Impression of Photonics West

Many years ago, around the time of the “dot-com” bubble, photonics had a similar bubble popping episode which in many ways was worse than the dot-com one.

The reason it was worse was that many companies at that time had real sales, profits, factories around the world and employed many thousands and thousands of people.

Going into the bubble Finisar was one of the smaller companies but we had been consistently profitable since our founding in 1988.  And our drop in sales while strong (about 30%) was much less than the biggest companies in the space which dropped 90%+ or simply went out of business completely.

One of the problems at that time was that people felt that “photonics” or the engineering discipline of the precise making, manipulating, amplifying and such with photons was the next wave to follow the “electronics” wave that had been so successful for venture investors and startup founders for the past 20-30 years.  Beyond similar sounding names there was promise in the late 1990s.  Fibers began to carry information at 2.5 and 10 Gb/s over distances of 1000s of kilometers using optical amplifiers, dispersion could be managed and corrected precisely, large numbers of wavelengths could be place onto a single fiber pushing the information carrying capabilities even further … and all these wavelengths could be amplified and dispersion corrected in parallel and as optical signals without conversion back to electron-based information.  Magic seemed everywhere in the field.  And to many people at the time, especially Venture investors, the world looked like this.

What we thought

 

Unfortunately this was not true.

What was true emerged over the next 5-8 years after the bubble burst.  And then we found out the world needs did not call for photonics to be so important.  So then the real world looked like this.

What was true

 

In reality, the most important trend coming to communications was to be wireless radio based communications that we see today as 3G/4G wireless phones, WiFi for LANs and Bluetooth for personal devices, etc.  It is technology that touches us in so many ways all the time.  And then we underestimated the resillance of copper based communications – on circuit boards, for inter and intra cabinet communications, for cable television and internet services, and even the upgrading of our phone lines to DSL.  So instead of being the whole square with electron based comms being marginalized and becoming insignificant, phontonics really only continued to dominate in long haul telecom and then it moved into data centers and buildings for datacom.  [Datacom is where Finisar got its start, today it dominates both spaces of photonic communications.]

Unfortunately so much money was invested that the field continued to be over funded for the next decade and consolidation took longer than everyone thought and many companies have very long periods of losses and shareholder disappointment.

Now lets move into the present day.  This week San Francisco hosted the Photonics West conference and exhibition.  It allowed 1000s of companies to show their products, to hear talks on various aspects of photonics technology and so forth.

And what I saw was that so much of the investment money has poured into optical fiber communications but that is beginning to change.  We are seeing the maturation of that field of applications for photonics but the fields of sensing are just beginning to find their strides.  And photonic engineering and sensing of the world around us – for self driving cars, more accurate medical diagnostics, better ways to make, manage, store and transfer energy are all being explored.  And these can bring a renassiance to the field and an even brighter future.

But we must not over invest in fields which have been good in the past.  As those mature they are less interesting for venture like investments and more the domains of larger companies that can invest the large sums needed to make smaller incremental improvements.

 

Posted in Essays, Investing, Optical Technology, Personal Stories

Abundance and Exponential Growth

I am reading a book entitled Abundance.    Here is a synopsis of the book –

“We will soon be able to meet and exceed the basic needs of every man, woman and child on the planet. Abundance for all is within our grasp. This bold, contrarian view, backed up by exhaustive research, introduces our near-term future, where exponentially growing technologies and three other powerful forces are conspiring to better the lives of billions. An antidote to pessimism by tech entrepreneur turned philanthropist, Peter H. Diamandis and award-winning science writer Steven Kotler.”

This is clearly a vision worth understanding.

And it is a vision supported by many different elements of our technologically based society.

But the foundational argument for the book is that technology and the progress it brings is essentially growing at exponential rates.  This compounding is manifest in many areas –

  • computing power grows along with Moore’s law so that every 18 months or so computers get 2x as fast or capable and you pay the same price
  • disk drive storage expands at rates also aligned with Moore’s law
  • shipped fiber optic bandwidth has been expanding at a rate of 10x every 5 years for more than 25 years

There are many more examples.

Similar facts helped NASA in the early 1950s feel it was possible to land a man on the moon in ~15 years even though no human had ever left the earth beyond flying in an airplane!

