Opticalcomputers.pdf

8/10/2019 Opticalcomputers.pdf

http://slidepdf.com/reader/full/opticalcomputerspdf 1/73

School of Electrical & Computer Engineering OPTICAL COMPUTERS

1

SCHOOL OF ELECTRICAL & COMPUTER

ENGINEERING

OPTICAL COMPUTERS




2

Contents

Overview of Optical computers1 Components of Optical computers. . . . . . . . . . . . . . . . . . . . . . 9

1.1 Hard Disk1.2 CPU1.3 Memory1.4 Cache Memory1.5 Main Memory1.6 Screen1.7 Power Supply

2 Need of Optical Computers . . . . . . . . . . . . . . . . . . . . . . . . . 15

3 Optical Components for Computing . . . . . . . . . . . . . . . . . . 203.1 VCSEL3.2 SLM3.3 WDM3.4 Optical Memory

4 Fibre Optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

4.1 Use of Fibre Optics in Computing4.2 Why use Fibre Optics

5 An Optical Computer Powered by Germanium Laser. . . . 40

6 Concept of Picosecond (By NASA) . . . . . . . . . . . . . . . . . . . . 44

7 Optical computer Bus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Application

Merits

Drawback

Some current research

Future Trends

References




3

PREFACE

An optical computer (also called a photonic computer) is a device that uses thephotons of visible light or infrared (IR) beams, rather than electric current, toperform digital computations. An electric current creates heat in computer systems.As the processing speed increases, so does the amount of electricity required; thisextra heat is extremely damaging to the hardware. Light, however, createsinsignificant amounts of heat, regardless of how much is used. Thus, thedevelopment of more powerful processing systems becomes possible.

An optical desktop computer could be capable of processing data up to 100,000times faster than current models because multiple operations can be performedsimultaneously.

Optical computing where the processing of electrical energy is replaced by lightquanta is very attractive for future technologies. The replacement of wires byoptical pathways is of special interest because light can cross without interferenceand thus, the complex wiring of modern computers may be appreciably simplified.Moreover, optical computers can operate at very high rates because there are not

the problems of electrical computers such as inductivities of wires and loading ofparasitic capacitors.




4

AN OVERVIEW OF OPTICAL COMPUTING

Computers have become an indispensable part of life. We needcomputers everywhere, be it for work, research or in any such field. As the use ofcomputers in our day-to-day life increases, the computing resources that we needalso go up. For companies like Google and Microsoft, harnessing the resources asand when they need it is not a problem. But when it comes to smaller enterprises,affordability becomes a huge factor. With the huge infrastructure come problemslike machines failure, hard drive crashes, software bugs, etc. This might be a bigheadache for such a community. Optical Computing offers a solution to thissituation.

An Optical Computer is a hypothetical device that uses visible light orinfrared beams, rather than electric current, to perform digital computations. Anelectric current flows at only about 10 percent of speed of light. Byapplying some of the advantages of visible and/or IR networks at the device andcomponent scale, a computer can be developed that can perform operations verymuch times faster than a conventional electronic computer.

Optical computing describes a new technological approach forconstructing computer’s processors and other components. Instead of the currentapproach of electrically transmitting data along tiny wires etched onto silicon.Optical computing employs a technology called silicon photonics that uses laserlight instead.

This use of optical lasers overcomes the constraints associated withheat dissipation in today’s components and allows much more information tobe stored and transmitted in the same amount of space. Optical computing meansperforming computations, operations, storage and transmission of data using light.Optical technology promises massive upgrades in the efficiency and speed ofcomputers, as well as significant shrinkage in their size and cost. An optical

desktop computer is capable of processing data up to 1,00,000 times faster thancurrent models.

An optical computer (also called a photonic computer) is a device thatuses the photons of visible light or infrared (IR) beams, rather than electric current,to perform digital computations. An electric current creates heat in computersystems. As the processing speed increases, so does the amount of electricityrequired; this extra heat is extremely damaging to the hardware.




5

For decades, silicon, with its talent for carrying electrons, has beenthe mainstay of computing. But for a variety of reasons (see "The Coming LightYears"), we're rapidly approaching the day when electrons will no longer cut it.Within 10 years, in fact, silicon will fall to the computer scientist's triple curse: "It'sbulky, it's slow, and it runs too hot." At this point, computers will need a newarchitecture, one that depends less on electrons and more on... well...what else?

Computer of 2010




6

Optics. With the assistance of award-winning firm frogdesign (the geniuses behind

the look of the early Apple and many of today's supercomputers and workstations),Forbes ASAP has designed and built (virtually, of course) the computer of 2010.

Whenever possible, our newly designed computer replaces stodgy oldelectrons with shiny, cool-running particles of light--photons. Electrons remain,doing everything they do best (switching), while photons do what they do best(traveling very, very fast). In other words, we've brought the speed and bandwidthof optical communications inside the computer itself. This mix is calledoptoelectronics, another buzzword we encourage you to start using immediately.

The result is a computer that is far more reliable, cheaper, and morecompact—the entire thing, believe it or not, is about the size of a Frisbee--than theall-electronic solution. But above all, optoelectronic computing is faster thanwhat's available today.How fast ? In a decade, we believe, you will be able to buyat your local computer shop the equivalent of today's supercomputers.

How likely is it that this computer will be built ? Some of its componentsare slightly pie-in-the-sky. But many others have already been developed or are

being developed by some of the best scientific minds in the country. Sooner orlater, and probably sooner, an optoelectronic computer will exist .

Okay, so we've built a revolutionary new optical computer just intime for 2010. What do we do with it now? Everything. Because it's small (aboutthe size of a Frisbee) and because it has the power of today's supercomputer, the2010 PC will become the repository of information covering every aspect of ourdaily life. Our computer, untethered and unfettered by wires and electrical outlets,becomes something of a key that unlocks the safety deposit box of our lives.

When we plug our 2010 PC into the wall of our home, our house willbecome smart, anticipating our every desire. At work, we'll plug it into our desk,which will become a gigantic interactive screen. When it communicates wirelesslywith a small mobile device, we'll have a personal digital assistant—on steroids.




7

Standard, electrical-based, computers rapidly approach fundamentallimitation. Alternative principles should be explored in order to keep computing

developments at the current pace or even faster. Optical computing has majorpotential in providing a solution through its use of photons to performcomputations instead of electrons. This workshop will be an opportunity tobring people together from optics and computer science who are interested inestablishing important principles and in developing optical computers. This willalso be an opportunity to meet with pioneering figures and to discuss the future ofoptical supercomputing.

Computers have enhanced human life to a great extent. The speed ofconventional computers is achieved by miniaturizing electronic components to a

very small micron-size scale so that those electrons need to travel only very shortdistances within a very short time. The goal of improving on computer speed hasresulted in the development of the Very Large Scale Integration (VLSI) technologywith smaller device dimensions and greater complexity. Last year, the smallest-todate dimensions of VLSI reached 0.08 m by researchers at Lucent Technology.Whereas VLSI technology has revolutionized the electronics industry andestablished the 20th century as the computer age, increasing usage of the Internetdemands better accommodation of a 10 to 15 percent per month growth rate.Additionally, our daily lives demand solutions to increasingly sophisticated andcomplex problems, which requires more speed and better performance ofcomputers.

For these reasons, it is unfortunate that VLSI technology is approaching itsfundamental limits in the sub-micron miniaturization process. It is now possible tofit up to 300 million transistors on a single silicon chip. It is also estimated that thenumber of transistor switches that can be put onto a chip doubles every 18 months.Further miniaturization of lithography introduces several problems such asdielectric breakdown, hot carriers, and short channel effects. All of these 2 factorscombine to seriously degrade device reliability. Even if developing technologysucceeded in temporarily overcoming these physical problems, we will continue to

face them as long as increasing demands for higher integration continues.Therefore, a dramatic solution to the problem is needed, and unless we gear ourthoughts toward a totally different pathway, we will not be able to further improveour computer performance for the future.

Optical interconnections and optical integrated circuits will provide a wayout of these limitations to computational speed and complexity inherent inconventional electronics. Optical computers will use photons traveling on opticalfibers or thin films instead of electrons to perform the appropriate functions. In theoptical computer of the future, electronic circuits and wires will be replaced by a




8

few optical fibers and films, making the systems more efficient with nointerference, more cost effective, lighter and more compact. Optical components

would not need to have insulators as those needed between electronic componentsbecause they donot experience cross talk. Indeed, multiple frequencies (or differentcolors) of light can travel through optical components without interfacing witheach others, allowing photonic devices to process multiple streams of datasimultaneously.

SECURITY

The PC will be protected from theft, thanks to an advanced biometric scannerthat can recognize your fingerprint.INTERFACE

You'll communicate with the PC primarily with your voice, putting it truly atyour beck and call.

The Desktop as Desk Top

In 2010, a "desktop" will be a desk top...in other words, by plugging ourcomputer into an office desk, its top becomes a gigantic computer screen--aninteractive photonic display. You won't need a keyboard because files can be

opened and closed simply by touching and dragging with your finger. And forthose throwbacks who must have a keyboard, we've supplied that as well.A virtual keyboard can be momentarily created on the table top, only to

disappear when no longer needed. Now you see it, now you don't.Your Digital Butler

What do we do with our 2010 computer when we arrive home after a longday's work? The computer becomes the operating system for our house, and ourhouse, in turn, knows our habits and responds to our needs. ("Brew coffee at 7,play Beethoven the moment the front door opens, and tell me when I'm low on

milk.")Your Home

The PC of 2010 plugs into your home so your house becomes a smartoperating system.




9

Optical Computing Technology




10

An optical computer (also called a photonic computer) is a device that usesthe photons of visible light or infrared (IR) beams, rather than electric current, to

perform digital computations. An electric current creates heat in computer systems.As the processing speed increases, so does the amount of electricity required; thisextra heat is extremely damaging to the hardware. Light, however, createsinsignificant amounts of heat, regardless of how much is used. Thus, thedevelopment of more powerful processing systems becomes possible. By applyingsome of the advantages of visible and/or IR networks at the device and componentscale, a computer might someday be developed that can perform operations 10 ormore times faster than a conventional electronic computer.

Visible-light and IR beams, unlike electric currents, pass through each other

without interacting. Several (or many) laser beams can be shone so their pathsintersect, but there is no interference among the beams, even when they areconfined essentially to two dimensions. Electric currents must be guided aroundeach other, and this makes three-dimensional wiring necessary. Thus, an opticalcomputer, besides being much faster than an electronic one, might also be smaller.

Most research projects focus on replacing current computer componentswith optical equivalents, resulting in an photonic digital computer systemprocessing binary data. This approach appears to offer the best short-term

prospects for commercial optical computing, since optical components could beintegrated into traditional computers to produce an optical/electronic hybrid. Otherresearch projects take a non-traditional approach, attempting to develop entirelynew methods of computing that are not physically possible with electronics.

Optical computing where the processing of electrical energy is replaced bylight quanta is very attractive for future technologies . The replacement ofwires by optical pathways is of special interest because light can cross withoutinterference and thus, the complex wiring of modern computers may beappreciably simplified. Moreover, optical computers can operate at very high rates

because there are not the problems of electrical computers such as inductivities ofwires and loading of parasitic capacitors. Chemical structures are required for thehandling of light and this has to be done by suitable chromophores. Organicmaterials are preferred because of their chemical variability and uncriticalrecycling for mass production. There are mainly three obstacles for thedevelopment of optical computers: firstly the preservation of the optical energy,secondly the low light-fastness of many active optical components and thirdlythe comparably long wavelengths of light of about 0.5 m. The former twoproblems can be solved by the application of highly light-fast fluorescence dyes




11

where the fluorescence quantum yield is a measure of the preservation of light-energy; light fast fluorescent dyes with 100% fluorescence quantum yield are

known .The third problem sets a lower limit to the size of conventional opticalcomponents and hinders the construction of an optical computer on a molecularscale. However, the development of molecular optics would reduce the size of suchcomponents by a factor of 500.

The limitation of resolution by the wavelengths of light may be overcomeby the transport of the energy of light instead of the emission and absorption oflight quanta. This corresponds to the use of the alternating current (50 Hz) with aproblematic wavelength of some 6000 km where the electrical energy is handledon a human scale or even lower.

In analogy to such a transport of electrical energy an energy transferbetween chromophores can replace the absorption and emission of light quanta inoptical signal processing components. The transfer will proceed rapidly if thedistance between the two chromophores lies within the F¨orster radius, that meansbetween 2 and 3 nm for most combinations of similarly absorbing chromophores.On the other hand, this F¨orster radius would be the natural lower limit for the sizeof complex arrangements of switching components for handling energy transferbecause going below this limit would spread energy over many chromophoreswithout control; a solution of this limiting problem would be the prerequisite for

the development of optical computers with very high densities of integration.




12

COMPONENTS OF OPTICAL COMPUTER

• Hard Disk• CPU• Memory• Cache Memory• Main Memory• Screen• Power Supply

(1) HARD DISK (STORES PROGRAMS AND FILES)

To build our 2010 computer (see previous page) we first need tobuild the hard disk. The disk will be holographic and willsomewhat resemble a CD-ROM or DVD. That is, it will be aspinning, transparent plastic platter with a writing laser on one sideand reading laser on the other, and it will hold an astoundingterabyte (1 trillion bytes) of data, just a tad more than we get today--1,000 times more, to be exact. With such capacity, you'll be ableto store every ounce of information about your life. But beware.If your computer is stolen or destroyed , you might actually startwondering who you are.

WHERE ARE WE?

A holographic disk might be the easiestcomponent here to build since it exists in the lab today.

WHO'S WORKING ON IT?Stanford University, Lucent Technologies, and

cutting-edge Silicon Valley optics company Siros Technologies.

TIME OF COMPLETION? 2005, for a commercial product.




