In their own words “Bell Labs threw a party and everybody came” – and most journalists (and I confess that I was one of them) didn’t seem to appreciate that this was the birthday of a dry, scientific paper rather than the creation of a working device a few years later.

However, the “phoney fortieth”* did lead me, eventually, to the account of the very first laser. Ten years later I wrote an editorial column recounting the various anniversaries taking place in the summer of 2008. It was 150 years since the first transatlantic cable, 50 years since the publication of Schawlow and Townes’ famous paper on the theory of the laser, and 20 years since the two ideas – transatlantic cables and lasers – had come together in the form of the first fibre-optic cable to span the ocean.

Thanks to that editorial I met Larry Johnson, founder of US training firm The Light Brigade, who is working on a project to archive the history of optical communications. Larry wanted to set the record straight – he pointed out that Theodore Maiman was the name that everyone should remember in conjunction with the laser because he’d been first to actually build one. Larry put me in touch with Kathleen, wife of the late Dr Maiman, who died in 2007.

Kathleen met Maiman in 1984 on a flight back from the ceremony in which he was inducted into the National Inventors Hall of Fame. Now she speaks in his place at conferences and special events to celebrate the creation of the first laser. In our conversation, it became clear that she misses her late husband with the kind of intensity that time does little to diminish.

The following is an extract from my interview with Kathleen:

PR: Are you a scientist too?
KM: No I’m not. I really more know Ted on the personal level, but I saw him work and how he calculated and recalculated and how he checked. He didn’t just take what was written to be necessarily accurate and correct; he re-measured other peoples work and calculations. I saw him do that. He didn’t take anything at face value. That was one of his traits, to be a doubting Thomas, not to assume anything.

Since Ted died a couple of people have wanted to share with me what they observed about Ted, working with him as a scientist. And they all seem to say the same thing, that Ted was not stopped by what others thought, he wasn’t limited. Even with politics I could see how he could understand both sides of the issue, rather than take the side of one and not to understand or consider the side of another.

PR: What lead Ted to be interested in the laser work in the first place?
KM: Ted got the love of science from his father who was a very creative electrical engineer. In their home as a young boy he was taught science and he excelled. He and his father were always, I have letters between him and his father in college working out certain kinds of problems, and writing back and forth, of course that wasn’t the work that was about the laser, but it was part of his science background.

He had many interests, he had broad interests but there were certain areas where he could go so deep. It astounded me the perspectives, the understanding that he would go to, the level he could go. So in a way it doesn’t surprise me that he could take on a project like the laser that well known scientists had already been discouraged [from pursuing]. Because the physics changes about halfway between microwave and laser, there’s very different physics, and so for those reasons they were giving up. But Ted, there was something so tenacious about him. He didn’t like it when he heard, “Oh that’s not possible”. He’d want to know why that wasn’t possible.

PR: Can you describe some of the events leading up to Ted’s discovery of the laser?
KM: When he was at Hughes that was his first important job. He was hired as a research scientist, and he was interested in moving not just a small percent higher in the electromagnetic spectrum, but actually making that quantum leap, from microwave to laser. But Hughes didn’t believe that it could be done, because at the time the labs of the world were giving up on lasers, even Bell Labs was taking down their apparatus.

Ted had a very hard time at Hughes [pursuading them to let him do the work], but he was convinced that he had a workable way of making a laser, of making coherent light. He knew he could do it. So the Army Corps of Engineers at that time had asked for a maser to be made, the idea was to make it more practical because it was 5000 lb. The agreement made between Hughes and Ted Maiman, was he deliver the maser for the Army Corps of Engineers, and if he was successful, he would be given 9 months and $50,000 to actually make coherent light.

He went to work to make the maser more practical, took it from 5000 lb to 2.5 lb and also improved the bandwidth; it was because of that he was able to do a dedicated project on the laser [...] I think that the real story is of a man who, through his persistence, was able to defy the authoritative voices of the time who were saying the laser was not possible.

Read “And let there be light”, an account of the race to make a laser in Physics World magazine: May 2010 special issue (written by yours truly).

* Maiman’s words.