Perhaps the leading futurist today is Ray Kurzweil.  He published a series of books that have proven uncannily accurate – The Age of Intelligent Machines, The Age of Spiritual Machines, The Singularity is Near.

Clearly if you look around you there have been inventions or products introduced in each of the last 4 decades that never existed and before the end of the decade they were pervasively owned by many people throughout the first world – microwave ovens, VCRs, CD players, DVD movies, Video Games, Internet, Google, FaceBook.

Clearly some things that looked impossible in the past now are possible due to exponential growth in technology.

But …

Such growth does not apply everywhere.  It was predicted in the early 1960s that you would be soon flying on supersonic aircraft for intercontinental travel.  However as Boeing has introduced the 707, 727, 737, 747, 757, 767, 777 and now the 787.  The speed of the aircraft has not increased exponentially as earlier predicted or hoped.  In general it has not grown at all.  Why?  Because it was not practical to operate aircraft at such speeds economically.

Are there other such examples that are today not generally recognized?  Sure!

Optical fiber cable today carries about 1 Tb/s typically using individual channels of 10 or 40 Gb/s and then using multiple wavelengths or different colors of light on the same fiber.  But it is not clear at all that the growth of the last 25 years can be sustained.  Why?  Because optical fiber has bandwidth that is limited by Shannon’s law and other electrical and optical physics.  And in that technology and physics we are reaching the limits of what is possible.  So the growth will slow now over the next decade.  There can still be an enriching of the technology platform and clearly many more fibers can be put into a cable but the actual information carried on a single fiber is reaching its limit just at the amount of information carried on copper wires reached its limit about 20 years ago as modem technology matured.

Other examples are around as well –

  • automobiles are essentially the same as there were in the 1940s: same top speed, same basic construction
  • shipping is essentially the same
  • building size and construction technology has not changed much: we have today only slightly taller buildings than the empire state building where the design was complete and construction started in 1930!

Just to name a few.

It is easy to get caught up in this fervor but in reality exponential expansion in capability never lasts very long before running into limitations imposed by physics or economics.

So be careful when you read how these trends of the past 20 – 40 years are going to solve the problems of the next 40.

 

Posted in Essays, Green Perspectives, Investing, Optical Technology, Spiritual Threads

Amazing Korean Internet Access

Tonight from my hotel (The Grand Intercontinental Seoul Parnae) I achieved by far the fastest internet access I have ever seen.  Totally amazing.

Here is the graphic showing the actual results using www.speedtest.net –

 Korean Internet Speed Test - Click to Enlarge

Korean Internet Speed Test – Click to Enlarge

This is what fiber optic communications should be doing for us all.

Service providers have become very proficient about quoting speeds that they never deliver.  This is so totally the opposite.

Frickin’ Amazing.

Posted in Optical Technology, Personal Stories

Old Warrior Passes

Bruce Campbell, one of the founders of Raynet – a very early fiber to the home company – died earlier today.

Bruce was one of the most classical entrepreneurs I have ever known.  Irascible, difficult, intellectually honest, enthusiastic, outside the box thinker – all of these describe him.

But perhaps a couple of stories will do it best.

The first comes from a time when he worked for Raychem and there was a wall that separated their offices from the labs where they would do work.  Turns out there was a long walk from the office to the lab but only a wall geographically speaking.  So Bruce asked the corporate buildings and grounds group if they could put in a door for him and others of his group to be able to go to the lab.  The response was that it was possible but they were currently very busy and so it would take 6 months or more to get to it.  In further probing he found out that repairs were handled on a priority basis and often took a day or so.  So the next day Bruce brought in a large hand held circle saw and proceeded to cut a near door size hole in the wall.  Wham the cut out fell down and moments later Bruce called to report that they needed a repair.

The door was completed the next day.