13

(2) THE CENTRAL PROCESSING UNIT (CPU)

Our 2010 CPU will operate on the same principle as today's PCs.But instead of electronic microprocessors providing the brains andbrawn, our future CPU will have optoelectronic integrated circuits(chips that use silicon to switch but optics to communicate). Thiswill give us huge increases in speed and efficiency. Why? Becausethe CPU of today spends far too much time waiting around for datato process. Instantaneous on-chip optical communication, andmemory running as fast as the processor, will guarantee a

continuous stream of data processing within the CPU. Withcommunication between components no longer bottlenecked byelectronic transmission, we can probably push the clock rate to 100gigahertz.

Our universal appliance of tomorrow also has ahexagonal optoelectronic processor surrounded by a ring of fastcache, so that data for any part of the chip can be fetched from theclosest part of the cache. The result will be computer performance--or, at any rate, delivery of computational results--comparable totoday'ssupercomputers .

WHERE ARE WE?Optoelectronic integrated circuits do exist

today, on a small scale and for specialized purposes. Getting fromthe current state of the art to a complete and superfastoptoelectronic CPU will require tremendous effort and theaccumulation of an entirely new body of intellectual property.

WHO'S WORKING ON IT?

Scientific-Atlanta, Lucent, and Nortel. Advanced workin optical interconnection is now being done at Stanford. Intel,through its purchase of Danish optoelectronics company GIGA,intends to have the fast track outofthegate.

TIME OF COMPLETION? 2010,If we're really lucky.




14

(3) MEMORY(RAM)

When we stir optical communication into the old-fashionedelectronic computer, some of the greatest potential gains willinvolve your computer's short-term memory. In the long-gone days(1980) of the 80286, computers enjoyed a design advantage thatwe've never had since. The memory bus speed--that is, the speed atwhich data flowed between CPU and memory--was the same as theCPU's clock rate, or how fast it operates . (Of course, they wereboth 8 megahertz , but we said this was a long time ago.) Datareached the CPU as fast as the chip could process it, whichkept the CPU from waiting around being bored.We've never reached that pinnacle again, and since then, thesituation has gotten steadily worse. A reasonably fast computertoday has a CPU clock of 600 megahertz and a memory bus speedof 133 megahertz. Despite various clever technical feats, the CPUstill spends half to two-thirds of its time just waiting around fordata from memory.Optoelectronics will knock this problem out of the park. With aproperly designed optical memory bus, speed of fetch frommemory can once again equal CPU clock rate.

Of course, this also will require that processing in RAM be veryquick, so we'll need a faster RAM architecture, which luckily is--orwill be--available. A large cache (see below) made of superfast,nonvolatile magnetic RAM will hold information that the CPUneeds quickly and repeatedly. It will be backed up by a muchlarger area of holographic (pure optical) main RAM that will holdprograms, files, images, etc., while you work with them.

(4) FAST MEMORY (CACHE)

To build our new fast cache, we'll need to get rid of theinefficiencies of today's product, which requires the computer toconstantly refresh it, just like short-term memory in humans needsto be constantly refreshed or it's forgotten. The inefficiencies incache are so bad, in fact, that once you know the speed of yourcache you can assume that its real-world performance will be abouta third of that--the missing two-thirds being sacrificed to refreshcycles.






16

TIME OF COMPLETION? 2009,or maybe a tad earlier.

(6) POWER SUPPLY

One of the biggest advantages of photonic circuitry is an extremelylow power requirement. A long, sticklike lithium battery, bent intoa doughnut and installed in the periphery of the computer, will runit for a couple of weeks. But fresh power is as close as the chargingcradle on the nearest wall, which resembles the one for today's

cordless or cellular phones.WHERE ARE WE? Pretty close. We've come a long way inbattery development in the past few years.

WHO'S WORKING ON IT? Hewlett-Packard.

TIME OF COMPLETION? 2007.

(7) THE SCREEN

Size does matter in our 2010 computer screen. It will either be verylarge, literally the desk top of your desktop, or very small, amonocle you hold up to your eye. For the bigger version, ourcomputer screen will depend on some kind of photonically excitedliquid crystal, with power requirements significantly lower than

today's monitors. Colors will be vivid and images precise (thinkplasma displays). In fact, today's concept of "resolution" will belargely obsolete. Get ready for pay-per-view Webcasts.

WHERE ARE WE? This design, if fully realized, depends on atechnology that doesn't exist today. Optical excitement of a liquidcrystal is the stuff of research papers. More likely is that ourcomputer will end up with a less ambitious display, one like our




17

current PCs possess but much, much better. We've got 10 fruitfulyears to develop it, after all.

WHO'S WORKING ON IT? Sharp Electronics, a world leader incolor LCD technology, which is also investing heavily inoptoelectronics. Sony, Toshiba, and IBM are the current leaders inflat-panel displays.

TIME OF COMPLETION? 2010, if we're lucky.




18

NEED OF OPTICAL COMPUTERS

Optics has been used in computing for a number of years but the mainemphasis has been and continues to be to link portions of computers, forcommunications, or more intrinsically in devices that have some opticalapplication or component (optical pattern recognition, etc). Optical digitalcomputers are still some years away, however a number of devices that canultimately lead to real optical computers have already been manufactured,including optical logic gates, optical switches, optical interconnections, and opticalmemory. The most likely near-term optical computer will really be a hybridcomposed of traditional architectural design along with some portions that canperform some functional operations in optical mode.

With today’s growing dependence on computing technology, the need forhigh performance computers (HPC) has significantly increased. Many performanceimprovements in conventional computers are achieved by miniaturizing electroniccomponents to very small micron-size scale so that electrons need to travel onlyshort distances within a very short time. This approach relies on the steadilyshrinking trace size on microchips (i.e., the size of elements that can be ‘drawn’onto each chip). This has resulted in the development of Very Large ScaleIntegration (VLSI) technology with smaller device dimensions and greater

complexity. The smallest dimensions of VLSI nowadays are about 0.08 mm.Despite the incredible progress in the development and refinement of the basictechnologies over the past decade, there is growing concern that these technologiesmay not be capable of solving the computing problems of even the currentmillennium.

Technologies lead to breakthroughs in engineering and manufacturing in awide range of industries. With the help of virtual product design and development,costs can be reduced; hence looking for improved computing capabilities isdesirable. Optical computing includes the optical calculation of transforms and

optical pattern matching. Emerging technologies also make the optical storage ofdata a reality.

The speed of computers was achieved by miniaturizing electroniccomponents to a very small micron-size scale, but they are limited not only by thespeed of electrons in matter (Einstein’s principle that signals cannot propagatefaster than the speed of light) but also by the increasing density of interconnectionsnecessary to link the electronic gates on microchips. The optical computer comesas a solution of miniaturization problem. In an optical computer, electrons are




19

replaced by photons, the subatomic bits of electromagnetic radiation that make uplight.

Optics, which is the science of light, is already used in computing, mostoften in the fiber-optic glass cables that currently transmit data on communicationnetworks much faster than via traditional copper wires. Thus, optical signals mightbe the ticket for the fastest supercomputers ever. Compared to light, electronicsignals in chips travel at snail speed. Moreover, there is no such thing as a shortcircuit with light, so beams could cross with no problem after being redirected bypinpoint-size mirrors in a switchboard. In a pursuit to probe into cutting-edgeresearch areas, optical technology (optoelectronic, photonic devices) is one of themost promising, and may eventually lead to new computing applications as aconsequence of faster processor speeds, as well as better connectivity and higherbandwidth. The pressing need for optical technology stems from the fact thattoday’s computers are limited by the time response of electronic circuits. A solidtransmission medium limits both the speed and volume of signals, as well asbuilding up heat that damages components. For example, a one-foot length of wireproduces approximately one nanosecond (billionth of a second) of time delay.Extreme miniaturization of tiny electronic com- Optical computing includes theoptical calculation of transforms and optical pattern matching. Emergingtechnologies also make the optical storage of data.

These and other obstacles have led scientists to seek answers in light itself.Light does not have the time response limitations of electronics, does not needinsulators, and can even send dozens or hundreds of photon signal streamssimultaneously using different color frequencies. Those are immune toelectromagnetic interference, and free from electrical short circuits. They havelow-loss transmission and provide large bandwidth; i.e. multiplexing capability,capable of communicating several channels in parallel without interference. Theyare capable of propagating signals within the same or adjacent fibers withessentially no interference or cross talk. They are compact, lightweight, andinexpensive to manufacture, as well as more facile with stored information thanmagnetic materials. By replacing electrons and wires with photons, fiber optics,crystals, thin films and mirrors, researchers are hoping to build a new generation ofcomputers that work 100 million times faster than today’s machines.

The fundamental issues associated with optical computing, its advantagesover conventional (electronics-based) computing, current applications of optics incomputers are discussed in this part. In the second part of this article the problemsthat remain to be overcome and current research will be discussed.




20

Optical computing was a hot research area in the 1980s. But the worktapered off because of materials limitations that seemed to prevent optochips from

getting small enough and cheap enough to be more than laboratory curiosities.Now, optical computers are back with advances in self-assembled conductingorganic polymers that promise super-tiny all-optical chips.

[1]. Advances in optical storage device have generated thepromise of efficient, compact and large-scale storage devices

[2]. Another advantage of optical methods over electronicones for computing is that parallel data processing can frequently be done muchmore easily and less expensively in optics than in electronics

[3]. Light does not have the time response limitations ofelectronics, does not need insulators, and can even send dozens or hundreds ofphoton signal streams simultaneously usingdifferent color frequencies. Parallelism, the capability to execute more than oneoperation simultaneously, is now common in electronic computer architectures.But, most electronic computers still execute instructionssequentially; parallelism with electronics remains sparsely used. Its firstwidespread appearance was in Cray supercomputers in the early 1980’s when two

processors were used in conjunction with one shared memory. Today, largesupercomputers may utilize thousands of processors but communication overheadfrequently results in reduced overall efficiency

[4]. On the other hand for some applications in input-output(I/O), such as image processing, by using a simple optical design an array of pixelscan be transferred simultaneously in parallel from one point to another. Opticaltechnology promises massive upgrades in the efficiency and speed of computers,as well as significant shrinkage in their size and cost. An optical desktop computercould be capable of processing data up to 100,000 times faster than current models

because multiple operations can be performed simultaneously. Other advantages ofoptics include low manufacturing costs, immunity to electromagnetic interference,a tolerance for lowloss transmissions, freedom from short electrical circuits and thecapability to supply large bandwidth and propagate signals within the same oradjacent fibers without interference.

One oversimplified example may help to appreciatethe difference between optical and electronic parallelism. Consider an imagingsystem with 1000 1000 independent points per mm2 in the object plane whichare connected optically by a lens to a corresponding number of points per mm2 in




21

the image plane; the lens effectively performs an FFT of the image plane in realtime. For this to be accomplished electrically, a million operations are required.

Parallelism, when associated with fast switching speeds, would result in staggeringcomputational speeds. Assume, for example, there are only 100 million gates on achip, much less than what was mentioned earlier (optical integration is still in itsinfancy compared to electronics). Further, conservatively assume that Opticaltechnology promises massive upgrades in the efficiency and speed of computers,as well as significant shrinkage in their size and cost.

An optical desktop computer could be capable ofprocessing data up to 100,000 times faster than current models because multipleoperations can be performed simultaneously. Each gate operates with a switchingtime of only 1 nanosecond(organic optical switches can switch at sub-picosecondrates compared to maximum picosecond switching times for electronic switching).Such a system could perform more than 1017 bit operations per second. Comparethis to the gigabits (109) or terabits (1012) per second rates which electronics areeither currently limited to, or hoping to achieve. In other words, a computation thatmight require one hundred thousand hours (more than 11 years) of a conventionalcomputer time could require less than one hour by an optical one. But building anoptical computer will not be easy. A major challenge is finding materials that canbe mass produced yet consume little power; for this reason, optical computers maynot hit the consumer market for 10 to 15 years.

Another of the typical problems optical computershave faced is that the digital optical devices have practical limits of eight to elevenbits of accuracy in basic operations due to, e.g., intensity fluctuations. Recentresearch has shown ways around this difficulty. Thus, for example, digitalpartitioning algorithms, that can break matrix-vector products into lower-accuracysub-products, working in tandem with error-correction codes, can substantiallyimprove the accuracy of optical computing operations. Nevertheless, manyproblems in developing appropriate materials and devices must be overcomebefore digital optical computers will be in widespread commercial use. In the nearterm, at least, optical computers will most likely be hybrid optical/electronicsystems that use electronic circuits to preprocess input data for computation and topost-process output data for error correction before outputting the results.

The promise of all-optical computing remains highlyattractive, however, and the goal of developing optical computers continues to be aworthy one. Nevertheless, many scientists feel that an all-optical computer will notbe the computer of the future; instead optoelectronic computers will rule where theadvantages of both electronics and optics will be used. Optical computing can alsobe linked intrinsically to quantum computing. Each photon is a quantum of a wave




22

function describing the whole function. It is now possible to control atoms bytrapping single photons in small, superconducting cavities

[5]. So photon quantum computing could become a futurepossibility.

The pressing need for optical technology stems from the fact thattoday’s computers are limited by the time response of electronic circuits. A solidtransmission medium limits both the speed and volume of signals, as well asbuilding up heat that damages components. One of the theoretical limits on howfast a computer can function is given by Einstein’s principle that signal cannotpropagate faster than speed of light. So to make computers faster, their components

must be smaller and there by decrease the distance between them. This has resultedin the development of very large scale integration (VLSI) technology, with smallerdevice dimensions and greater complexity. The smallest dimensions of VLSInowadays are about 0.08mm. Despite the incredible progress in the developmentand refinement of the basic technologies over the past decade, there is growingconcern that these technologies may not be capable of solving the computingproblems of even the current millennium.

