ECOC 2024 industry reflections
Gazettabyte is asking industry figures for their thoughts after attending the recent 50th-anniversary ECOC show in Frankfurt. Here are the first contributions from Huawei's Maxim Kuschnerov, Coherent's Vipul Bhatt, and Broadcom's Rajiv Pancholy.
Maxim Kuschnerov, Director R&D, Optical & Quantum Communication Laboratory at Huawei.
At ECOC, my main interest concerned the evolution of data centre networking to 400 gigabits per lane for optics and electronics. Historically, the adoption of new optical line rates always preceded the serdes electrical interconnects but now copper cables are likely to drive much of the leading development work at 400 gigabit per lane.
Arista Networks argued that 448G-PAM6 works better for copper, while 448G-PAM4 is the better choice for optics – a recurring argument. While PAM6 signalling is certainly more suitable for longer copper cables, it will face even tougher challenges on the optical side with increasing reflection requirements in newly built, dusty data centres. Also, a linear drive option for future Ethernet will be imperative, given the DSP’s increasing share of the the consumption in pluggable modules. Here, a native 448G-PAM4 format for the serdes (the attachment unit interface or AUI) and optics looks more practical.
My most important takeaway regarding components was the initial feasibility of electro-absorption modulated lasers (EMLs) with a greater than 100GHz analogue bandwidth, presented by Lumentum and Mitsubishi publicly and other companies privately. Along with thin-film lithium niobate (TFLN) Mach–Zehnder modulators suited for Direct Reach (DR) applications with shared lasers, EMLs have historically offered low cost, small size and native laser integration.
For 1.6-terabit modules, everyone is waiting on the system availability of 224-gigabit serdes at a switch and network interface card (NIC) level. The power consumption of 1.6-terabit optical modules will improve with 3nm CMOS DSPs and native 200 gigabit per lane. Still, it gets into an unhealthy region where the network cable power consumption is in the same ballpark as the system function of switching. Here, the bet on LPO certainly didn’t pay off at 100 gigabits per lane and will not pay off at 200 gigabits per lane at scale. The question is whether linear receive optics (LRO)/ half-retimed approaches will enter the market. Technically, it’s feasible. So, it might take one big market player with enough vertical integration capability and a need to reduce power consumption to move the needle into this more proprietary, closed-system direction. Nvidia showcased their PAM4 DSP at the show. Just saying…
212G VCSELs are still uncertain. There is a tight initial deployment window to be hit if these high-speed VCSELS are to displace single-mode fibre-based optics at the major operators. Coherent’s results of 34GHz bandwidth are not sufficient and don’t look like something that could yet be produced at scale. Claims by some companies that a 400 gigabit per lane VCSEL is feasible sound hollow for now, with the industry crawling around the 30GHz bandwidth window.
Last but not least, co-packaged optics. For years, this technology couldn’t escape gimmick status. Certainly, reliability, serviceability, and testability of co-packaged optics using today’s methodology would make a deployment impractical. However, the big prize at 400 gigabit per lane is saving power – a significant operational expense for operators – something that is too attractive to ignore.
The targets of improving optics diagnostics, developing higher-performance dust-reflection DSP algorithms to deal with multi-path interference, adopting more resiliency to failure in the network, and introducing a higher degree of laser sparing are not insurmountable tasks if the industry sets its mind to them. Given the ludicrous goals of the AI industry, which is reactivating and rebranding nuclear power plants, a significant reduction in network power might finally serve a higher purpose than just building a plumber’s pipe.
Vipul Bhatt, Vice President of Marketing, Datacom Vertical, Coherent
ECOC 2024 was the most convincing testimony that the optical transceiver industry has risen to the challenge of AI’s explosive growth. There was hype, but I saw more solid work than hype. I saw demonstrations and presentations affirming that the 800-gigabit generation was maturing quickly, while preparations are underway for the next leap to 1.6 terabit and then 3.2 terabit.
This is no small feat, because the optics for AI is more demanding in three ways. I call them the three P’s of AI optics: performance, proliferation, and pace.
Performance because 200 gigabit PAM4 optical lanes must work with a low error rate at higher bandwidth. Proliferation because the drive to reduce power consumption has added new transceiver variants like linear packaged optics (LPO) and linear receive optics (LRO). And pace because the specifications of AI optics are evolving at a faster pace than traditional IEEE standards.
Rajiv Pancholy, Director of Hyperscale Strategy and Products, Optical Systems Division, Broadcom
As generative AI systems move to unsupervised, transformer-based parallel architectures, there is less time for resending packets due to data transmission errors. Improved bit error rates are thus required to reduce training times while higher interconnect bandwidth and data rates are needed to support larger GPU clusters. These compute networks are already moving to 224 gigabit PAM4 well before the previous generation at 112 gigabit PAM4 was allowed to reach hyperscale deployment volumes.
The problem is scalability with a high-radix supporting all-to-all connectivity. The power for a single rack of 72 GPUs is 120kW, and even with liquid cooling, this becomes challenging. Interconnecting larger scale-up and scale-out AI computing clusters requires more switching layers which increases latency.
Furthermore, after 224 gigabit PAM4, the losses through copper at 448 gigabit PAM4 make link distances from the ASIC too short. Moving to modulation schemes like PAM-6 or PAM-8 presents a problem for the optics, which would need to stay at 448 gigabit PAM4 to minimize crosstalk and insertion losses.