Some years later Bruce was instrumental in getting Raynet funded by Bell South.  As the money came in, Raychem spun out Raynet and brought over senior management that had been with Raychem as well.  The ideas for Raynet had come from Bruce’s group but now this was going to be managed.  Anyway, I came over to Raynet around that same time though I was had not been with Raychem for very long and I was certainly not a senior manager.  Eventually we had to deliver system that would be the prototype of a unique way to do fiber to the home.  And this system required that we build “taps” that could extract and inject light into the fiber without breaking it.  This can be done by bending the fiber carefully and Bruce was quick to see the potential for this approach.  Millions were raised and over time the approach was eventually abandoned.  But as a venture, it was classic.  Start with an idea and go find a customer.  Listen to the customer and innovate around their needs.

During my time at Raynet, Bruce and I clashed many times and I had to lick my wounds but in the end and for many years after, we remained friends.

Later he started a company called Wavelengths Lasers and Finisar did the engineering for the early products of that company.  One time we killed a $10,000 laser and Bruce and his team came over to hold a funeral for it and then they made it into a earring for one of the gals.   Always able to see the humor even in pretty lousy situations.

So if there is a heaven (or hell) you better watch out because there are likely to be new doors coming your way pretty soon.

Farewell Bruce.

Posted in Essays, Optical Technology, Personal Stories, Spiritual Threads

Big Picture

Every once in a while, there is something on the web that knocks my socks off.

As many of the readers of this Weblog know, I was once an astronomer; my PhD is in that field from the University of Virginia. During that time, I studied the structure of spiral galaxies (1), (2), (3) like our own Milky Way.

Our galaxy is a disk of stars with a very weak, perhaps only 5% mass enhancement in a spiral pattern. Gravitational disks are susceptible to such modes but such a mass enhancement if only that would not give the spirals such a striking and beautiful appearance. What makes the spiral pattern so very visible is the fact that with only such a small mass variation across the disk of the galaxy, the gas in the galaxy between the stars gets whipped into a spiral shock wave pattern. As the gas goes through this shock wave clumps of gas get compressed sufficiently that single and groups of stars are born in a spiral pattern. Most of the new star formation in these galaxies happens this way.

Stars that are born this way can have various sized from 2-3x bigger than our sun to perhaps only 60% the size of our sun. Small stars burn very dimly,are more reddish in color and live VERY long times compared to our sun but big stars burn blue, are very bright and live very short times compared to our sun. The largest stars have lives of perhaps only 1,000,000 years compared to the expected life of our sun which is 10,000,000,000 years. So these young bright blue stars are essentially born and die in almost the same place.

In a spiral pattern. So it is these big bright blue stars that give the spiral galaxies their striking visual appearance.

Seeing our own galaxy’s spiral pattern is tough because we live (our sun’s position) in the galactic plane so we see the spiral edge on and not from the face.

One more point about the blue stars … when they die, the blow up in an event known as a supernova and for a short period of time during the explosion they burn with more light than all of the 200,000,000,000 other stars in the galaxy. Wow. After the supernova dims down that explosive event can be seen as an expanding bubble of gas in our galaxy.  As these bubbles cool they can be quite visible in the infrared portions of the light spectrum.  And that is the end of this lecture.

Now for the striking picture. If you got here –> http://www.chromoscope.net/

This shows what we can see from earth of our galaxy in various colors of light.  If you move the slider from the starting “visual” position to Hydrogen Alpha you can see the gas in our galaxy and then you see many gigantic bubbles.  Each of these is the product of a supernova that happened some millions of years ago!

Enjoy!

Posted in Essays, Optical Technology, Personal Stories

Our Abstract World

We live in a world that is increasingly different from the one in which our ancestors received their genetic programming. Today, our minds constantly use abstractions to negotiate our way through the world.

An example of this is the process we use to steer an automobile.

(A fun fact to know is that when automobiles were first introduced, there was great debate on how to steer them. Some early cars had reins similar to horses, others had a tiller like small sailing boats, and still others used a wheel like those found on larger ships.)

Fun facts aside, when you drive a car today with a steering wheel, you are using a solid object to perform an abstraction, all within the blink of an eye. When you turn the steering wheel, you are not only turning the wheel, you are turning the wheels of the car and guiding the car in the direction you wish to go. The small movements you make are amplified and strengthened through a power steering mechanism and then onto some rack and pinion structure to move the wheels that contact the road and keep the vehicle on the right track.