The speed of computers was achieved by miniaturizing electroniccomponents to a very small micron-size scale, but they are limited not only by thespeed of electrons in matter but also by the increasing density of interconnectionsnecessary to link the electronic gates on microchips. The optical computer comesas a solution of miniaturization problem.Optical data processing can performseveral operations in parallel much faster and easier than electrons. Thisparallelism helps in staggering computational power. For example a calculationthat takes a conventional electronic computer more than 11 years to completecould be performed by an optical computer in a single hour. Any way we canrealize that in an optical computer, electrons are replaced by photons, thesubatomic bits of electromagnetic radiation that make up light.




23

Optical Components for Computing

The major breakthroughs on optical computing have been centered on thedevelopment of micro-optic devices for data input. Conventional lasers are knownas ‘edge emitters’ because their laser light comes out from the edges. Also, theirlaser cavities run horizontally along their length. A vertical cavity surface emittinglaser (VCSEL – pronounced ‘vixel’), however, gives out laser light from itssurface and has a laser cavity that is vertical; hence the name. VCSEL is asemiconductor vertical cavity surface emitting microlaser diode that emits light ina cylindrical beam vertically from the surface of a fabricated wafer, and offerssignificant advantages when compared to the edge-emitting lasers currently used in

the majority of fiber optic communications devices. They emit at 850 nm and haverather low thresholds (typically a few mA). They are very fast and can give mW ofcoupled power into a 50 micron core fiber and are extremely radiation hard.VCSELS can be tested at the wafer level (as opposed to edge emitting lasers whichhave to be cut and cleaved before they can be tested) and hence are relativelycheap. In fact, VCSELs can be fabricated efficiently on a 3-inch diameter wafer. Aschematic of VCSEL is shown in Figure 1.

Fig.- 1




24

The principles involved in the operation of a VCSEL are

very similar to those of regular lasers. As shown in Figure , there are two specialsemiconductor materials sandwiching an active layer where all the action takesplace. But rather than reflective ends, in a VCSEL there are several layers ofpartially reflective mirrors above and below the active layer. Layers ofsemiconductor with differing compositions create these mirrors, and each mirrorreflects a narrow range of wavelengths back into the cavity in order to cause lightemission at just one wavelength.

Spatial light modulators (SLMs) play an important role inseveral technical areas where the control of light on a pixel-bypixel basis is a keyelement, such as optical processing, for inputting information on light beams, anddisplays. For display purposes the desire is to have as many pixels as possible in assmall and cheap a device as possible. For such purposes designing silicon chips foruse as spatial light modulators has been effective. The basic idea is to have a set ofmemory cells laid out on a regular grid. These cells are electrically connected tometal mirrors, such that the voltage on the mirror depends on the value stored inthe memory cell.

A layer of optically active liquid crystal is sandwiched

between this array of mirrors and a piece of glass with a conductive coating. Thevoltage between individual mirrors and the front electrode affects the opticalactivity of the liquid crystal in that neighborhood. Hence by being able toindividually program the memory locations one can set up a pattern of opticalactivity in the liquid crystal layer.

Figure 2(a) shows a reflective 256x256 pixel device basedon SRAM technology. Several technologies have contributed to the developmentof SLMs. These include micro-electro-mechanical devices, such as, acousto-opticmodulators (AOMs), and pixelated electrooptical devices, such as liquid-crystalmodulators (LCMs).

Figure 2(b) shows a simple AOM operation in deflectinglight beam direction. Encompassed within these categories are amplitudeonly,phase-only, or amplitude-phase modulators. Broadly speaking, an optical computeris a computer in which light is used somewhere. This can means fiber opticalconnections between electronic components, free space connections, or one inwhich light functions as a mechanism for storage of data, logic or arithmetic.Instead of electrons in silicon integrated circuits, the digital optical computers willbe based on photons. Smart pixels, the union of optics and electronics, both




25

expands the capabilities of electronic systems and enables optical systems withhigh levels of electronic signal processing. Thus, smart pixel systems add value to

electronics through optical input/output and interconnection, and value is added tooptical systems through electronic enhancements which include gain, feedbackcontrol, and image processing and compression.

Fig.- 2 (a)




26

Fig.- 2 (b)

Smart pixel technology is a relatively new approach tointegrating electronic circuitry and optoelectronic devices in a commonframework. The purpose is to leverage the advantages of each individualtechnology and provide improved performance for specific applications. Here, theelectronic circuitry provides complex functionality and programmability while theoptoelectronic devices provide high-speed switching and compatibility withexisting optical media. Arrays of these smart pixels leverage the parallelism ofoptics for interconnections as well as computation. A smart pixel device, a light




27

emitting diode (LED) under the control of a field-effect transistor (FET), can nowbe made entirely out of organic materials on the same substrate for the first time. In

general, the benefit of organic over conventional semiconductor electronics is thatthey should (when mass-production techniques take over) lead to cheaper, lighter,circuitry that can be printed rather than etched. Scientists at Bell Labs have made300-micron-wide pixels using polymer FETs and LEDs made from a sandwich oforganic materials, one of which allows electrons to flow, another which acts ashighway for holes (the absence of electrons); light is produced when electrons andholes meet. The pixels are quite potent, with a brightness of about 2300candela/m2, compared to a figure of 100 for present flat-panel displays . ACambridge University group has also made an all-organic device, not as bright asthe Bell Labs version, but easier to make on a large scale .

VCSEL (VERTICAL CAVITY SURFACE EMITTING LASER)

VCSEL (pronounced ‘vixel’) is a semiconductor verticalcavity surface emitting laser diode that emits light in a cylindrical beam verticallyfrom the surface of a fabricated wafer, and offers significant advantages whencompared to the edge-emitting lasers currently used in the majority of fiber opticcommunications devices. The principle involved in the operation of a VCSEL isvery similar to those of regular lasers.

There are two special semiconductor materials sandwichingan active layer where all the action takes place. But rather than reflective ends, in aVCSEL there are several layers of partially reflective mirrors above and belowthe active layer. Layers of semiconductors with differing compositions createthese mirrors, and each mirror reflects a narrow range of wavelengths back in tothe cavity in order to cause light emission at just one wavelength. 4

OPTICAL INTERCONNECTION OF CIRCUIT BOARDS USING VCSEL AND

PHOTODIODE

VCSEL convert the electrical signal to optical signal whenthe light beams are passed through a pair of lenses and micromirrors. Micromirrorsare used to direct the light beams and this light rays is passed through a polymerwaveguide which serves as the path for transmitting data instead of copper wires inelectronic computers. Then these optical beams are again passed through a pair oflenses and sent to a photodiode. This photodiode convert the optical signal back tothe electrical signal.5




28

SLM (SPATIAL LIGHT MODULATORS)

SLM play an important role in several technical areas where the control of

light on a pixel-by-pixel basis is a key element, such as optical processing anddisplays.

SLM FOR DISPLAY PURPOSES

For display purposes the desire is to have as many pixels as possible in assmall and cheap a device as possible. For such purposes designing silicon chips foruse as spatial light modulators has been effective. The basic idea is to have a set ofmemory cells laid out on a regular grid. These cells are electrically connected tometal mirrors, such that the voltage on the mirror depends on the value stored inthe memory cell. A layer of optically active liquid crystal is sandwiched between

this array of mirrors and a piece of glass with a conductive coating. The voltagebetween individual mirrors and the front electrode affects the optical activity ofliquid crystal in that neighborhood. Hence by being able to individually programthe memory locations one can set up a pattern of optical activity in the liquidcrystal layer.6 SMART PIXEL TECHNOLOGY

Smart pixel technology is a relatively new approach to integrating electroniccircuitry and optoelectronic devices in a common framework. The purpose is toleverage the advantages of each individual technology and provide improved

performance for specific applications. Here, the electronic circuitry providescomplex functionality and programmability while the optoelectronic devicesprovide high-speed switching and compatibility with existing optical media.Arrays of these smart pixels leverage the parallelism of optics for interconnectionsas well as computation. A smart pixel device, a light emitting diode under thecontrol of a field effect transistor can now be made entirely out of organicmaterials on the same substrate for the first time. In general, the benefit of organicover conventional semiconductor electronics is that they should lead to cheaper,lighter, circuitry that can be printed rather than etched.

WDM (WAVELENGTH DIVISION MULTIPLEXING)

Wavelength division multiplexing is a method of sending many differentwavelengths down the same optical fiber. Using this technology, modern networksin which individual lasers can transmit at 10 gigabits per second through the samefiber at the same time. WDM can transmit up to 32 wavelengths through a singlefiber, but cannot meet the bandwidth requirements of the present daycommunication systems. So nowadays DWDM (Dense wavelength division




29

multiplexing) is used. This can transmit up to 1000 wavelengths through a singlefiber. That is by using this we can improve the bandwidth efficiency.8

ROLE OF NLO IN OPTICAL COMPUTING

The role of nonlinear materials in optical computing has become extremelysignificant. Non-linear materials are those, which interact with light and modulateits properties. Several of the optical components require efficient nonlinearmaterials for their operations. What in fact restrains the widespread use of alloptical devices is the in efficiency of currently available nonlinear materials, whichrequire large amount of energy for responding or switching. Organic materialshave many features that make them desirable for use in optical devices such as

1. High nonlinearities2. Flexibility of molecular design3. Damage resistance to optical radiations

Some organic materials belonging to the classes of phthalocyanines andpolydiacetylenes are promising for optical thin films and wave guides. Thesecompounds exhibit strong electronic transitions in the visible region and have highchemical and thermal stability up to 400 degree Celsius. Polydiacetylenes areamong the most widely investigated class of polymers for nonlinear opticalapplications. Their subpicosecond time response to laser signals makes themcandidates for high-speed optoelectronics and information processing.

To make thin polymer film for electro-optic applications, NASA scientistsdissolve a monomer (the building block of a polymer) in an organic solvent. Thissolution is then put into a growth cell with a quartz window, shining a laserthrough the quartz can cause the polymer to deposit in specific pattern.

The field of optical computing is considered to be the most multidisciplinaryfield and requires for its success collaborative efforts of many disciplines, ranging

from device and optical engineers to computer architects, chemists, materialscientists, and optical physicists. On the materials side, the role of nonlinearmaterials in optical computing has become extremely significant. Nonlinearmaterials are those, which interact with light and modulate its properties. Forexample, such materials can change the color of light from being unseen in theinfrared region of the color spectrum to a green color where it is easily seen in thevisible region of the spectrum. Several of the optical computer components requireefficient nonlinear materials for their operation. What in fact restrains the wide-spread use of all optical devices is the inefficiency of currently available nonlinear




30

optical materials, which require large amounts of energy for responding orswitching. In spite of new developments in materials, presented in the literature

daily, a great deal of research by chemists and material scientists is still required toenable better and more efficient optical materials. Although organic materials havemany features that make them desirable for use in optical devices, such as highnonlinearities, Flexibility of molecular design, and damage resistance to opticalradiation, their use in devices has been hindered by processing difficulties forcrystals and thin films. Our focus is on a couple of these materials, which haveundergone some investigation in the NASA/MSFC laboratories, and were alsoprocessed in space either by the MSFC group, or others. These materials belong tothe classes of phthalocyanines and polydiacetylenes. These classes of organiccompounds are promising for optical thin films and waveguides. Phthalocyaninesare large ring-structured porophyrins for which large and ultrafast nonlinearitieshave been observed. These compounds exhibit strong electronic transitions in thevisible region and have high chemical and thermal stability up to 400°C. Wemeasured the third order susceptibility of phthalocyanine, which is a measure of itsnonlinear efficiency to be more than a million times larger than that of the standardmaterial, carbon disulfide. This class of materials has good potential forcommercial device applications, and has been used as a photosensitive organicmaterial, and for photovoltiac, photoconductive, and photoelectrochemicalapplications.

ADVANCES IN PHOTONIC SWITCHES

Logic gates are the building blocks of any digital system. An optical logicgate is a switch that controls one light beam by another; it is ON when the devicetransmits light and it is OFF when it blocks the light.To demonstrate the AND gatein the phthalocyanine film, two focused collinear laser beams are wave guidedthrough a thin film of phthalocyanine. Nanosecond green pulsed Nd:YAG laser

was used together with a red continuous wave (cw) He-Ne beam. At the output anarrow band filter was set to block the green beam and allow only the He-Nebeam. Then the transmitted beam was detected on an oscilloscope. It was foundthat the transmitted He-Ne cw beam was pulsating with a nanosecond duration andin synchronous with the input Nd:YAG nanosecond pulse. This demonstrated thecharacteristic table of an AND logic gate.




31

OPTICAL NAND GATE

In an optical NAND gate the phthalocyanine film is replaced by a hollowfiber filled with polydiacetylene. Nd:YAG green picosecond laser pulse was sentcollinearly with red cw He-Ne laser onto one end of the fiber. At the other end ofthe fiber a lens was focusing the output on to the narrow slit of a monochrometerwith its grating set for the red He-Ne laser. When both He-Ne laser and Nd:YAGlaser are present there will be no output at the oscilloscope. If either one or noneof the laser beams are present we get the output at the oscilloscope showingNAND function.11

OPTICAL MEMORY

In optical computing two types of memory are discussed. One consists ofarrays of one-bit-store elements and other is mass storage, which is implementedby optical disks or by holographic storage systems. This type of memory promisesvery high capacity and storage density. The primary benefits offered byholographic optical data storage over current storage technologies includesignificantly higher storage capacities and faster read-out rates. This research isexpected to lead to compact, high capacity, rapid-and random-access, and low

power and low cost data storage devices necessary for future intelligent spacecraft.The SLMs are used in optical data storage applications. These devices are used towrite data into the optical storage medium at high speed. More conventionalapproaches to holographic storage use ion doped lithium niobate crystals to storepages of data.

For audio recordings ,a 150MBminidisk with a 2.5- in diameter has beendeveloped that uses special compression to shrink a standard CD’s640-MB storagecapacity onto the smaller polymer substrate. It is rewritable and uses magnetic fieldmodulation on optical material. The mini disc uses one of the two methods to writeinformation on to an optical disk. With the mini disk a magnetic field placedbehind the optical disk is modulated while the intensity of the writing laser is heldconstant. By switching the polarity of the magnetic field while the laser creates astate of flux in the optical material digital data can be recorded on a single layer.As with all optical storage media a read laser retrieves the data.