Winston Churchill famously said that history is written by the victors. In the case of the laser it might be more accurate to say that history was written by those with the best public relations team.

This weekend, on 16 May to be precise, the laser celebrates its 50th anniversary. On this day in 1960, Theodore “Ted” Maiman, a junior employee at the Hughes Aircraft Company, observed the first evidence of laser action.

His device, which was built out of a photographic flash lamp coiled around a stubby crystal of artificial ruby, emitted short pulses of red light. The optical signal was faint – it had to be captured and amplified using a photomultiplier tube – but measurements proved that it was a laser nonetheless. Improving his design over the next few weeks, Maiman created a laser bright enough to burn through a Gillette razor blade.

The significance of the laser was obvious to Maiman’s employers, who wanted to publicize the achievement right away, and so convened a press conference in New York on 7 July 1960. Far from being described as “a solution in search of a problem”, the press release issued on that day contained a laundry list of possible applications for a laser.

I think it’s safe to say that the outcome of the press conference wasn’t quite what Hughes intended: sensationalist newspaper headlines screaming “LA Man Invents Scientific Death Ray”; loss of credibility because Maiman hadn’t been given a chance to publish his results in a scientific journal; confusion in the scientific community because the press photo that was circulated showed a later design for the laser using a much larger flash lamp and crystal than the original; and having gone public in advance of filing any patents, Hughes forfeited the international patent rights to Maiman’s laser design.

Once word was out, research laboratories across the US dispatched scientists to Malibu to try to find out what Maiman had actually done, while simultaneously redoubling their own efforts to make a laser. Many researchers tried – and some soon succeeded – in reproducing his work using the brief description from the press release and the publicity photographs. One of those research teams was at Bell Telephone Labs, now known as Bell Labs, which succeeded in making a pulsed ruby laser on 1 August 1960.

The Bell Labs public relations team persuaded the researchers to haul their apparatus 25 miles away to the top of a hill and then aim it back at an old, very tall radar tower behind at the company’s headquarters in Murray Hill. This provided a powerful demonstration of the laser’s potential for optical communications, which was perfect material for a press release. Time magazine said the device could be as important as the transistor; like many reports, they didn’t seem to realise that this wasn’t the first laser.

The positive reaction may have been influenced by the fact that the Bell Labs’ name was already strongly associated with lasers. Who in the physics community hasn’t heard of the December 1958 Physical Review paper describing the properties of “Infrared and optical masers” by Bell Labs’ researchers Arthur Schawlow and Charles Townes? This paper, containing a proposal for making a laser using alkali gas, was the catalyst for the worldwide efforts to make a laser. Townes had already made a name for himself by inventing the maser, the precursor to the laser, which amplifies microwave radiation rather than visible light.

To be continued…

And so to Vienna for the 35th year of ECOC, which is billed — quite rightly in my view — as the leading optical networking event in Europe. Personal highlights from last year included the post-deadline paper from Alcatel-Lucent and Draka, which reported 40 Gbit/s transmission over transoceanic distances for the first time, JDSU’s photonic integrated amplifier, and Rod Alferness’ plenary lecture on predictions for the next 10 years of telecoms. I’m sure 2009 will provide a similar combination of interesting technical results, innovative new product announcements, and insightful debate on the future of optical networking technology. In short, pure heaven for technophiles like me.

But while rejoicing in our love of technology, can we ignore the fact that since last year the world has become gripped by a recession deeper than anything seen before in our lifetimes? In my view, we can’t afford to. As an fledgling reporter it was drummed into me that there are always three aspects to every idea: technology, money and people. In the context of ECOC, technology and the sharing of ideas (“communication” in the personal sense) are the event’s raison d’etre; that’s two out of three covered. But when it comes to money, how does it all fit together?

Here’s a cautionary tale: at the time of writing the vendor with the most optical networking market share doesn’t have an announced 100 Gbit/s field trial; while the vendor with the most 100 Gbit/s field trials, is currently reorganizing under Chapter 11 bankruptcy protection and looks unlikely to survive as a stand-alone company. Industry analyst Mark Lum, who originally pointed out this odd juxtaposition of circumstances, is quick to add that there is no cause and effect going on here. Nevertheless, this goes to show that having the best technology is no guarantee of financial reward.