Nevertheless, your mind translates all of this for you so that you can turn the wheel in various motions and you don’t really have to think about the concrete business of guiding the car. Your mind abstracts the motion of the steering wheel into the motion of the car.

Now you might be asking yourself: “Where are we going with all this? What does this have to do with Finisar primary focus in designing, manufacturing and marketing the best possible optical subsystems?”

It is true that Finisar designs, manufactures and markets optical subsystems, and we do that by working in abstractions. Abstractions are really just constructs of our minds, despite the fact that we often accept them as reality and don’t question them sufficiently.

Let me describe how we at Finisar use abstractions as tools, and present some arguments on why we should test and validate common abstracts more carefully.

How One Abstraction Begets Another

Finisar’s optical subsystems include optical transmitters/receivers (or transceivers) that either transform electrical signals into optical signals (transmitters) or optical signals into electrical currents (receivers). These devices are usually digital in nature and operate at speeds above one billion bits per second or one Gigabit per second (Gb/s).

We observe these optical or electrical signals using a device called an oscilloscope. An oscilloscope produces an image on a screen that looks like this:

The picture above is what we call an “eye diagram,” which is simply a picture of the bits that are sent on the fiber optic glass strand. In this picture, you see the results of many different bits each overlaid on top of one another.

How this picture is interpreted and how we use it represents the crux of what Finisar does and why Finisar’s products are known to produce some of the very best eye diagrams in the world.

Measuring the Eye

The figure below shows how we create the eye pattern. Imagine an idealized string of bits comprised in a 1 – 0 – 1 pattern (the first diagram in the figure).

In an optical fiber, a laser emits a bright light to represent a ‘1′ and sends no light (the fiber is dark) for a ‘0.’ Within the fiber the light turns on and off to form the sequence of 1, 0, 1, and so on. The laser cannot turn on and off instantaneously, so there are some instants in time where the light is growing brighter or fainter as it goes on or off. This is the transition point between a 0 to a 1 or from a 1 to a 0.

If we use the oscilloscope with a careful “trigger,” we can see what these bits look like, which is shown in the third diagram in the figure above. Here we see the oscilloscope trying to “trace” out the bit to measure the light as it emerges from the fiber.

If we use the oscilloscope to look at the light emerging from the fiber over many bits and if the bits are coming out are random, we sometimes will overlay 1s on top of 0s and vice-versa. When this happens we get an eye pattern, shown in the fourth diagram of the figure above. Notice that this diagram is just like the one above it with the pattern sometimes inverted.

Look again at the LCD oscilloscope eye pattern (below) and you will see how this really looks to us as we measure a Finisar transceiver.

Abstractions As Building Blocks

Here’s a step by step description of how we link abstractions together to create the eye patterns in an oscilloscope:

  1. The end of the fiber is connected to a photodiode that converts the light to a current of electricity.
  2. This current is transformed to a voltage by means of a transimpedance amplifier.
  3. The amplifier provides gain whereby the signal level is increased typically by more than a factor of 1,000.
  4. The output of the amplifier is presented to the oscilloscope which “samples” the voltage periodically. The voltage is turned into a number by means of an analog to digital converter.
  5. An internal clock in the oscilloscope determines how often the voltage is sampled.
  6. As the samples are digitized, the output is fed into computer memory for storage.
  7. A computer takes the data out of memory and displays it on the LCD display screen.
  8. Data taken at different times and voltages is painted across the screen one dot at a time. Thousands of dots are required to make an eye pattern.
  9. Data is typically taken from one point every few hundred bits. This means that the actual data on the eye pattern represents single dots per bit taken every few hundred bits. Thus, the data is taken over millions and billion of bits.

This is pretty confusing, but the point here is every step must be precisely correct. The eye pattern will be severely distorted if any of the following problems occur:

  1. The end of the fiber is not well connected to the photodiode so that some of the light is lost by not hitting the photodiode.
  2. The photodiode transforms the light from the fiber into a current but does not do this linearly so the light represented by the current is inaccurate.
  3. The amplifier gain varies over time, temperature or something else or is not linear so that the gain of 1000 is sometimes 900 or 1500.
  4. The analog to digital converter is inaccurate.
  5. The clock in the oscilloscope is inaccurate.
  6. The samples are not stored properly in the computer memory.
  7. The software in the computer is not programmed correctly.