32

Fiber Optics: -

Definition: A basic fiber optic system consists of a transmitting device, which generatesthe light signal; an optical fiber cable, which carries the light; and a receiver, whichaccepts the light signal transmitted. The fiber itself is passive and does not containany active, generative properties.

History: Many individuals throughout the history of the world have recognized the

value of using light to to communicate. Early defense warning systems were set upon the Great wall of China with signal fires to warn of enemies approaching. In the

late 1700's the "optical telegraph" was invented by a French engineer namedClaude Chappe which, similar to the fire signals, used semaphores mounted ontowers, where human operators relayed messages from one tower to the next. In1870, John Tyndal demonstrated the principle of total internal reflection by shininga light into a water tank, poking a hole in the side, and as the water ran out in anarc, the light took the shape and followed the water down. Ten years later,Alexander Graham Bell patented an optical telephone system "Photophone" whichhe imagined sound waves carried by light. It wasn't until many years later throughnumerous advances in thinking and technical discovery's that Tyndal's and Bell's

basic concepts came together to what we now know as fiber optics. Through theinvention of the continuouswave helium-neon laser and enhancements to opticalfiber, researchers Dr. Robert Maurer, Peter Schultz, and Donald Keck of CorningIncorporated lead the way in development of Silica manufactured fiber optics andin 1970 were successful in manufacturing 20dB/km, cable that was tested and usedsuccessfully in Britain. Today optical fiber is manufactured at .25dB/km, which isan indicator of the purity of the silica and how much loss of light occurs overdistance.

Technical Info:

Optical fiber for telecommunications is made up of three parts including thecore, cladding & coating. The core is the central part of the fiber which transmitsthe light. The cladding surrounds the core and keeps the light in the core because itis made of material with a lower index of refraction. The core and cladding areinseparable because they are made up of a single piece of glass silica, treated tocreate the differences needed in refraction. Finally, a coating generally made ofUV protective acrylate is put on a fiber during the draw process to protect it.




33

Fiber optic systems can carry both analog and digital signals over

light waves. A system consists of a signal generator, (e.g. computer, video, audio)an encoder, a fiber optic cable, and a decoder, and a receiving device (e.g. tv,computer network, etc.) Fiber optics have many advantages over copper cable.They have become a desired standard for networking backbones and hubs becauseof the advantages they have over copper to achieve the speed and bandwidthcapacity. A single fiber optic cable can transmit the same amount of data asapproximately 600 pair traditional copper telecommunications wire, an transmitdata further with less boosting of the signal, it is not effected by electricalanomalies such as lightning, it is small, light weight and easy to install.




34

Year2000:

With the highly purified and streamlined manufacturing process, the currentspeeds of data transfer are around 5millionbps. The biggest challenge remaining isthe economic challenge. Today telephone and cable television companies generallybring in fiber links (backbones)to remote sites serving many customers, but thenuse twisted wire pair or coaxial cables from optical network units to individualhomes. This technology is often referred to "broadband" and is becomingincreasingly popular, but considerably limited to the potential of complete fiberoptic networks directly linked to individual homes. Only time will tell how long it

will take before the technology becomes reasonably economical and enoughdemand is given to take that next step.

1

Uses of Optics in Computing

Currently, optics is used mostly to link portions of computers, or moreintrinsically in devices that have some optical application or component. Forexample, much progress has been achieved, and optical signal processors have

been successfully used, for applications such as synthetic aperture radars, opticalpattern recognition, optical image processing, fingerprint enhancement, and opticalspectrum analyzers. The early work in optical signal processing and computingwas basically analog in nature.

In the past two decades, however, a great deal of effort has been expended inthe development of digital optical processors. Much work remains before digitaloptical computers will be widely available commercially, but the pace of researchand development has increased through the 1990s. During the last decade, therehas been continuing emphasis on the following aspects of optical computing:

Optical tunnel devices are under continuous developmentvarying from small caliber endoscopes to characterrecognition systems with multiple type capability.

Development of optical processors for asynchronoustransfer mode.

Development architectures for optical neural networks. Development of highaccuracy analog optical processors, capable of processing large amounts of data inparallel.




35

Since photons are uncharged and do not interact with one another as readilyas electrons, light beams may pass through one another in full-duplex operation,

for example without distorting the information carried. In the case of electronics,loops usually generate noise voltage spikes whenever the electromagnetic fieldsthrough the loop changes. Further, high frequency or fast switching pulses willcause interference in neighboring wires.

On the other hand, signals in adjacent optical fibers or in optical integratedchannels do not affect one another nor do they pick up noise due to loops. Finally,optical materials possess superior storage density and accessibility over magneticmaterials. The field of optical computing is progressing rapidly and shows manydramatic opportunities for overcoming the limitations described earlier for currentelectronic computers. The process is already underway whereby optical deviceshave been incorporated into many computing systems. Laser diodes as sources ofcoherent light have dropped rapidly in price due to mass production.

Also, optical CD-ROM discs are now very common in home and officecomputers. Current trends in optical computing emphasize communications, forexample the use of free-space optical interconnects as a potential solution toalleviate bottlenecks experienced in electronic architectures, including loss ofcommunication efficiency in multiprocessors and difficulty of scaling down the ICtechnology to sub-micron levels. Light beams can travel very close to each other,and even intersect, without observable or measurable generation of unwanted

signals. Therefore, dense arrays of interconnects can be built using optical systems.In addition, risk of noise is further reduced, as light is immune to electromagneticinterferences. Finally, as light travels fast and it has extremely large spatialbandwidth and physical channel density, it appears to be an excellent media forinformation transport and hence can be harnessed for data processing. This highbandwidth capability offers a great deal of architectural advantage and flexibility.Based on the technology now available, future systems could have 1024 smartpixels per chip with each channel clocked at 200MHz (a chip I/O of 200Gbits persecond), giving aggregate data capacity in the parallel optical highway of morethat 200Tbits per second; this could be further increased to 1000Tbits. Free-spaceoptical techniques are also used in scalable crossbar systems, which allow arbitraryinterconnections between a set of inputs and a set of outputs. Optical sorting andoptical crossbar inter-connects are used in asynchronous transfer modes or packetrouting and in shared memory multiprocessor systems.

In optical computing two types of memory are discussed. One consists ofarrays of one-bit-store elements and the other is mass storage, which isimplemented by optical disks or by holographic storage systems. This type ofmemory promises very high capacity and storage density. The primary benefits




36

offered by holographic optical data storage over current storage technologiesinclude significantly higher storage capacities and faster read-out rates. This

research is expected to lead to compact, high-capacity, rapid- and random-access,radiation-resistant, low-power, and low-cost data storage devices necessary forfuture intelligent spacecraft, as well as to massive-capacity and fast-accessterrestrial data archives. As multimedia applications and services become more andmore prevalent, entertainment and data storage companies are looking at ways toincrease the amount of stored data and reduce the time it takes to get that data outof storage. The SLMs and the linear array beam steerer are used in optical datastorage applications. These devices are used to write data into the optical storagemedium at high speed.

The analog nature of these devices means that data can be stored at muchhigher density than data written by conventional devices. Researchers around theworld are evaluating a number of inventive ways to store optical data whileimproving the performance and capacity of existing optical disk technology. Whilethese approaches vary in materials and methods, they do share a commonobjective: expanded capacity through stacking layers of optical material. For audiorecordings, a 150-MB minidisk with a 2.5-in. diameter has been developed thatuses special compression to shrink a standard CD’s 640-MB storage capacityonto the smaller polymer substrate. It is rewritable and uses magnetic fieldmodulation on optical material. The minidisk uses one of two methods to write

information onto an optical disk. With the minidisk, a magnetic field placed behindthe optical disk is modulated while the intensity of the writing laser head is heldconstant. By switching the polarity of the magnetic field while the laser creates astate of flux in the optical material, digital data can be recorded on a single layer.As with all optical storage media, a read laser retrieves the data. Along withminidisk developments, standard magneto-optical CD technology has expanded thecapacity of the 3.5-in. diameter disk from 640 MB to commercially available 1 GBstorage media. These conventional storage media modulate the laser instead of themagnetic field during the writing process. Fourth-generation 8,5.25 in.diameterdisks that use the same technology have reached capacities of 4 GB per disk. Thesedisks are used mainly in ‘jukebox’ devices. Not to be confused with the musical jukebox, these machines contain multiple disks for storage and backup of largeamounts of data that need to be accessed quickly.

Beyond these existing systems are several laboratory systems that usemultiple layers of optical material on a single disk. The one with the largestcapacity, magnetic super-resolution (MSR), uses two layers of optical material.The data is written onto the bottom layer through a writing laser and magnetic fieldmodulation (MFM). When reading the disk in MSR mode, the data is copied from




37

the lower layer to the upper layer with greater spacing between bits. In this way,data can be stored much closer together (at distances smaller than the read beam

wavelength) on the bottom layer without losing data due to averaging across bits.This method is close to commercial production, offering capacities of up to 20 GBon a 5.25 in. disk without the need for altering conventional read-laser technology.Advanced storage magnetic optics (ASMO) builds on MSR, but with oneexception.

Standard optical disks, including those used in MSR, have grooves and lands just like a phonograph record. These grooves are used as guideposts for the writingand reading lasers. However, standard systems only record data in the grooves, noton the lands, wasting a certain amount of the optical material’s capacity. ASMOrecords data on both lands and grooves and, by choosing groove depthsapproximately 1/6 the wavelength of the reading laser light, the system caneliminate the crosstrack crosstalk that would normally be the result of recording onboth grooves and lands. Even conventional CD recordings pick up data fromneighboring tracks, but this information is filtered out, reducing the signal-to-noiseratio. By closely controlling the groove depth, ASMO eliminates this problemwhile maximizing the signal-to-noise ratio. MSR and ASMO technologies areexpected to produce removable optical disk drives with capacities between 6 and20 GB on a 12-cm optical disk, which is the same size as a standard CD that holds640 MB. Magnetic amplifying magneto-optical systems (MAMMOS) use a

standard polymer disk with two or three magnetic layers. In general terms,MAMMOS is similar to MSR, except that when the data is copied from the bottomto the upper layer, it is expanded in size, amplifying the signal. According toArchie Smith of Storagetek’s Advanced Technology Office (Louisville, CO),MAMMOS represents a two-fold increase in storage capacity over ASMO.Technology developed by Call/Recall Inc. (San Diego, CA) could help bridge thegap between optical disk drives and holographic memories. Called 2-photonoptical storage technology (which got its start with the assistance of the Air Forceresearch laboratories and DARPA), the Call/Recall systems under development usea single beam to write the data in either optical disks with up to 120 layers, or into100-layer cubes of active-molecule-doped MMA polymer. In operation, a maskrepresenting data is illuminated by a mode-locked Nd:YAG laser emitting at 1064nm with pulse durations of 35 ps. The focal point of the beam intersects a secondbeam formed by the second harmonic of the same beam at 532 nm. The secondbeam fixes the data spatially and temporally. A third beam from a He Ne laseremitting at 543 nm reads the data by causing the material to fluoresce. Thefluorescence is read by a chargecoupled device (CCD) chip and converted throughproprietary algorithms back into data. Newer versions of the system use a




38

Ti:Sapphire laser with 200-fs pulses. Call/Recall’s Fredrick McCormick said thenewer and older approaches offer different strengths. The YAG system can deliver

higher-power pulses capable of storing megabits of data with a single pulse, but atmuch lower repetition rates than the Ti:Sapphire laser with its lower-power pulses.Thus, it is a trade-off. Call/Recall has demonstrated the system using portableapparatus comprised of a simple stepper-motor-driven stage and 200-microwattHeNe laser in conjunction with a low-cost video camera. The company estimatesthat an optimized system could produce static bit error rates (BER) of less than 910–13. McCormick believes that a final prototype operating at standard CDrotation rates would offer BERs that match or slightly exceed conventional opticaldisk technology. Researchers such as Demetri Psaltis and associates at theCalifornia Institute of Technology are also using active-molecule-doped polymersto store optical data holographically.

Their system uses a thin polymer layer of PMMA doped withphenanthrenequinone (PQ). When illuminated with two coherent beams, thesubsequent interference pattern causes the PQ molecules to bond to the PMMAhost matrix to a greater extent in brighter areas and to a lesser extent in areas wherethe intensity drops due to destructive interference. As a result, a pair of partiallyoffsetting index gratings is formed in the PMMA matrix. After writing thehologram into the polymer material, the substrate is baked, which causes theremaining unbounded PQ molecules to diffuse throughout the polymer, removing

the offsetting grating and leaving the hologram. A uniform illumination is the finalstep, bonding the diffuse PQ throughout the matrix and fixing the hologram in thepolymer material.

Storagetek’s Archie Smith estimates that devices based on this method couldhold between 100 and 200 GB of data on a 5.25-in diameter polymer disk.

More conventional approaches to holographic storage use irondopedlithium niobate crystals to store pages of data. Unlike standard magneto-opticalstorage devices, however, the systems developed by Pericles Mitkas at ColoradoState University use the associative search capabilities of holographic memories.Associative or content-based data access enables the search of the entire memoryspace in parallel for the presence of a keyword or search argument. Conventionalsystems use memory addresses to track data and retrieve the data at that locationwhen requested. Several applications can benefit from this mode of operationincluding management of large multimedia databases, video indexing, imagerecognition, and data mining.