Of course, cause and effect do tend to operate more reliably in the other direction — there does need to be investment in R&D to secure the future of a business. For the optical components community, which has been working hard to build viable businesses after the tech crash of the early 2000s, the credit crisis could not have come at a worse time. Many optical components and equipment vendors are still struggling with difficult balance sheets, thin product margins, and risk-averse CFOs, all of which can put the squeeze on R&D spending.

A recent report from the OECD confirmed what we perhaps feared in our hearts: that innovation is already under threat. Historically, business R&D spending and patent filings have moved in parallel with GDP, slowing markedly during the economic downturns of the early 1990s and early 2000s. Recent evidence, based on corporate reports from the first quarter of 2009, confirms this is happening again, with R&D spending declining in many cases. US venture capital investments plunged 60% in the first quarter of 2009 and the same is true in Europe and in China. Patent applications are down. What now?

But while the recession will undoubtedly change the world, it is not the end. One bright spot is the research and education community, which seems to display a certain amount of “crisis immunity” — probably due to the fact that their focus is on the potential value of technology, rather than simply revenue generation.

As a result, R&D decision makers in the education sector are more eager to try “bleeding edge” technologies in order to assess their usefulness. In fact, a project leader from the Czech research and education network CESNET told me that if they didn’t deploy the latest kit then they wouldn’t get the funding. Hence CESNET was one of the first outfits to operate a reconfigurable optical add-drop multiplexer (ROADM) enabled network in Europe. Similarly, the UK research network JANET was one of the first to test and deploy 40 Gbit/s wavelengths in collaboration with Nortel. These kind of efforts help new technologies gain acceptance.

What’s more, academic R&D funding is typically planned on longer timescales than vendor budgets. Within Europe CELTIC is a unique organization that brings together service providers, manufacturers, universities and research institutions to fulfil short to medium-term research goals relating to telecoms (the moniker stands for Cooperation for a sustained European Leadership in Telecommunications). The total budget that has been defined for CELTIC between 2004 and 2011 is €1 billion, a number that has not been impacted by recent economic uncertainty.

Collaboration between industry and academia is nothing new of course. When universities and research institution select the topics on which to perform fundamental research, industrial roadmaps frequently play a role. Conversely, optical vendors can often secure European or national grant funding by working with an academic partner. But the recession looks set to push collaboration to new levels.

While industrial R&D budgets may be feeling the pain, the worldwide recession has opened up new opportunities for external funding, particularly in light of the renewed focus on the value of a “digital economy” at both European and national levels.

In the UK for example, the Technology Strategy Board (TSB), which is a government body that promotes collaborative R&D and technology transfer, has made Digital Britain a core programme. £1m of feasibility projects on next-generation optical access are already underway, with the hope that these efforts can evolve into larger projects funded by the European Commission’s Framework 7 program or the proposed Photonics21 ERANET+ fund. In total £30m has been earmarked for Digital Britain themed research, which includes not just optical infrastructure but a range of related topics from internet security to new methods of content delivery over digital test-beds.

Clearly, co-operation between industry and academia will be an important source of technological progress in optical networking for the foreseeable future. And that’s why an event like ECOC, which brings the two worlds together with its technical conference and exhibition, should take on a new significance for all concerned.

Pauline Rigby is a science and technology writer specialising in optical communications, and former editor of FibreSystems Europe magazine.

If there’s one company I look forward to talking to, it’s the UK’s Centre for Integrated Photonics (CIP), because they always seem to have something interesting going on. At ECOC last September, CIP wowed attendees with its demonstration of a 32-channel multi-wavelength laser. The component contained two 16-channel laser arrays, with each channel being directly modulated.

This device is aimed at WDM-PON applications, where a single transmitter could replace 32 separate devices and a modulator, allowing all the optics at the PON headend to be collapsed down onto a single linecard. One of the benefits of GPON and EPON technologies is that they simplify fibre management and economize on equipment space in the central office; the multi-wavelength laser could bring both of those benefits to WDM-PON equipment.