The changing of the light from faint to bright or the reverse takes about 100 trillionths of a second. It’s extremely fast. No ‘blinks of an eye’ or other analogies will work here. It is just plain fast, faster than anyone can possibly imagine.

You get the idea. Every abstraction we make introduces a potential source of error. If any one of these is wrong it is difficult to determine which one it is.

More Fun with Abstractions

So now you say, “How does this affect me”? I don’t create eye patterns on my job and driving a car doesn’t seem so abstract when I’m actually in the car.

There are other abstractions everywhere. Here are some fun historical ones for you to think about:

  • Around 1500 it was determined by Copernicus, Galileo and Columbus that the earth was a sphere and that it orbited about the sun (though there was some discomfort on this for most involved). Did you know that it took another 400 years for us to guess that our sun was NOT at the center of the universe? (This happened in the early 1900s when we observed certain globular star clusters and found that they orbited some 50,000 light years distant from us in the center of our Milky Way Galaxy.) We could not easily break free from our abstraction that mankind was the center of creation.
  • In both the old and new testaments of the Bible it is written that human generation occurs when a man plants his seed in a woman. Putting in modern terms, the ancients thought that men were entirely responsible for generation, and women were the “dirt” into which the seed was sown. Perhaps this explains why men were dominant in many cultures for thousands of years (Even today we still speak of men sowing their wild oats). We were stuck in an abstraction that affected much of how men and women related to each other. Today we know that men and women both contribute genetic material to create new little human beings. This corrected abstraction undoubtedly helps us have a more equal partnership between men and women.
  • During the energy crisis of the 1970s, we abstracted our energy supply such that many were convinced that we would “run out” very soon. That did not happen, and within a few years the world was again awash in oil. Now we are trying to understand global warming and greenhouse gases. There are abundant abstractions in this area, and we do not understand them very well at all.
  • For a long time, computers, memories, disks and such were sized for their expected usages. To overbuild a computer was really expensive. Over time, components like memory have dropped vastly in cost, but it took time for us to understand that the potential uses for memory were virtually unlimited. Early PCs could only support 640K of memory. Such hardware limitations were specifically designed into even the Microsoft software architecture, and we allowed this abstraction to limit us for several years.

Using Abstractions to Foster Innovation and Success

Sometimes we embellish our view of the world by creating abstractions that have a worth that is not real. Other times, unimaginative abstractions limit our ability to invent or progress as a society.

Abstractions are everywhere because they are the tools that humans use to form mental images of their changing and evolving world. They are products of our minds, whether we are driving a car, trying to understand human conception or trying to measure the transmission of light as it travels down an optical fiber.

Today, because of the complexity of our world, most abstractions come in layers, and the failure of any one layer can alter your view of what really is there. Be aware of the abstractions you accept. When something seems not to fit or work, look where you have abstracted the world from what reality is there physically. Then go look at each abstraction between what is wrong and what is reality. It is such investigation and questioning the moves us forward.

(Original essay published in September, 2001 when at Finisar)

Posted in Optical Technology, Spiritual Threads

Optical Networks and Earth Limits

George Gilder is a widely published columnist and pundit. His thoughts about man’s technological future are widely respected. Lately, he has been evangelizing a vision of an all-optical Internet, plentiful bandwidth and the demise of many traditional Internet technologies like Cisco routers. Some of this vision is underpinned with assumptions that are not applicable and as clarity in assumptions is critical for any future vision, I’d like to use this forum to work through these.

First of all let me quote Gilder directly so that my jumble of opinions do not dilute his message:

“… the Internet is a computer on a planet. Like a computer on a motherboard, it faces sever problems of memory access. The Internet communications depend on ingenious hierarchical memory management, analogous to a computer’s registers, buffers and latches, its three tiers of speculative caches, its bulk troves of archives, its garbage management systems to filter and weed out redundant or dated data, and it direct-memory access controllers to bypass congested nodes.

In a world of bandwidth abundance, an ever-increasing share of roundtrip delay for a message is attributable to speed of light latency. No matter how capacious the transmission pipes, how large in numbers of bits per second the data stream, the first bit in the message cannot move from source to terminal any faster than light speed allows, plus the time waiting in queues and buffers at all the switches or other nodes along the way.”

