not serial, fashion," the forum says.
"That is, it can process and evaluate many possible permutations of a problem simultaneously. This capability will enable extremely fast solutions to a wide range of problems currently considered intractable even on the fastest available supercomputers."
A key disruption will be in cryptography: quantum computing blows apart current encryption techniques, which are effective because it takes a long time - hundreds of years - to factor a large number and break an encryption scheme.
"A quantum computer can do that in seconds," it says. "When that day comes, everything that depends on encryption, from credit card transactions to email, will be wide open and unprotected until new security techniques are created."
Task: Recognising objects within an image. Humans can easily recognise and label a large range of interesting objects in an image - trees, cars, people and landmarks.
This task remains beyond the reach of any current automated image-labelling system. A technology that can recognise and name objects in an image will be able to solve many problems standing in the way of true machine intelligence, including learning and automated reasoning.
Reality check: D-Wave Systems is pioneering the use of quantum computers for commercial applications. It is working with Google to test a proof-of-concept structured-classification application using a machine running at 28 qubits.
"Our computers will be able to solve a wide range of high-value classification problems in areas such as image matching, machine learning, scheduling and complex database search," D-Wave's Jonathan Silverman says.
"Once the technology scales, it will be able to solve very hard problems that are beyond the scope of traditional systems."
D-Wave expects to have a 512-qubit system running by the end of this year, and a 1024-qubit system available commercially in 2009.
Optical developments may one day deliver computing at the speed of light.
Researchers are working on tapping light - already used in fibre-optics for high-speed telecommunications - to power computers.
Using photons instead of electrons is theoretically much faster, because photons travel much faster than an electric current.
Photons also have two unique properties that give them a boost: entanglement (instantaneous communication regardless of the distance) and superposition (the ability to process multiple options simultaneously).
Other benefits include lower cross-talk, better signal integrity at high frequencies and lower power requirements at higher transmission rates.
Task: Energy-efficient systems. IBM has announced a "green" prototype optical interconnect technology that could enable very large file transfers using extremely low power.
Sun Microsystems has launched a high-risk project to use light instead of wires between chips to speed data transfer.
If successful, the technique will eliminate data transfer bottlenecks in supercomputers using hundreds or thousands of processors.
NEC has announced an advance in optical connections between chips that would allow supercomputers to reach speeds of 10 petaflops, roughly 20 times faster than the world's fastest computer.
Reality check: Optical computing is still in its early stage in labs.
A key challenge is that it is very difficult to use light to control the state of a beam of light (either on or off), although researchers say they have developed a workable approach that uses only one photon, rather than large bursts of photons, to switch states.