Now CIP has taken that idea further by adding a modulator array to its multi-wavelength laser to obtain better transmission performance. The device that CIP showed at OFC today was a 10-channel laser array, containing two arrays of five lasers mounted on the same substrate. An array size of five was picked to optimize yield.

The company expects that multi-channel devices could replace CWDM transceivers in metro networks. “Most people think that for low cost you need to do CWDM,” explained chief technology officer David Smith. “The approach we’ve taken is intrinsically narrow linewidth, and locks onto the [DWDM] ITU grid. This could compete with CWDM, but you end up with potentially much higher capacity.”

Other advantages include lower cost, smaller size, and better power efficiency. In other words, this technology boosts the three key performance metrics that are important to vendors.

Reducing power consumption has also become increasingly important over the last year, according to Smith. It’s not just environmental awareness; carriers like Verizon are demanding more energy efficient equipment from their suppliers. The problem is that for every Watt of power that’s produced by the device, it needs five to six times as much power to remove it, Smith explains.

“One of the critical things we’re going to try to do with our technology is get it coolerless,” he said. The company is developing uncooled lasers as part of a project funded by the UK Technology Strategy Board, and hopes to partner with another vendor to gain access to athermal (uncooled) arrayed waveguide gratings (AWGs), which combine and separate wavelengths.

CIP is hoping that it can apply its multi-wavelength laser technology to multiple markets — metro and access — to spread the development cost, and access higher volumes to bring manufacturing costs down. The company’s challenge now is to make that transition from a small outfit with really cool R&D, to a components vendor with (hopefully) much higher revenues.

UPDATE: 30/06/09 David Smith, CIP’s chief technology officer, pointed out to me that most of his company’s revenues are generated through commercial R&D. I have corrected the final sentence.

For some years now, Intel has been looking for a way to “siliconize” photonics. The chip giant wants to build optical devices on a silicon substrate to drive the manufacturing process to higher volumes and lower cost. Now the company says it has made a breakthrough in one of the key components that would be required — a silicon-based optical detector.

Intel has added germanium to silicon to create an avalanche photodiode (APD) that is better at detecting high-speed, low-intensity signals than existing devices. The results have been published in the journal Nature Photonics.

“This is the first time that a silicon photonics device has better performance than any recorded performance from an equivalent device in III-V materials, specifically indium phosphide,” claims Mario Paniccia, Intel fellow and director of the company’s photonics technology lab.

That’s a big deal because the performance of silicon-based components is usually worse than their counterparts in other materials. Paniccia said the target performance in the company’s silicon photonics work is typically 90% of the performance for an order of magnitude in cost reduction. The APD is an exception to this rule.

Most optical components are made from III-V semiconductor materials, such as gallium arsenide and indium phosphide, because these materials are good at creating and detecting light. Silicon is not good at these functions — hence the addition of germanium to the device to allow light to be “caught” as it impinges on the device.

However, although silicon is not good at absorbing light, it has very good electronic properties and this can be exploited in an APD to create gain.

In a standard PIN detector, light arriving at the absorption region is converted into an electron-hole pair. An applied voltage separates the electron-hole pair, which move towards the electrodes to generate current. The APD operates on the same principle, but with one key difference — the multiplication region. When the electron is pulled into this region it creates additional electron-hole pairs, and these in turn create further electron-hole pairs. Now a single photon incident on the device can produce tens or even hundreds of electrons.

The performance of the APD is measured as its gain-bandwidth product. This parameter, which is the gain of the device multiplied by the speed at which it operates, is fixed for a particular type of material. For an indium phosphide APD the figure is typically around 120 GHz.

In the Nature Photonics paper, the researchers report devices operating at a wavelength of 1300 nm and a data rate of 10 Gbit/s, which had a gain of “over 30″. The best result was a “world record” gain-bandwidth product of 340 GHz.

Real-world applications
Intel’s silicon photonics work is mostly aimed at short-reach connections, between or even inside electronic chips, but the silicon APD would probably find its first applications in mainstream telecoms, where it could provide a low-cost replacement for detectors in long-haul links. Other applications such as sensing, imaging, quantum cryptography or bio-chips have also been mooted.