Different types of data such as formatted and unformatted text, gray scaleand binary images, video frames, alphanumeric data tables, and time signals can be




39

interleaved in the same medium and we can search the memory with either datatype. The system uses a data and a reference beam to create a hologram on one

plane inside the lithium niobate. By changing the angle of the reference beam,more data can be written into the cube just like pages in a book. The currentsystems have stored up to 1000 pages per spatial location in either VGA or VGAresolutions. To search the data, a binary or analog pattern that represents the searchargument is loaded into a spatial light modulator and modulates a laser beam. Thelight diffracted by the holographic cube on a CCD generates a signal that indicatesthe pages that match the sought data. Recent results have shown the system canfind the correct data 75 percent of the time when using patterns as small as 1 to 5percent of the total page. That level goes up to 95 to 100 percent by increasing theamount of data included in the search argument.2

Why Use Optics for Computing?

Optical interconnections and optical integrated circuits have severaladvantageous over their electronic counterparts. They are immune toelectromagnetic interference, and free from electrical short circuits. They havelow-loss transmission and provide large bandwidth; i.e. multiplexing capability,capable of communicating several channels in parallel without interference. Theyare capable of propagating signals within the same or adjacent fibers with

essentially no interference or cross-talk. They are compact, lightweight, andinexpensive to manufacture, and more facile with stored information than magneticmaterials.

We are in an era of daily explosions in the development of optics and opticalcomponents for computing and other applications. The business of photonics isbooming in industry and universities worldwide. It is estimated that photonicdevice sales worldwide will range between $12 billion and $100 billion in 1999due to an ever-increasing demand for data traffic.

According to KMI corp., data traffic is growing worldwide at a rate of 100%

per year, while, the Phillips Group in London estimates that the U.S. data trafficwill increase by 300% annually. KMI corp. also estimates that sales of dense-wavelength division multiplexing equipment will increase by more than quadrupleits growth in the next five years, i.e. from $2.2 billion worldwide in 1998 to $9.4billion 2004. In fact, Future Communication Inc., London, announced this year toupgrade their communication system accordingly. The companyÕs goal is to usewavelength division multiplexing at 10 Gb/s/channel to transmit at a total rate ofmore than 1000 Tb/s.




40

Most of the components that are currently very much in demand are electro-optical (EO). Such hybrid components are limited by the speed of their electronic

parts. All-optical components will have the advantage of speed over EOcomponents. Unfortunately, there is an absence of known efficient nonlinearoptical materials that can respond at low power levels. Most alloptical componentsrequire a high level of laser power to function as required. A group of researchersfrom the university of southern California, jointly with a team from the universityof California Los Anglos, have developed an organic polymer with a switchingfrequency of 60 GHz. This is three times faster than the current industry standard,lithium niobate crystal-based devices. The California team has been working toincorporate their material into a working prototype. Development of such a devicecould revolutionize the information superhighway and speed data processing foroptical computing. Another group at Brown University and the IBM.

Almaden Research Center (San Jose, CA) have used ultrafast laser pulses tobuild ultrafast datastorage devices. This group was able to achieve ultrafastswitching down to 100ps. Their results are almost ten times faster than currentlyavailable Òspeed limitsÓ. Optoelectronic technologies for optical computers andcommunication hold promise for transmitting data as short as the spacebetween computer chips or as long as the orbital distance between satellites. AEuropean collaborative effort demonstrated a high-speed optical data input andoutput in free-space between IC chips in computers at a rate of more than 1 Tb/s.

Astro Terra, in collaboration with Jet Propulsion Laboratory (Pasadena, CA) hasbuilt a 32-channel 1-Ggb/s earth Ðto Ðsatellite link with a 2000 km range. Manymore active devices in development, and some are likely to become crucialcomponents in future optical computer and networks.

The race is on with foreign competitors. NEC (Tokyo, Japan) havedeveloped a method for interconnecting circuit boards optically using VerticalCavity Surface Emitting Laser arrays (VCSEL). Researchers at Osaka CityUniversity (Osaka, Japan) reported on a method for automatic alignment of a set ofoptical beams in space with a set of optical fibers.

As of last year, researchers at NTT (Tokyo, Japan) have designed an opticalback plane with free Ðspace optical interconnects using tunable beam deflectorsand a mirror. The project had achieved 1000 interconnections per printed-circuitboard, with throughput ranging from 1 to 10 Tb/s.

Optics has a higher bandwidth capacity over electronics, which enables moreinformation to be carried and data to be processed arises because electroniccommunication along wires requires charging of a capacitor that depends onlength. In contrast, optical signals in optical fibers, optical integrated circuits, andfree space do not have to charge a capacitor and are therefore faster.




41

Another advantage of optical methods over electronic ones for computing isthat optical data processing can be done much easier and less expensive in parallel

than can be done in electronics. Parallelism is the capability of the system toexecute more than one operation simultaneously. Electronic computer architectureis, in general, sequential, where the instructions are implemented in sequence. Thisimplies that parallelism with electronics is difficult to construct. Parallelism firstappeared in Cray super computers in the early 1980s.

Two processors were used in conjunction with the computer memory toachieve parallelism and to enhance the speed to more than 10 Gb/ s. It was laterrealized that more processors were not necessary to increase computational speed,but could be in fact detrimental. This is because as more processors are used, thereis more time lost in communication. On the other hand, using a simple opticaldesign, an array of pixels can be transferred simultaneously in parallel from onepoint to another. To appreciate the difference between both optical parallelism andelectronic one can think of an imaging system of as many as 1000x1000independent points per mmin the object plane which are connected optically by alens to a corresponding 1000x 1000 points per mm in the image plane. For this tobe accomplished electrically, a million nonintersecting and properly isolatedconduction channels per mm would be required.

Parallelism, therefore, when associated with fast switching speeds, wouldresult in staggering computational speeds. Assume, for example, there are only 100

million gates on a chip, much less than what was mentioned earlier (opticalintegration is still in its infancy compared to electronics). Further, conservativelyassume that each gate operates with a switching time of only 1 nanosecond(organic optical switches can switch at sub-picosecond rates compared tomaximum picosecond switching times for electronic switching). Such a systemcould perform more than 1017 bit operations per second. Compare this to thegigabits (109) or terabits (1012) per 6 second rates which electronics are eithercurrently limited to, or hoping to achieve.

In other words, a computation that might require one hundred thousandhours (more than 11 years) of a conventional computer could require less than onehour by an optical one.

Another advantage of light results because photons are uncharged and do notinteract with one another as readily as electrons. Consequently, light beams maypass through one another in fullduplex operation, for example without distortingthe information carried. In the case of electronics, loops usually generate noisevoltage spikes whenever the electromagnetic fields through the loop changes.Further, high frequency or fast switching pulses will cause interference inneighboring wires. Signals in adjacent fibers or in optical integrated channels do




42

not affect one another nor do they pick up noise due to loops. Finally, opticalmaterials possess superior storage density and accessibility over magnetic

materials.Obviously, the field of optical computing is progressing rapidly and showsmany dramaticopportunities for overcoming the limitations described earlier forcurrent electronic computers.

The process is already underway whereby optical devices have beenincorporated into many computing systems. Laser diodes as sources of coherentlight have dropped rapidly in price due to mass production. Also, optical CD-ROMdiscs have been very common in home and office computers.

OPTICAL DISK13

WORKING

The 780nm light emitted from AlGaAs/GaAs laser diodes is collimated by a lensand focused to a diameter of about 1micrometer on the disk. If there is no pit wherethe light is incident, it is reflected at the Al mirror of the disk and returns to thelens, the depth of the pit is set at a value such that the difference between the pathof the light reflected at a pit and the path of light reflected at a mirror is an integralmultiple of half-wavelength consequently, if there is a pit where light is incident,

the amount of reflected light decreases tremendously because the reflected lightsare almost cancelled by interference. The incident and reflected beams passthrough the quarter wave plate and all reflected light is introduced to thephotodiode by the beam splitter because of the polarization rotation due to thequarter wave plate. By the photodiode the reflected light, which has a signalwhether, a pit is on the disk or not is changed into an electrical signal.




43

An Optical Computer Powered by Germanium Laser

One of the issues of current chip design is the excessive power needed totransport and store ever increasing amounts of data. A possible solution is to useoptics not just for sending data, but also to store information and performcalculations, which would reduce heat dissipation and increase operating speeds.Disproving previous beliefs in the matter, MIT researchers have demonstrated thefirst laser built from germanium which can perform optical communications... andit's also cheap to manufacture.

As Moore's law keeps giving us faster and faster computers, chip buildersalso need higher-bandwidth data connections. But excessive heat dissipation and

power requirements make conventional wires impractical at higher frequencies,which has lead researchers to develop new ways to store, transmit and elaborateoptically-encoded information.

If optical-based data elaboration is to have a future, researchers will need tofind a cheap and effective way to integrate optical and electronic components ontosilicon chips.

The solution found by the MIT team and detailed in a paper published in the journal Optics Letters is notable not only because it achieves these objectives, butalso because it changes the way physicists have been looking at a class of materialsthat were previously thought to be unsuitable for manufacturing lasers.

In a semiconductor, electrons that receive a certain amount of energy enter a"conduction band" and are free to conduct electrical charge. Once they fall out ofthis excited state, the electrons can either release their energy as heat or as photons.Materials such as the expensive gallium arsenide are thought to be the best formanufacturing lasers, because their excited electrons tend to go fall back into thephoton-emitting state.

However, the MIT team demonstrated that materials such as germanium,whose electrons would normally tend to go in the heat-emitting state, can bemanipulated to emit photons and used to produce lasers that are cheap not onlybecause of the cost of the materials, but also because the processes used to buildthem are already very familiar to chip manufacturers.




44

The researchers found two ways to make germanium "optics-friendly". Thefirst is a technique called "doping," which involves implanting very low

concentrations of a material such as phosphorous to force more electrons in theconduction band and modify the electrical properties of the material.

The second strategy was to "strain" the germanium, pulling its atoms slightlyfarther apart than they would be naturally by growing it directly on top of a layerof silicon. This makes it easier for electrons to jump into the photon-emitting state.

The team now needs to find a way to increase the concentration ofphosphorus atoms in the doped germanium to increase the power efficiency of thelasers, making them more attractive as sources of light for optical data

connections and, one day, for computing as well.

First germanium laser could pave way for optical computers

By Dario Borghino

18:34 February 14, 2010

First germanium laser could pave way for optical computers




45

One of the issues of current chip design is the excessive power needed totransport and store ever increasing amounts of data. A possible solution is to use

optics not just for sending data, but also to store information and performcalculations, which would reduce heat dissipation and increase operating speeds.Disproving previous beliefs in the matter, MIT researchers have demonstrated thefirst laser built from germanium which can perform optical communications... andit's also cheap to manufacture.

As Moore's law keeps giving us faster and faster computers, chip buildersalso need higher-bandwidth data connections. But excessive heat dissipation andpower requirements make conventional wires impractical at higher frequencies,which has lead researchers to develop new ways to store, transmit and elaborate

optically-encoded information.

If optical-based data elaboration is to have a future, researchers will need tofind a cheap and effective way to integrate optical and electronic components ontosilicon chips.

The solution found by the MIT team and detailed in a paper published in the journal Optics Letters is notable not only because it achieves these objectives, butalso because it changes the way physicists have been looking at a class of materialsthat were previously thought to be unsuitable for manufacturing lasers.

In a semiconductor, electrons that receive a certain amount of energy enter a"conduction band" and are free to conduct electrical charge. Once they fall out ofthis excited state, the electrons can either release their energy as heat or as photons.Materials such as the expensive gallium arsenide are thought to be the best formanufacturing lasers, because their excited electrons tend to go fall back into thephoton-emitting state.

However, the MIT team demonstrated that materials such as germanium,

whose electrons would normally tend to go in the heat-emitting state, can bemanipulated to emit photons and used to produce lasers that are cheap not onlybecause of the cost of the materials, but also because the processes used to buildthem are already very familiar to chip manufacturers.

The researchers found two ways to make germanium "optics-friendly". Thefirst is a technique called "doping," which involves implanting very lowconcentrations of a material such as phosphorous to force more electrons in theconduction band and modify the electrical properties of the material.




46

The second strategy was to "strain" the germanium, pulling its atoms slightlyfarther apart than they would be naturally by growing it directly on top of a layer

of silicon. This makes it easier for electrons to jump into the photon-emitting state.The team now needs to find a way to increase the concentration of

phosphorus atoms in the doped germanium to increase the power efficiency of thelasers, making them more attractive as sources of light for optical data connectionsand, one day, for computing as well.

The work is part of the Si-Based-Laser Initiative of the MultidisciplinaryUniversity Research Initiative (MURI), and was sponsored by the Air Force Officeof Scientific Research (AFOSR).




47

Concept of Picosecond (By NASA)

NASA scientists are working to solve the need for computer speed usinglight itself to accelerate calculations and increase data bandwidth.

Watches tick in seconds. Basketball games are timed in 10ths of a second,and drag racers in 100ths. Computers used to work in milliseconds (1,000ths), thenmoved up to microseconds (millionths), and now are approaching nanoseconds(billionths) for logic operations - and picoseconds (trillionths!) for the switches andgates in chips.

"That's great in theory," says Dr. Donald Frazier of NASA's Marshall SpaceFlight Center. "Except that electronic signals, even with Very Large ScaleIntegration (VLSI) and maximum miniaturization, are bogged down by manyaspects of the solid materials they travel through. So we've had to find a fastermedium for the signals - and the answer seems to be light itself!" Above: Dr.Donald Frazier monitors a blue laser light used with electro-optical materials.

Light travels at 186,000 miles per second. That's 982,080,000 feet persecond -- or 11,784,960,000 inches. In a billionth of a second, one nanosecond,

photons of light travel just a bit less than a foot, not considering resistance in air orof an optical fiber strand or thin film. Just right for doing things very quickly inmicrominiaturized computer chips.

"Entirely optical computers are still some time in the future," says Dr.Frazier, "but electro-optical hybrids have been possible since 1978, when it waslearned that photons can respond to electrons through media such as lithiumniobate. Newer advances have produced a variety of thin films and optical fibersthat make optical interconnections and devices practical. We are focusing on thin

films made of organic molecules, which are more light sensitive than inorganics.Organics can perform functions such as switching, signal processing and frequencydoubling using less power than inorganics. Inorganics such as silicon used withorganic materials let us use both photons and electrons in current hybrid systems,which will eventually lead to all-optical computer systems."