He then makes the case for strong caching and dispersed assets … here are some of the supporting assertions:

“Even with no hops or other delays, the light-speed limit alone means that Internet users outside the North American continent are at lest 200 milliseconds away from the vast majority of websites.” (80% of web hosts are in continental USA)

“To fetch a web object using the Internet protocol – whether a frame, image, logo, or banner – takes two to seven round trips between the end user and the Web server. With each page comprising as many as twenty-five objects, those round-trip-speed-of-light milliseconds keep adding up even for entirely static material.”

Perhaps a little table will help us keep track of what physics foundation Gilder is operating from.

Description value units

speed of light 300000 km/sec

diameter of earth 12500 kilometers
circumference of earth 40000 kilometers

max intercontinent hop 30000 kilometers
round trip time 200 milliseconds

typ intercontinent hop 12000 kilometers
round trip time 80 milliseconds

typical US only hop 2500 kilometers
round trip time 17 milliseconds

So, his assertion of 200 ms for the worst case hop just based on speed of light arguments is well justified. The figure hinges on the following underlying assumptions worst-case distance is:

  • 50% of earth’s circumference
  • 50% longer than even this due to routing inefficiencies
  • really 2x this because a web page fetch involves a request and an answer so the distance is traversed twice.

This worst case figure is really exaggerated for the point of his article … a more typical intercontinental distance is likely to be 12000 km and then the time drops to 80 ms. And for normal intra-continental distances of 3000 km this is then 17 ms. Actually most traffic will be of this latter category since each citizen/user will most of the time be fetching information localized for their country.

But then his assertion that this is the root of all our problems does not wash. He assumes that web pages will always be transmitted as individually requested objects on a page. This assumption is weak … a much simpler solution is for HTTP and its cousins to migrate towards a single transmission request for all information and for the sender to gather together the information and send it just one time as a single group or stream. No doubt this is better for the server and the Internet infrastructure to see a larger single flow instead of so many little (25) pieces. Again, it is the multiplication of so many separate requests to assemble a single web page that causes the problem.

In a sense, such progress assumes that Internet protocols will continue to evolve … sounds reasonable. Moreover, once the streaming has started the end user will not notice the effects of distance provided there is sufficient bandwidth between the sites and buffer at the receive end to smooth out any potholes in the transmission

Which is the better approach?

Reasonably local/short distance widely dispersed caching or more aggregated page serving?

Let’s use Gilder’s assumption of 25 objects per web page for our starting place. We shall assume the cache resides at an effective distance of 250 km for a round trip time of 1.7 ms and therefore 25 objects will take 43 ms. This is substantially slower than the direct sending of the page over typical intra-continental distances of 2500 km but sent as a single object (17 ms).

The point of this is that direct bundled sending is substantially better than local caching of pages like Akamai or Inktomi can provide.

As our agents and web servers get smarter, we will find that their ability to coordinate the information flow to us so as to minimize our waits and keep their responses more in line with our normal cadence of work will remain acceptable.

So where is the problem of Internet scaling? It is primarily in the core of the network where the global computer must continue to scale to packet per second forwarding levels measured in terms of 1012 and then 1015 packet per second over the next few years. Today the biggest, baddest layer 3 switches or routers have ability to do about 108 packets per second. Collections of these cannot do any better than this because they are ultimately placed into chains and the chain is only as good as its weakest link.

The problem with Internet scaling is not with typical web pages or the migration of television or even telephones to the web; it is with distributed computing. As we seek to create larger clusters of computing, especially as clusters are formed and broken down dynamically as their user bases require this distance and light travel time strongly affect the ability to do cluster computing.

Interestingly enough the problem in trying to create a world wide computer system (like that envisioned in the movie, The Matrix) is that the messages will only propagate at the speed of light and the network in between cannot beat this limitation. Responses that require inputs or outputs from disparate points in the system will not respond in computer time but in earth time … milliseconds not nanoseconds.

(this essay was originally written in March, 2000. fortunately the physics has not changed!)

Posted in Optical Technology