The enhancement in gain-bandwidth product can be useful in a couple of ways. For starters, the extra gain could be used to extend the transmission distance, because the device can detect weak signals that have lost intensity as they travel down an optical fibre. Alternatively, it could be used to provide leeway in system design, allowing cheaper, low-power lasers to be used on the transmit end of the link, which could be particularly useful for applications like fibre-to-the-home.

The third option is to use the device to provide more moderate gain at higher speeds. Intel suggests that the silicon APD could help to lower the cost of 40 Gbit/s systems, although it hasn’t yet built a device that can operate that fast.

Paniccia emphases the fact that this is a research result, although he doesn’t see any particular obstacles to commercialization. “Now it’s about optimizing those devices for performance, packaging the devices and getting the manufacturing process qualified,” he says.

“We have work to do in terms of reducing the dark current and improving the sensitivity, which is based on the dark current,” Paniccia adds. The dark current, which flows even when no light is present, is caused by the slight lattice mismatch between the silicon and germanium layers of the device.

The prototype APDs were produced on a commercial CMOS production line alongside memory chips being made by Intel spin-out Numonyx, where much of Intel’s silicon photonics work is now carried out. Experts at the University of Virginia and the University of California, Santa Barbara, provided consultation and assisted with testing, Intel said.

Advancing the cause
This result is the latest in a string of silicon photonics advances from the chip giant. Back in 2004 Intel pushed the performance of silicon modulators, which at the time only ran at 20 MHz, out to 10 Gbit/s. Two years later it boosted the data rate to 40 Gbit/s, and then last summer demonstrated an array of eight modulators on a single chip, giving an aggregate capacity of 200 Gbit/s.

Intel’s most recent achievement, in 2006, was the demonstration of a hybrid silicon laser, which integrated a tiny piece of indium phosphide onto a silicon-based structure to provide light emission.

According to Paniccia, Intel now believes that the hybrid silicon laser is the way forward for silicon photonics, rather than the all-silicon laser based on the Raman effect that Intel demonstrated back in 2005.

Of course, Intel is not alone in its pursuit of silicon photonics. Other companies in the field include CyOptics, Lightwire and Luxtera. In fact, Luxtera reported a monolithic detector based on silicon in March 2007. Like Intel, Luxtera used germanium on silicon, but the device had a lower sensitivity than Intel’s silicon APD.

Reproduced with permission. © Institute of Physics and IOP Publishing Ltd.

This summer the Optical Internetworking Forum (OIF) decided to forge ahead with the standardization of optical modules for 100 Gbit/s networking based on a modulation format called dual-polarization quadrature phase-shift keying (DP-QPSK).

DP-QPSK cuts the symbol rate on the fibre by a factor of four, by transmitting two bits of data per clock cycle on each of two polarizations. Thus a 100 Gbit/s data rate is reduced to a symbol rate of just 25 Gbaud on the fibre, which mitigates the impact of chromatic dispersion — an effect that causes high-speed pulses to spread out as they travel down the fibre.

“We have selected an implementation approach supported by a critical mass of photonic component vendors and users,” claimed David Stauffer of IBM, and the OIF’s PLL Working Group chair.

However, it seems that no sooner did the OIF give its blessing to DP-QPSK, than the dissenters started to crawl out of the woodwork.

One of the first vendors to actively back other options was ADVA Optical Networking, which is taking part in a research initiative called 100 Gigabit Ethernet Transport (100GET). The partners of the 100GET-METRO sub-project, which include optical software company VPIsystems, will use numerical simulations to investigate the robustness of different modulation formats to fibre impairments including higher-order effects like second-order polarization mode dispersion (PMD) and chromatic dispersion slope. Their aim is to find the best approach for 100 Gbit/s transport in metro networks, which may well have different requirements from long-haul networks.

Similarly, DWDM vendor Ericsson, which recently started talking about its 100 Gbit/s ambitions, is taking part in a different sub-project under the 100GET umbrella. The project participants are keeping a totally open mind on which modulation format would be best for DWDM transport: orthogonal frequency division multiplexing (OFDM), sub-carrier modulation (SCM), and multi-level phase shift keying (M-PSK) are all under investigation.