"What we are accomplishing in the lab today will result in development ofsuper-fast, super-miniaturized, super-lightweight and lower cost optical computingand optical communication devices and systems," Frazier explained.




48

The speed of computers has now become a pressing problem as electronic

circuits reach their miniaturization limit. The rapid growth of the Internet,expanding at almost 15% per month, demands faster speeds and larger bandwidthsthan electronic circuits can provide. Electronic switching limits network speeds toabout 50 Gigabits per second (1 Gigabit (Gb) is 109, or 1 billion bits).

Dr. Hossin Abdeldayem, a member of Frazier's optical technologiesresearch group, states that Terabit speeds (1 Terabit, abbreviated "Tb", is 1012, or1 trillion bits) are needed to accommodate the growth rate of the Internet and theincreasing demand for bandwidth-intensive data streams. Optical data processingcan perform several operations simultaneously (in parallel) much faster and easierthan electronics. This "parallelism" when associated with fast switching speedswould result in staggering computational power. For example, a calculation thatmight take a conventional electronic computer more than eleven years to completecould be performed by an optical computer in a single hour.

All-optical switching using optical materials can relieve the escalatingproblem of bandwidth limitations imposed by electronics," says Dr. Abdeldayem."In 1998, Lucent Technologies introduced a lithographic submicron technology tofurther miniaturize electronic circuits and enhance computer speed. Additional

miniaturization of electronic components only provides a short-term solution to theproblem. There are also physical problems accompanied by miniaturization thatmight affect the computer's reliability. "

Drs. Frazier and Abdeldayem and their group in Huntsville, AL, havedesigned and built all-optical logic gate circuits for data processing at Gigabit andTerabit rates, and they are also working on a system for pattern recognition.

Dr. Hossin Abdeldayem of NASA/Marshall works with lasers to develop asystem for pattern recognition. "We have also developed and tested nanosecondoptical switches, which can act as computer logic gates," says Dr. Abdeldayem,who recently presented the group's research paper entitled "All-Optical LogicGates for Optical Computing" at The Pittsburgh Conference in New Orleans, LA."Picosecond and nanosecond all-optical switches, which act as AND and partialNAND logic gates were demonstrated in our laboratory," explains Dr.Abdeldayem. "Such logic gates are members of a large family of gates incomputers that perform logic operations such as addition, subtraction andmultiplication. They are vital for the development of optical computing and optical




49

communication. Our all-optical logic gates were made using a thin film of metal-free phthalocyanine compound and a polydiacetylene polymer in a hollow fiber"

Logic gates are the building blocks of any digital system," he continues."An optical logic gate is a switch that controls one light beam with another. It is"on" when the device transmits light, and "off" when it blocks the light."

"Our phthalocyanine switch operates in the nanosecond regime (i.e.,Gigabits per second), functioning as an all-optical AND logic gate. To demonstrateit, we waveguided a continuous (cw) laser beam co-linearly with a nanosecondpump beam through a thin film of metal-free phthalocyanine. The output was sentto a fast photo-detector and to an oscilloscope. The cw beam was found to pulsatesynchronously with the pump beam, showing the characteristic table of an ANDlogic gate."

A schematic of the nanosecond all-optical AND logic gate setup. Moreschematics and illustrations are available in "Recent Advances in Photonic Devicesfor Optical Computing" by NASA/Marshall's Hossin Abdeldayem, Donald O.Frazier, Mark S. Paley, and William K. Witherow.

"Our setup for the picosecond switch was similar, except that the

phthalocyanine film was replaced with a hollow fiber coated from inside with athin polydiacetylene film. Both collinear laser beams were focused on one end ofthe tube, and a lens at the other end focused the output onto a monochrometer witha fast detector attached. The product of the two beams demonstrates three of thefour characteristics of a NAND logic gate."

"Optical bistable devices and logic gates such as these are the equivalent ofelectronic transistors," concludes Dr. Abdeldayem. "They operate as very highspeed on-off switches and are also useful as optical cells for information storage."

According to Dr. Frazier, these all-optical computer components and thin-films developed by NASA are essential to the current worldwide work in electro-optical hybrid computers - and will help to make possible the astounding organicoptical computers that will be the standard of future terrestrial and spaceinformation, operating and communication systems.




50

Recent Advances in Photonic Switches at NASA/MSFC

Logic gates are the building blocks of any digital system. An optical logicgate is a switch that controls one light beam by another; it is ÒONÓ when thedevice transmits light and it is ÒOFFÓ when it blocks the light. Recently wedemonstrated in our laboratory at NASA/Marshall Space Flight Center two fast all-optical switches using phthalocyanine thin films and polydiacetylene fiber. Thephthalocyanine switch is in the nanosecond regime and functions as an all-opticalAND logic gate, while the polydiacetylene one is in the picosecond regime andexhibits a partial all-optical NAND logic gate.

To demonstrate the AND gate in the phthalocyanine film, we waveguidedtwo focused collinear beams through a thin film of metal-free phthalocyanine film.The film thickness was ~ 1 m and a few millimeters in length. We used the secondharmonic at 532 nm from a pulsed Nd:YAG laser with pulse duration of 8 ns along with a cw He-Ne beam at 632.8 nm. The two collinear beams were thenfocused by a microscopic objective and sent through the phthalocyanine film.

At the output a narrow band filter was set to block the 532 nm beam andallow only the He-Ne beam. The transmitted beam was then focused on a fastphoto-detector and to a 500 MHz oscilloscope. It was found that the transmittedHe-Ne cw beam was pulsating with a nanosecond duration and in synchronouswith the input Nd:YAG nanosecond pulse. The setup described above

demonstrated the characteristic table of an AND logic gate.The setup for the picosecond switch was very much similar to the setup infigure 3 except that the phthalocyanine film was replaced by a hollow fiber filledwith a polydiacetylene. The polydiacetylene fiber was prepared by injecting adiacetylene monomer into the hollow fiber and polymerizing it by UV lamps. TheUV irradiation induces a thin film of the polymer on the interior of the hollow fiberwith a refractive index of 1.7 and the hollow fiber is of refractive index 1.2. In theexperiment, the 532 nm from a mode locked picosecond laser was sent collinearlywith a cw He-Ne laser and both were focused onto one end of the fiber. At theother end of the fiber a lens was focusing the output onto the narrow slit of amonochrometer with its grating set at 632.8 nm. A fast detector was attached to themonochrometer and sending the signal to a 20 GHz digital oscilloscope. It wasfound that with the He-Ne beam OFF, the Nd:YAG pulse is inducing a weekfluorescent picosecond signal (40 ps) at 632.8 nm that is shown as a picosecondpulse on the oscilloscope. This signal disappears each time the He-Ne beam isturned on. These results exhibit a picosecond respond in the system anddemonstrated three of the four characteristics of a NAND logic gate .




51

A comparison of a scanning electron micrographs of 1 m thick films ofcopper phthalocyanine deposited by physical vapor transport in the 3M PVTOS

flight (STS-20) and ground control experiments. In microgravity the filmÕsmicrostructure is very dense compared to that produced in unit gravity in thepresence of convection. This difference in microstructure has a significant affecton the macroscopic film optical properties.

A comparison of a ground-grown polydiacetylene film with a microgravity-grown one. The aggregates are impeded into the film by the fluid convection on theground, while the microgravity film is almost free of these aggregates whereconvection is almost absent.

A schematic of the nanosecond all-optical AND logic gate setup. Aschematic of the all-optical NAND logic gate setup.




52

Optical computer bus

with dynamic bandwidth allocationA signal communication device for use within a computer includes a set of

optical fibers configured to form an optical computer bus between a set ofcomputer sub-system elements of a computer. A set of input optical connectorcards are connected to the set of optical fibers. Each of the input optical connectorcards includes a transmitting dynamic bandwidth allocator responsive to an opticalbus clock signal operating at a multiple of a computer system clock signal suchthat a set of bus time slots are available for each computer system clock signal

cycle. The transmitting dynamic bandwidth allocator allows a light signal to beapplied to the optical computer bus during a dynamically assigned bus time slot. Inthis way, the optical computer bus bandwidth can be dynamically allocated todifferent computer sub-system elements during a single computer system clocksignal cycle.

1. A method of signal communication within a computer, the methodcomprising the following steps:

(a) operating a bus clock signal at a multiple of a computer systemclock signal, said multiple being greater than one, such that a set of bus time slots

are available for each computer system clock signal;

(b) dynamically assigning at least one bus time slot to given ones of aplurality of computer sub-system elements during a single computer system clocksignal cycle;

(c) applying a signal from a computer sub-system element to acomputer bus during a dynamically assigned bus time slot.

2. The method of claim 1, wherein step (c) includes: applying a signal fromone of said plurality of computer sub-system elements to said computer bus duringsaid at least one bus time slot, a plurality of dynamically assigned bus time slotsbeing divisible among said plurality of computer sub-system elements inaccordance with a bandwidth requirement of one of the computer sub-systemelements.




53

3. The method of claim 1, wherein step (c) includes a step of alternatelyproducing a uniform, a random, and a dedicated division of bus time slots between

said plurality of computer sub-system elements.

4. The method of claim 1, wherein step (c) includes a step of converting saidsignal from said computer sub-system element to a light signal.

5. The method of claim 4, wherein step (c) includes coupling said lightsignal to an optical computer bus.

6. The method as recited in claim 1 further comprising determining areceiver skew value measured between a received signal and said bus clock signal.

7. The method as recited in claim 6 further comprising reducing a transmitterskew value between a transmitted signal and said bus clock signal based upon saidreceiver skew value.

BRIEF DESCRIPTION OF THE INVENTION

This invention relates generally to computer buses. More particularly, thisinvention relates to a computer bus that is implemented with optical fibers to avoid

the physical limitations associated with traditional computer bus designs.

BACKGROUND OF THE INVENTION

A computer bus is a communication link used to connect multiple computersubsystems. For example, a computer bus is used to link the memory andprocessor, and to link the processor with input/output (I/O) devices. Computerbuses are traditionally classified as follows: processor-memory buses, I/O buses, orbackplane buses. Processor-memory buses are short, generally high speed, andmatched to the memory system so as to maximize memory-processor bandwidth.I/O buses, by contrast, can be lengthy, can have many types of devices connectedto them, and often have a wide range in the data bandwidth of the devicesconnected to them. Backplane buses are designed to allow processors, memory,and I/O devices to coexist on a single bus. Backplane buses balance the demandsof processor-memory communication with the demands of I/O device-memorycommunication. Backplane buses received their name from the fact that they aretypically built into a computer backplane--the fundamental interconnection




54

structure within the computer chassis. Processor, memory, and I/O boards plug intoa backplane and then use the backplane bus to communicate.

Processor-memory buses are often design-specific, wile both I/O buses andbackplane buses are frequently standard buses with parameters established byindustry standards. The distinction between bus types is becoming increasinglydifficult to specify. Thus, the present application generically refers to computerbuses to encompass all processor-memory buses, I/O buses, and backplane buses.

The problem with computer buses is that they create a communicationbottleneck since all input/output must pass through a single bus. Thus, thebandwidth of the bus limits the throughput of the computer. Physical constraintsassociated with existing computer buses are beginning to limit the availableperformance improvements generally available in computers.

The physical operation and constraints of existing computer bus designs aremost fully appreciated .A computer bus 20 positioned on a backplane 22. Thecomputer bus 20 is a set of wires, effectively forming a transmission line. Arandom number of system cards (or cards) 24A-24N are attached to the computerbus 20. By way of example, the cards 24 may include a video processing card, amemory controller card, an I/O controller card, and a network card. Each card 24 is

connected to the computer bus 20 through a connector 26. Thus, each card 24 iselectrically connected to the set of wires forming the computer bus 20. As a result,one card, say card 24A, can communicate with another card, say card 24N, bywriting information onto the computer bus 20. Only one card 24 can writeinformation onto the computer bus 20 at a time, thus a computer bus 20 cangenerate a performance bottleneck as different cards 24 wait to write informationonto the bus 20.

Another problem associated with a traditional computer bus 20, is that itsperformance is constrained by complicated electrical phenomenon. For example,the connectors 26 effectively divide the bus into transmission line segments,resulting in complicated transmission line effects. Note that the transmission linesegments will vary depending upon the number of cards 26 connected to the bus20. This periodic loading of the bus 20 makes it difficult to optimize busperformance. In addition, each connector 26 produces a lumped discontinuity withparallel capacitance and series inductance, thereby complicating the electricalcharacteristics of the bus 20. Note also that "T-connections" are formed between




55

the wires of a computer bus 20 and the wires to a connector 26. The T-connectionscomplicate the electrical characteristics of the computer bus 20.

Each card 24 includes a transceiver circuit 28 connected to a card logiccircuit 30, which performs the functional operations associated with the card 24.The transceiver circuit 28 is used to read and write information on the bus 20. Thatis, the transceiver circuit 28 reads information from the bus 20, the card logiccircuit 30 processes the information, and then the transceiver circuit 28 writesprocessed information to the bus 20. Additional electrical complications arise withthe transceiver circuits 28. For example, transmission line segments are formedbetween each connector 26 and each bus transceiver 28 circuit. In addition, thetransceiver circuits 28 present an impedance at their package pins that dependsupon the circuit design, the electrical state of the transceiver, and the packaging.

In sum, the computer bus 20 constitutes a transmission line with complicatedelectrical interactions caused by such factors as transmission line segments andconnectors forming lumped discontinuities with parallel capacitance and seriesinductance. The bus 20 may be terminated with termination resistors (R) to reducetransmission line effects, such as reflections and mismatches. Nevertheless,solutions of this sort do not overcome all transmission line problems associatedwith a computer bus 20.