“100G transmission presents considerable challenges,” said Rodolfo Di Muro, optical product marketing manager for Ericsson’s business unit networks. “It is anticipated that 100 Gbit/s will coexist with 40 Gbit/s or 10 Gbit/s in the same network with minimum architectural change. Technology is not mature yet, and we can anticipate that one size does not fit all networks.”

Sensitive issues
Even Len Bosack, founder and owner of optical firm XKL, is ready to stick his oar in. He believes that DP-QPSK isn’t going to be up to the job at 100 Gbit/s because the optical signal will be too sensitive to chromatic dispersion and PMD.

“I keep suggesting to people that what you really want is polarisation-diverse 8-QAM [quadrature amplitude modulation, a name for advanced phase-shift keying formats], and that will get you a dandy 100G transmission system, which will go quite considerable distances,” he told fibresystems.org. 8-QAM is a coding scheme that transmits three bits of data per symbol. Combined with polarization multiplexing, it has a symbol rate on the fibre of 15 Gbaud.

“Yes you’ve given up some receiver sensitivity to get there, and the transport is a bit more complicated, but in my view it’s a tradeoff that is both feasible and desirable,” he adds.

Ironically Nortel, which is an ardent supporter of DP-QPSK at 40 Gbit/s, says it won’t be using the same modulation format at 100 Gbit/s. Nortel wants its 100 Gbit/s solution to be an overlay solution for 10 Gbit/s-engineered networks, which means slashing the symbol rate on the fibre from 100 Gbit/s right down to 10 Gbit/s.

In fact, so far fibresystems.org has only found one vendor that’s prepared to back DP-QPSK at 100 Gbit/s. Alcatel-Lucent doesn’t favour DP-QPSK for 40 Gbit/s transmission, due to the possibility of optical crosstalk with 10 Gbit/s signals running on the same fibre, which would have the same baud rate as a 40 Gbit/s DP-QPSK signal. However, the giant vendor thinks that DP-QPSK has potential at 100 Gbit/s.

“At 100 Gbit/s DP-QPSK suffers much less impact from nonlinear effects, so a modulation format that experiences issues at 40G becomes recommended at 100G,” said Paolo Ottolenghi, senior product manager at Alcatel-Lucent.

Reproduced with permission. © Institute of Physics and IOP Publishing Ltd.

The standardization of higher data rates is vital if Ethernet is to continue as a ubiquitous end-to-end protocol. Pauline Rigby finds out how standards are progressing.

Ethernet has traditionally evolved in multiples of 10, from the first successful commercial version of Ethernet at 10 Mbit/s through Fast Ethernet (100 Mbit/s), Gigabit Ethernet (1 Gbit/s) to 10 Gigabit Ethernet (10 Gbit/s) – the highest speed available today. But the next multiple — 100 Gigabit Ethernet (100 GbE) — hit a speed bump when disagreement arose between different interests within the Institute of Electrical and Electronic Engineers (IEEE) Higher Speed Study Group (HSSG).

Eventually vendors reached an agreement, deciding to add 40 Gbit/s data rate to the proposed standard for 100 GbE, which allowed things to move on to the next stage. On 5 December 2007, the HSSG officially became the IEEE P802.3ba Task Force, doing the technical work to define the standard. Here Pauline Rigby talks to John D’Ambrosia, chair of the Task Force, whose full-time job is to guide the process and ensure the work finishes on time.

PR: What’s driving the demand for higher data rates?
JD’A: Quite simply traffic was growing at such a rate in network aggregation applications, especially with the proliferation of 10 GbE, that it needed a higher-speed solution. People were talking about aggregating 16 or 32 links of 10 GbE to handle all of the servers they were using. Link aggregation was becoming very unwieldy; we needed something faster.

Naturally the Ethernet switch vendors were the most vocal about the need for 100 GbE, but it wasn’t just them. One of the big differences between this project and its predecessors was the involvement with the end-user community. We had people from Google, EDS, Microsoft Networks and AMS-IX [Amsterdam Internet Exchange]. We had content providers coming in because video drives the bandwidth through the roof.