Given these complicated electrical interactions, signals on the bus 20 do notexperience a uniform rise. That is, if the bus 20 was a perfect transmission line,then high signals (digital ONES) written to the bus 20 would experience a uniformrise. However, in view of the complicated electrical interactions on the bus 20,high signals frequently experience one or more spurious signal transitions beforereaching a final peak value that can be processed. Waiting for signals to settlecauses delays. Another problem is that the complicated electrical interactions onthe computer bus 20 require higher powered drive signals, and thus more powerdissipation.

It is difficult to avoid these problems by changing the electricalcharacteristics of the bus 20. That is, it is difficult to design a bus with improvedtransmission line properties in view of the complicated factors that establish bus 20performance. Thus, it would be highly desirable to design a new type of bus whoseperformance is not contingent upon complicated transmission line effectsassociated with prior art buses.




56

SUMMARY OF THE INVENTION

An embodiment of the invention includes a set of optical fibers configuredto form an optical computer bus between a set of computer sub-system elements ofa computer. A set of input optical connector cards are connected to the set ofoptical fibers. Each of the input optical connector cards includes a transmittingdynamic bandwidth allocator responsive to an optical bus clock signal operating ata multiple of a computer system clock signal such that a set of bus time slots areavailable for each computer system clock signal cycle. The transmitting dynamicbandwidth allocator allows a light signal to be applied to the optical computer busduring a dynamically assigned bus time slot. In this way, the optical computer busbandwidth can be dynamically allocated to different computer sub-system elementsduring a single computer system clock signal cycle.

The optical computer bus is extremely fast, with bus signals moving atapproximately the speed of light ((index of refraction of the fiber)31 1×the speedof light). The operation of the bus is not compromised by transmission line effectsassociated with prior art computer buses. Further, the optical computer bus of theinvention does not suffer from electrical noise problems. The optical computer busis compact and is therefore ideal for space-constrained modern computers. Despiteits radically different design and configuration, the computer bus of the invention

otherwise operates in a standard manner. Thus, the computer bus can be used inexisting computers and system designers can still rely upon known bus designtechniques.

DETAILED DESCRIPTION OF THE INVENTION

A digital gate computer bus 40, also called a chip bus, in accordance withthe invention. The chip bus 40 of the invention uses digital circuits 30 to performthe function executed by a conventional computer bus. That is, the chip bus 40 ofthe invention is used to perform a set of logical OR operations with digital gates sothat these operation do not have to be performed as wired OR operations on thewires of a computer bus. In this way, the transmission line problems associatedwith prior art computer buses are eliminated.

The operation of the invention is more fully appreciated with a simple example.Typically, each card attached to a computer bus has N communication bitscorresponding to the N wires forming the computer bus. Thus, for example, if fourcards are attached to a computer bus, then each card has a designated bit that reads




57

and writes signals to a designated wire of the computer bus. If any card on the buswrites a digital ONE to this designated wire of the computer bus, then all cards on

the bus read a digital high signal for this designated bit. This is a logical ORoperation performed by a hardwired circuit (the wire of the bus). The presentinvention eliminates the physical wires of traditional computer buses and executesthe operation associated with such wires with digital gates. That is, the chip bus 40of the invention performs logical OR operations with digital gates in order toeliminate the transmission line problems associated with prior art computer buses.

The chip bus 40 is positioned on a backplane 22. Chip bus communicationlines 42 are electrically connected to the chip bus 40. In one embodiment, chip businput lines 44 carry input signals to the chip bus 40, the chip bus performs logicalOR operations on the input signals and generates output signals which are appliedto chip bus output lines 46. The chip bus communication lines 42 are electricallyconnected to connectors 48, which in turn are electrically connected to cards 49.The connectors 48 and cards 49 may be of the type known in the art. Thus, the chipbus 40 of the invention can be used with prior art computer configurations.

Therein is a single bit embodiment of the chip bus 40 of the invention. Inparticular, the figure illustrates a chip bus bit processor 50. The chip bus bitprocessor 50 includes a logical OR circuit 51, illustratively shown as a wired OR

circuit. In this embodiment of the invention, the chip bus bit processor 50 alsoincludes a card signal driver 52 with a bus input signal driver 54, implemented asan inverter, and a bus output signal driver 56, also implemented as an inverter.

Thus, it can be appreciated that the chip bus bit processor 50 of receives asingle bit input signal from four cards (49A, 49B, 49C, 49D). In particular, eachsingle bit input signal is driven by the bus input signal driver 54 and applied to thelogical OR circuit 51. If any single bit input signal is a digital ONE on the logicalOR circuit 51, then a high output is generated at all output nodes. For theembodiment, the high output signal is seen by the card logic circuit 66 afterprocessing by inverters 56 and 64.

In one embodiment of the invention, the card 49B may include a cardtransceiver 60B. In this embodiment, the card transceiver 60B includes a logicoutput signal driver 62, implemented as an inverter, and a logic input signal driver64, also implemented as an inverter. The signals from the card transceiver 60B arethen processed by a logic circuit 66 in a conventional manner.




58

Therein are the same components shown , but the components arerearranged to more fully describe the invention. In addition, the logical OR circuit

51 as being implemented with a four input OR gate. Thus, it is seen that each card(49A, 49B, 49C, 49D) generates a single bit signal that is respectively applied tothe chip bus input lines (44A, 44B, 44C, 44D). The four signals are routed to thefour input OR gate 51. The output of the four input OR gate 51 is then routed backto the cards (49A, 49B, 49C, 49D) through their respective chip bus output lines(46A, 46B, 46C, 46D).

A four bit digital gate computer bus in accordance with the invention. Thefour bit digital gate computer bus is used in conjunction with four processing cards(49A, 49B, 49C, 49D). The four bit digital gate computer bus includes a chip buspackage 70 with package pins 72. Standard packaging techniques may be used toform this structure. Within the package 70 are four chip bus bit processors (50A,50B, 50C, 50D). The package 70 is positioned on a backplane 22.

Each processing card (49A, 49B, 49C, 49D) generates a single bit signal thatis applied to one of the chip bus bit processors 50. In particular, each processingcard generates a single bit signal that is applied to a chip bus input line 44 formedon backplane 22. The signal reaches a package pin 72 and is then routed to a chipbus bit processor 50 via a package internal trace 74. After processing by the chip

bus bit processor 50 is completed, the output signals are applied to chip bus outputlines 46 formed in the backplane 22. The chip bus output lines 46 route the outputsignals to their respective cards for processing in a standard fashion.

The invention has now been fully described. Attention presently turns to adiscussion of various implementation issues. Implementations of the chip bus 40 ofthe invention will have the shared portion of a physical bus implemented withdigital gates and will use point-to-point wiring to connect the daughterboards(cards 24). As used herein, point-to-point wiring refers to wiring running directlybetween pins of two packages, without "T-connections", "Y-connections", orrelated configurations or sources which complicate signal transmission.

The preferred embodiment of the invention uses separate chip bus input lines44 and chip bus output lines 46. However, it is possible to use bidirectional wiresto make these connections. The bidirectional wires save a factor of two in signals,but cannot reach the speed attainable by the unidirectional technique, unlessspecial transceivers are used that can simultaneously send and receive on the same




59

line. For the highest speed systems, it may be advantageous to use differentialsimultaneous bidirectional signaling to reduce system noise.

Simultaneous bidirectional transceiver technology has been available inemitter coupled logic for many years. The technology depends on having very highperformance differential amplifiers to subtract the outgoing signal from the signalon the pin to recover the incoming signal. Simultaneous bidirectional signaling hasbeen demonstrated in CMOS, but is harder to implement because the closematching and high gain of the bipolar devices is not available.

If the chip bus 40 is used in a synchronous system which is properlyarbitrated, it is not necessary to provide any control signals to control who isdriving the bus. While additional control signals are not required, the chip bus 40of the invention does require more wires on the backplane than the traditional busstructure, and also requires IC packages with many pins.

The logical OR circuit 51 may be implemented in any number of manners.For example, wide fan-in pseudo-NMOS gates with four to six inputs have beensuccessful. For a chip bus bit processor 50 that process more than six signals, agate tree is generally required.

If the electrical distance from the card 49 to the chip bus 40 is less thanabout half the transition time, the signal can be unterminated, and the driver can bequite small. If the line is long enough to be terminated, it is possible to operate inthe 50 to 100 Ohm regime, rather than the sub-20 Ohm regime associated with aheavily loaded conventional bus. Note that the termination can be done bycorrectly sizing the driver transistors.

One way to use the chip bus 40 is as a drop-in replacement for a traditionalbus structure. In this mode, the chip bus provides the advantages of lower powerbecause it is easier to drive the lines, there are smaller propagation delays becausethe point-to-point wiring is not periodically loaded, and the bus topology isdecoupled from the electrical behavior. The chip bus does not suffer from themultiple reflection noise and settling delays associated with classical busimplementations.

In a second implementation of the chip bus 40, a constraint is placed on thewire lengths. It is easiest to think about this in the context of the unidirectionalimplementation with all wire lengths equal. In this case, under the assumption of




60

no clock skew, the signal duration may be set to the minimal width to ensurerecognition. The electrical distance from the card 49 to the chip bus 40 enters into

latency, but no longer influences the maximum signaling speed.

The chip bus 40 of the invention is extremely fast. Simulated chip bus 40designs have shown bit rates of 2.4 Gbits/sec per line. The delay through the chipbus 40 is only 330 pS. A portion of the chip bus's speed is attributable to the factthat input signals to the bus 40 can be pipelined, four input signals A, B, C, and Dare respectively carried by chip bus input lines 44A, 44B, 44C, and 44D at timeTo. The pulse width of each signal is equivalent to the pulse width of the signalclock, shown as Tp.. 6B illustrates the progression of the four input signals after aclock cycle, that is, at time T=To Tp. the same signals on the chip bus output lines46A, 46B, 46C, and 46D. The signals appear on the chip bus output lines at a timeT=To nTp, where n is the number of clock cycles required to drive the signalsthrough the chip bus 40. that it is possible to have an input signal to the bus and anoutput signal from the bus every cycle. This pipelining capability results inextremely high processing speeds that are not possible with traditional busarchitectures.

Current processor designs have about twenty gates between latches. The sumof the setup and hold time of the latches is around 10% of the cycle time, or two

unit gate delays. The bus chip can be modeled as a pure delay, it doesn't changepulse width. This implies that up to ten bus signals could be stacked in oneprocessor cycle. If some margin is allowed for timing tolerances, a practical limitnear eight transactions per cycle might be obtained with very careful design.

The real limitations on the speed of the system are clock skew and bit-to-bitskew within a single package. Careful design of the wiring on the backplane 22allows wire skew to be reduced to below all other skews in the system. Clock skewcan be kept low by using self-compensating clock drivers.

If the bus is wide enough that more than one package 70 is required, twoelements will contribute to the bit-to-bit skew. One is the difference in averagetotal delay between the parts, and the other is the spread in delay between the pinswithin a single part. The traditional way of coping with part-to-part variations is tobin the parts. Note that this does not cause yield loss, it just requires that anyparticular board be populated with parts that have the same total delay dash-number. It is also possible to build active compensation circuits into the parts to




61

force the average delay to match, for example, a reference delay printed on theboard.

Bit-to-bit skew within a part is controlled by a combination of the vendor'sprocess control, and what special efforts were taken during the design and layoutof the chip to minimize the sensitivity of the part to random variations in theprocessing.

The clock protocol also influences the effect of delay variations. If thesignalling is source synchronized on a chip-by-chip basis, that is each group of bitsthat is carried by a single bus chip carries its own clock, the sensitivity to interchipdelay variations may be minimized. This does add some complexity to the receiverdesign to ensure that all the bit groups are correctly realigned. The source clockmay be used to provide the reference input for delay lock loops to compensatethese errors. Errors in arrival time of signals at the inputs of a single bus chip candirectly subtract from the signal pulse width.

Each signal can carry its own clock, for example, by using Manchestercoding as the synchronization protocol. Any method that carries the clock on thesame line as the signal will pay some overhead in bandwidth and latency. Oneadvantage of using a self clocking protocol is that all inputs to a chip bus can be

individually actively delay compensated by choosing one of the inputs as areference for all the others. This can be made to work both for the chip buses andthe system chips, and provides a global clock synchronization method as a sideeffect of minimizing skew errors in the interconnect.

The available bandwidth in a chip bus system in accordance with theinvention can be reduced by parasitic reactances in the IC packages and in theinterconnect. Reactances in the package can reduce the bandwidth by twomechanisms: low pass filtering the signals and introducing noise. If the ICpackages are designed with close attention to the parasitics, it is possible to resolvethese problems. For example, a flip chip circuit can be used for very low seriesinductance, and to maintain a controlled impedance right up to the pads.

Simultaneous switching noise caused by inductance in the ground returnpath in the chip packages (ground bounce) and crosstalk between signals alsointroduce uncertainty in when the transitions are recognized. The same measuresthat are used to reduce package parasitics to avoid bandwidth reduction will alsohelp reduce these noise sources.




62

The ability to swap out boards in servers without powering down the system

or stopping the clock has become a requirement for new server designs. This israther difficult to implement using a traditional bus structure because both theinsertion and removal of the board produces electrical transients on the backplane.The chip bus of the invention provides an elegant solution to this problem. Adisable pin for each port on a bus chip can be provided to force the correspondingport into an idle state where the output is not driven and the input is ignored. Thisisolates a board being removed or inserted from the bus. The control of thesedisable signals can be derived from variable length fingers on the backplaneconnectors.

It is possible to provide small state machines on the chip bus to performarbitration or protocol functions. If protocol or arbitration logic is embedded in thechip bus, two problems arise. The first is that the gate depth rises beyond theminimum needed to accommodate the fanout. The second is that connections arerequired between the chip buses to coordinate control. Both factors increaselatency and reduce bandwidth. These problems may be reduced by pipelining thebus protocol and arbitration. The pipelining can be done through central ordistributed arbitration. In the case of central arbitration, a special arbiter chip isplaced on the backplane near the chip buses. To match the performance of the chip

buses, all connections to the arbiter must be point-to-point, and the length matchedto the signal lines. The shortest pipeline sequence is: request, resolve, grant,transfer. Distributed arbitration can be accomplished by running the same statemachine on each of the devices present on the bus. This usually requires dedicatingN request lines, where N is the number of devices. Pipelining the arbitration is stillrequired.

When state machines or other intelligence is not used, the chip bus islogically equivalent to passive wires on a backplane. This allows them to run at themaximum speed that the technology will support and permits bit-slicing the bus toaccommodate real world packaging constraints.

In the bi-directional communication line implementation, a package wouldrequire control pins to control the signal direction. A package would typicallyrequire one power or ground pin per two signal pins. Standard pin versus speedtradeoffs may be made when designing a package 70. The chip bus 40 of theinvention may be clocked at up to eight times the processor clock speed.




63

To control clock skew it may be advantageous to use a commerciallyavailable clock distribution chip. Such chips compensate for skew by measuring

the phase of a reflected signal relative to an internal reference clock.

The pin count required for a package 70 can be reduced as far as desirableby relying upon multiple chips. A 70 bit bus supporting 16 cards can beimplemented with 8 chip buses 40 of the type that use separate input and outputlines. Each chip bus 40 could be formed in a 432 pin package with the processorand bus running at the same clock speed. This has the advantage of requiring nocontrol signals to the bus chips and can provide more bandwidth if the bus wererun at a multiple of the processor clock.

An optical bus bit processor 80 in accordance with another embodiment ofthe invention. From a logic standpoint, the optical bus bit processor 80 operates inthe same manner as the previously described bus bit processor 50. The differencebetween the devices is that the optical bus bit processor 80 processes light signals.That is, a digital high state is represented by a light pulse, while a digital low stateis represented by an absence of a light pulse. The optical bus bit processor 80performs a logical OR operation on light signals.

The optical bus bit processor, also called a star coupler, 80 includes a set of

N input optical fibers 82A-82N carry input signals to an optical fiber link 84,shown in this embodiment as a fiber ring. A set of N output optical fibers 86A-86Ncarry output signals. Consistent with previous embodiments of the invention, if asingle input signal is digitally ON (equivalent to a light pulse), then the fiber link84 will cause each output optical fiber 86A-86N to carry a digitally ON signal.Thus, the apparatus 80 performs a logical OR operation, consistent with theprevious embodiments of the invention.

An optical computer bus 90 in accordance with an embodiment of theinvention. The optical computer bus 90 includes a stack of optical bus bitprocessors 80. Each optical bus bit processor 80 is formed on a substrate 92. A setof substrates are combined to form a stack 93. A stack of eight substrates toprovide a system that transmits eight bit words. A stack structure is a convenientconfiguration, other physical configurations are also possible.

A system card (not shown) is attached to a connector 96. The system cardmay be memory card, a local input/output card, a network input/output card,graphics card, etc. Thus, each system card is typically in the form of a computer




64

sub-system. Alternately, all computer sub-systems can be contained on a singlecard.

A set of signal lines 98 electrically link the connector 96 to an input opticalconnector card 100. The input optical connector card 100 includes a set of signaldrivers 102. The signal drivers 102 process electrical signals from the connector 96and convert them into appropriate drive signals for an array of light producingdevices 104. The array 104 is preferably implemented as a set of Vertical CavitySurface Emitting Lasers (VCSELs). Each VCSEL is optically connected to a singleinput fiber 82 of a single substrate 92. Thus, the system card attached to connector96 can apply eight separate signals to the optical computer bus 90. In particular,each signal of the eight separate signals is applied to the first input fiber 82 of eachoptical bus bit processor 80 of the stack 93 of optical bus bit processors formingthe optical computer bus 90. Alternately, the output of the array 104 may beapplied to a fiber optic ribbon cable, which is connected to the first input fiber 82of each optical bus bit processor 80 of the stack 93.

It should be appreciated that the embodiment allows for three additionalinput connectors to be respectively linked to the three input fibers 82B, 82C, and82D. Thus, in the example system , four elements are connected to the optical bus90, thus forming a four word bus that processes four eight bit words. Given this

configuration, the optical computer bus 90 can be considered a "two-dimensionalbus".

Each optical bus bit processor 80 is capable of processing four input bits andproducing four output bits. The output signals on output fibers 86A-86D areapplied to a set of output connector cards. For the sake of simplicity, a singleoutput optical connector card 110. The output optical connector card 110 includesa light signal receiver array 112, which may be implemented using a set ofphotodiodes or a set of VCSELs. An array of drivers 114 is connected to the lightsignal receiver array 112. The array of drivers 114 generates a set of electricalsignals that are applied to the connector 96 for processing in a standard manner.

The optical bus bit processor 80 is positioned on a substrate 92. The opticalbus bit processor 80 may also be implemented as a fused coupler or in free space.In the free space embodiment, an individual light source of a set of light sources ata sending end is capable of transmitting a signal through space. The single signalgenerates an output signal at a set of light receiving sources. The light sources and




65

light receiving sources are controlled by dynamic bandwidth allocators of the typedescribed below.

The optical bus bit processor 80 may also be implemented aslithographically produced polymer or silica planar waveguides. Preferably, theoutput optical connector card 110 is DC coupled and has fast recovery fromoverload. This is important because several input pulses may overlap at thereceiver array 112 during system start-up or during a fault.

Those skilled in the art will appreciate that the optical computer bus 90 ofthe invention is very fast. The optical and electrical implementations of theinvention allow the bus clock to operate at a multiple of the system clock. Thisoperation illustrates a four word bus which makes connections to four input opticalconnector cards 100, although only one is shown for the sake of simplicity.Similarly, the bus is connected to four output optical connector card 110, althoughonly one is shown. Relying upon this example, if the clock for the optical bus 90 isoperated at four times the speed of the system clock, then four bus time slots exist.Each of the four connector cards can transmit an eight bit data word in a bus timeslot.

Waveform 120 illustrates the system clock signal. Waveform 122 illustrates

the bus clock signal, which is four times faster than the system clock signal.Waveform 124 illustrates that a first input optical connector card transmits data (aneight bit word in this example) during the first bus time slot, which corresponds tothe first bus clock signal cycle. The second input optical connector card transmitsdata during the second bus clock cycle, the third input optical connector cardtransmits data during the third bus clock cycle, and the fourth input opticalconnector card transmits data during the fourth bus clock cycle, during one systemclock cycle and four bus clock cycles (bus time slots), each optical connector cardis allowed to transmit data on the system bus. This process may be repeated forsubsequent clock cycles.

A flat or uniform allocation of optical bus bandwidth. Observe that duringthe course of a single computer system clock signal cycle, every node gets to sendone message in its own bus time slot. This functionality is a superset of a crossbarbecause every node can observe all of the transmissions. This allows for theimplementation of snoopy cache coherence methods that have a lower overheadthan the directory based methods required for a classic crossbar.




66

The invention can also be implemented by dynamically allocating opticalbus bandwidth. During the first system clock cycle, all available bus bandwidth is

assigned to the first input optical connector card, as illustrated with waveform 132.This allocation may be viewed as a dedicated division of bus bandwidth resourcesto a single computer sub-system. During the second system clock cycle, the busbandwidth is split between the second input optical connector card and the thirdinput optical connector card, as respectively shown with waveforms 134 and 136.This division of bandwidth resources may be viewed as being random. In the finalsystem clock cycle, the bus bandwidth is assigned to the fourth input opticalconnector card, as shown with waveform 138. Thus for each system clock cyclethe bus bandwidth can be divided among the input nodes in any number of ways.This feature allows nodes with heavy traffic to dominate the bus bandwidth forimproved overall system performance.

One technique for implementing the foregoing functionality. An inputoptical connector card 100 of the type. The input to the card 100 is from the signallines 98, which are linked to the connector 96. The output from the card 100 isapplied to the optical bus 90. As previously discussed, the input optical connectorcard 100 includes a laser array 104 and a set of signal drivers 102. In accordancewith an embodiment of the invention, the signal drivers 102 may be implementedas a set of transmitting dynamic bandwidth allocators 150A-150N. The signal

drivers 102 are connected to a transmission mask register array 148, which is anarray of registers, with each register storing a transmission mask signal indicatingwhich computer sub-system signals are to be transmitted during a bus clock cycle.The signal drivers 102 are also attached to a buffer array 144, which stores datafrom the signal lines 98. In particular, each buffer in the buffer array 144 storesdata for a corresponding transmitting dynamic bandwidth allocator 150.

A transmitting dynamic bandwidth allocator 150, in accordance with anembodiment of the invention. The circuit 150 includes a transmission circuit 152which receives a single data bit from the buffer array 144, a transmission mask bitfrom the transmission mask register array 148, and a bus clock signal. The data bitis applied to an input node of a flip-flop 160. The flip-flop 160 is enabled if the busclock signal is high and the transmission mask bit is set to a digital high value. Inthis case, the logical AND gate 162 generates a digital high value, or flip-flopenable signal, to enable the flip-flop 160. Thus, it can be appreciated that thetransmission mask bit controls the output from the transmission circuit 152. Theoutput from the transmission circuit 152 is used as a drive signal for the laser array




67

104. Preferably, a deskew circuit 154 and a drive circuit 156 are used at the outputend of the transmission circuit 152.

An output optical connector card 110 has a similar configuration to that ofthe input optical connector card 100 . In particular, an output optical connectorcard 110 has a receiver array 112 connected to a driver array 114, which includes aset of receiving dynamic bandwidth allocators, which are controlled by receivemask signals stored in receive mask registers. The output from the driver array 114is applied to a buffer array.

The processing of a bit signal between a transmitting dynamic bandwidthallocator 150 of an input optical connector card 100 and a receiving dynamicbandwidth allocator 170 of an output optical connector card 110. As discussedabove, the transmitting dynamic bandwidth allocator 150 includes a transmissioncircuit 152, a transmission signal mask register 148A of the transmission maskarray 148, a deskew circuit 154 and a driver 156. Similarly, the receiving dynamicbandwidth allocator 170 includes a driver 172 and a receiver circuit 174, which iscontrolled by a receiving mask bit from the receiving mask register 176 of areceiving mask register array (not shown). The receiver circuit 174 operates in thesame manner as the transmission circuit 152. The receiving dynamic bandwidthallocator 170 also includes a skew compare circuit 178, which identifies skew

between the received signal form the optical bus and the bus clock signal. Theskew value is then sent to the deskew circuit 154 of the transmitting dynamicbandwidth allocator 150 so that future signals are sent with reduced skew.

It should be appreciated that the disclosed dynamic bandwidth allocationconcept of the invention is also applicable to the disclosed digital gate computerbus. When implemented in reference to the digital gate computer bus embodimentof the invention, light array transmitters 104 and receivers 112 are omitted.

The foregoing descriptions of specific embodiments of the present inventionare presented for purposes of illustration and description. They are not intended tobe exhaustive or to limit the invention to the precise forms disclosed, obviouslymany modifications and variations are possible in view of the above teachings. Forexample, a traditional backplane 22, connectors 48, and cards 49 need not be used.The embodiments were chosen and described in order to best explain the principlesof the invention and its practical applications, to thereby enable others skilled inthe art to best utilize the invention and various embodiments with variousmodifications as are suited to the particular use contemplated.




68

APPLICATIONS :-

1. High speed communications :The rapid growth ofinternet,expanding at almost15% per month, demands faster speedsand larger bandwidth than electronic circuits can provide. Terabitsspeeds are needed to accommodate the growth rate of internetsince in optical computers data is transmitted at the speed of lightwhich is of the order of 3 10*8 m/sec hence terabit speedsareattainable.

2. Optical crossbar interconnects are used in asynchronous transfermodes and shared memory multiprocessor systems.

3. Process satellite data.




69

MERITS :-

1. Optical computing is at least 1000 to 100000 times faster thantoday’s silicon machines.

2. Optical storage will provide an extremely optimized way to storedata, with space requirements far lesser than today’s silicon chips.

3. Super fast searches through databases.

4. No short circuits, light beam can cross each other withoutinterfering with each other’s data.

5.

Light beams can travel in parallel and no limit to number ofpackets that can travel in the photonic circuits.6. Optical computer removes the bottleneck in the present day

communication system

1




70

DRAWBACKS

1. Today’s materials require much high power to work in consumerproducts, coming up with the right materials may take five years ormore.

2. Optical computing using a coherent source is simple to computeand understand, but it has many drawbacks like any imperfectionsor dust on the optical components will create unwantedinterference pattern due to scattering effects. Incoherent processingon the other hand cannot store phase information.

17






72

FUTURE TRENDS

The Ministry of Information Technology has initiated a photonicdevelopment program. Under this program some funded projects are continuingin fiber optic high-speed network systems. Research is going on for developingnew laser diodes, photodetectors, and nonlinear material studies for faster switches.Research efforts on nanoparticle thin film or layer studies for display devices arealso in progress. At the Indian Institute of Technology (IIT), Mumbai, efforts are inprogress to generate a white light source from a diodecase based fiber amplifiersystem in order to provide WDM communication channels.




REFERENCES

1. Debabrata Goswami , “ article on optical computing, optical componentsand storage systems,” Resonance- Journal of science education pp:56-71

July 2003

2. Hossin Abdeldayem,Donald. O.Frazier, Mark.S.Paley and William.K,

“Recent advances in photonic devices for optical computing,”

science.nasa.gov Nov 2001

3. Mc Aulay,Alastair.D , “Optical computer architectures and the application

of optical concepts to next generation computers”

4. John M Senior , “Optical fiber communications –principles and practice”

5. Mitsuo Fukuda “Optical semiconductor devices”

6. www.sciam.com

7. www.msfc.com

Documents

Opticalcomputers.pdf