One of the examples that really blew people away was from Yahoo! Asia Pacific. They were offering a service for streaming classic major-league baseball games, and filled a 40 Gbit/s pipe. They don’t know what the capacity request actually was because they had to cap the service at 40 Gbit/s.

I was talking with an individual who was at Google at the time of the HSSG, and he said that: “People point at Google as the anomaly and say well that’s just Google. I can tell you that where Google was a year and a half ago is where all of the other data centres are coming to now, they’re all starting to hit these problems.”

If the demand for bandwidth is growing so rapidly, how did the need for an intermediate rate at 40 Gbit/s arise?
There’s one chart that really helps to explain this question (figure 1). This is a case where a picture is worth a thousand words. This chart was based on input from a server vendor. He was talking about Moore’s Law and how the server I/O [input/output] doubles every 24 months. When the growth curve for network aggregation was added, what we see is two different growth rates for two different applications. Core networking doubles every 18 months, server I/O doubles every 24 months. When you plot this out over time you can see this leads to a huge difference.

We mapped different Ethernet standards across the lines, and looked at how that relates to the needs. You can see why Gigabit Ethernet was pretty much an instant success when it was released in 1998. It instantly hit the requirements for network aggregation, and a couple of years later it hit the servers.

Computing and networking bandwidth needs for 40 and 100?GbE.

In general people say Gigabit Ethernet was a huge success, but when you ask about 10 GbE, it’s a different story. If you were talking to someone who was judging 10 GbE by the value it brought to networking, those individuals say it was a raging success, and you can see that from the chart. However, if you were talking to people who were judging by ports shipped, i.e. server vendors, they say they don’t really need it and won’t need it until around 2010.

To move forward within the IEEE you need to have 75% consensus. Considering the two different backgrounds of the people present in the room, you can imagine the difficulty the group at that time had in coming to grips with this problem so that it could move ahead.

You must have been very happy that a consensus was reached. Is it unusual to have two different rates being developed for a single standard?
Yes it is unusual; it’s the first time history that we’d had two new rates at one time, so saying it’s a big deal is an understatement. But while it was unusual to do the two rates, when you look at the chart it becomes apparent that two rates was the right decision.
Some people are saying that 100 GbE is going to be late. Is that the case?

There are many people that will tell you that 100 GbE is going to be late, and if you look at the chart you can understand why they think that. When you look at 100 GbE in relation to the networking applications, it just barely meets the curve. We’re going to finish the specification at around the time the need is there, so there’s going to be a lot of pressure on the group to produce this thing as fast as we can. On the other hand, if you look at the relation to the computing applications, 100 GbE will be seven to eight years early.

Is there a risk that by trying to please everyone you end up pleasing no-one?
Let me answer that from the standards approach. There was essentially a one meeting delay before we were able to become a task force, which impacted on the schedule. However, as the chair my primary job is to make sure that a standard gets done in the proper time frame, and as a result we’ve scheduled accordingly. It really hasn’t been an issue since we made a decision to do it this way.

In fact, we’ve come up with a unique architecture that will satisfy both rates at the same time, and we’ve adopted most of the baseline proposals around that architecture. There are a few more pieces that we need to fill in, but we are making incredible progress.

That’s interesting. Tell us how the architecture can support both rates.
Several of our key objectives concern the physical layer specifications, or PMD [physical medium dependent]. Right now the solutions that are being considered are based on ganging four lanes or 10 lanes: four lanes of 10 Gbit/s, 10 lanes of 10 Gbit/s or four lanes of 25 Gbit/s. Those lanes could be physical lanes or they could be wavelengths. There are also some individuals from a transport background who proposed 40 Gbit/s serial, and others who suggested two lanes.

The next layer in the structure is the PMA, which means physical medium attachment, and that’s where the multiplexing happens. We came up with a unique approach that allows us to adjust the multiplexing to support 10 lanes, four lanes, two lanes — whatever we need.

One of the main areas we had a lot of debate on was in the physical coding. We have just adopted an approach called MLD — multilane distribution — which is based on the standard 64B/66B encoding that we do now in Ethernet. Simply put, it’s a mechanism to distribute the data across those different lanes. So now we have a lane-distribution mechanism that carves up the signal into the right number of lanes for whatever PMD we’re going to target.

The task force met in May. How did that move things along?
The May meeting was phenomenal, and I attribute that to the hard work that the task force is doing in terms of the consensus building and due diligence with the technical issues. We actually wound up making quite a few baseline proposal decisions at that point, which is ahead of schedule.

On the multimode fibre objective for 40 GbE and 100 GbE we’ve adopted a proposal that’s n × 10.3125 Gbit/s across parallel fibres. For the 40 km 100 GbE and the 10 km by 100 GbE we’ve adopted a 4 × 25 Gbit/s architecture. The decision on 40 GbE over 10 km hasn’t been made yet.

There are two proposals out at this point — one is for 40 Gbit/s serial and the second is for a 4 × 10 Gbit/s CWDM [coarse wavelength division multiplexed] approach. That will be one of the key decisions we’ll need to make at the July meeting so that we can move ahead.

What’s the timetable for completing the 40/100 GbE standard?
To keep to our schedule we aim to generate draft 1.0 after the September 2008 meeting. The next phase will be a Working Group Ballot — that’s where you open it up to the 802.3 Ethernet Working Group for comment, which is scheduled for the March meeting in 2009. After that comes the Sponsor Ballot in November 2009, where you open it up to the members of the IEEE Standards Association. The standards release is scheduled for June 2010.

For more information visit www.ieee802.org/3/ba/.

Reproduced with permission. © Institute of Physics and IOP Publishing Ltd.

We’ve all been taught that electronics and water don’t mix, but researchers at IBM’s Zurich Research Laboratory say that doesn’t have to be the case. In fact, they are proposing that tiny rivers of water be used to cool three-dimensional stacks of silicon chips in future generations of server processors and communications ICs.

Not only would this enhance the performance of the silicon, the heat harvested from the chips could be piped into heating and hot water systems to create the ultimate “green” data centre.

To this end, IBM researchers in collaboration with the Fraunhofer Institute in Berlin have demonstrated a prototype that integrates a cooling system into a 3D chip by piping pressurized water between each layer in the stack.

So-called 3D chip stacks — which take chips and memory devices that traditionally sit side-by-side on a silicon wafer and stack them together on top of one another — represent one of the most promising approaches to enhancing chip performance beyond its predicted limits.

These limits are expected to arise as chip makers pack more processing power onto a chip by scaling down the transistors — following the well-known Moore’s Law.

First, there’s the fact that the wiring between chips does not scale with the transistors on them, because the width of wires is shrinking but their length is not. 3D chip stacks exploit the third dimension to create much shorter vertical connections between the chips in each layer — up to 1000 times shorter according to IBM.

Using 3D chip stacks alleviates the wiring crisis but introduces another problem — cooling. IBM estimates that a 3D chip stack with an area of 4 cm and a thickness of about 1 mm would generate close to 1 kW in heat — more than enough to melt the components.

Conventional coolers attached to the back of a chip simply don’t scale to meet this requirement. What’s more, the 3D stacking process exacerbates the problem because each layer poses an additional barrier to heat removal.

“In order to exploit the potential of high-performance, 3D chip stacking, we need interlayer cooling,” explains Thomas Brunschwiler, project leader at IBM’s Zurich Research Laboratory. “Until now, nobody has demonstrated viable solutions to this problem.”

The key technical challenge was to design a system that maximizes the water flow through the layers, yet hermetically seals the interconnects to prevent water from causing electrical shorts. IBM researchers packaged the chip stacks in a sealed pressurized silicon housing with an inlet reservoir on one side and an outlet reservoir on the other. The cooling layer in the prototype was just 100 µm thick and contained 10,000 vertical interconnects per cm2.

IBM says the complexity of such a system resembles that of a human brain, wherein millions of nerves and neurons for signal transmissions are intermixed but do not interfere with tens of thousands of blood vessels for cooling and energy supply. IBM hopes to commercialize the technology in the next five years.

Reproduced with permission. © Institute of Physics and IOP Publishing Ltd.