Supporting 448 gigabit PAM4 with optics then potentially requires new materials to be integrated into silicon, like thin-film lithium niobate (TFLN) and Barium Titanate (BaTiO3), electro-optic (EO) polymers, and III-V materials like Indium Phosphate (InP) and Gallium Arsenide (GaAs). So now we have a gearbox and, potentially, a higher forward error correction (FEC) coding gain is required, adding more power and latency before the signal even gets to the transmit-side optics.
There were 1.6-terabit OSFP transceivers operating with eight lanes of 212.5 gigabit PAM4 while vendors continue to work towards a 3.2-terabit OSFP-XD. With 32 x 3.2Tbps pluggables operating at 40W each, the optical interconnect power would be 1.3kW for a 102.4Tbps switch. And if you use 64 x 1.6Tbps OSFP at 25W each, the optical interconnect power will be eben higher, at 1.6kW. I wonder how linear pluggable optics can compensate for all the path impairments and reflections at high data rates from pluggable solutions. Perhaps you can relax link budgets, temperature requirements, and interoperability compliance.
The best session this year was the last ECOC Market Focus panel on the Tuesday, which kept everyone a bit longer before they could figure out where in Frankfurt Oktoberfest beer was on tap. The panel addressed “Next-Gen Networking Optics like 1.6T or 3.2T”. All but one of the participants discussed the need and a migration to co-packaged optics, which we at Broadcom first demonstrated in March 2022.
It was great to also present at the ECOC Market Focus forum. My presentation was titled “Will you need CPO in 3 years?” Last year in Glasgow, I gave a similar presentation: “Will you need CPO in 5 years?”
ECOC celebrates its 50th anniversary
- The European Conference on Optical Communications (ECOC) is celebrating its 50th anniversary.
- The conference and exhibition will take place in Frankfurt, Germany, from September 22-26.
- Key themes at the show include satellite optical communication, artificial intelligence, networking for AI within the data centre, photonic integration, quantum and Green ICT.
ECOC will celebrate its 50th anniversary this month.
The event will include a special session highlighting the progress made in photonics over the last half century and will feature luminary speakers.
There will also be a celebratory event with food stalls from different countries.
“There will be an opportunity to mingle, and we also plan an exhibition to look back at what has been achieved combined with what directions we need to take for the future,” says Jörg-Peter Elbers, an ECOC General Chair this year.
“Photonics is now very important for digitalising different protocols,” adds Ronald Freund, another of the General Chairs. “It is not just telecommunications but also photonic networks for the industrial production of almost anything you can imagine.”
Conference themes
The use of photonics for communications has become so diverse that no one theme captures the show. Rather, the General Chairs highlight several themes at this year’s event.
“There are hot topics that we have tried to cover, while also recognising that we have a broad audience,” says Elbers. The audience ranges from communications service providers to component makers and chip foundry operators.
“One topic is satellite network integration in the optical domain,” says Professor Carmen Mas Machuca, the third ECOC General Chair. “Satellite networks using all this optical technology is one of the new topics.”
There is ongoing research work to combine terrestrial and space networks to address capacity gaps and bolster networking resiliency. Satellite can take over for any failures in the terrestrial core network. Satellite networks also offer latency benefits which can benefit niche applications like high-frequency trading.
Combining satellite and terrestrial networks represents a tremendous opportunity but has it challenges, says Elbers. An optical satellite symposium will be hosted at ECOC.
Another topic is sensing that will be addressed in multiple sessions. Examples include using optical fibre network for sensing, to wearables such as brain-computer interfaces (see Conference keynotes, next section).
Freund highlights other networking symposia addressing the data centre, and Green ICT. “How photonics can contribute to reach net zero targets,” says Freund.
“We wanted to highlight the breadth of the topic, how many different problems we can solve with optical technologies,” says Elbers.”
Conference keynotes
ECOC will feature four keynote speakers.
This year, the topics are AI interconnect requirements in the data centre, optical networking, past and future, for telecom, silicon photonics, and using photonics for computation.
- Andy Bechtolsheim, co-founder and Chief Architect at Arista Networks, will address whether interconnects can keep up with the demands of AI. Bechtolsheim will outline ways to improve the power, cost, and reliability of optics for interconnects needed for AI, the boundaries between copper and optics, and packaging innovations.
- Thomas van Briel is the Senior Vice President of Architecture and Strategy at the communications service provider Deutsche Telekom. His will address the evolution of IP and optical networks and the technologies that have driven and will drive optical networking.
- Roel Baets is an Emeritus Professor at Ghent University and former Group Leader at imec. His talk is on ‘Silicon Photonics 4.0’. If the title puzzles you, then it is well chosen, quips Professor Baets. “What I mean is that it will be important for silicon photonics to make use of smart and agile manufacturing, a notion associated with Industry 4.0,” says Baets. “Silicon photonics can be expected to be an enabling technology for Industry 4.0 since it can enable more performant AI, sensors for IoT, and more.”
- The final speaker is Joyce Poon, Head of Photonics Architecture at Lightmatter. She will address future computing examples involving integrated photonics. Lightmatter believes the most significant opportunities and challenges in computing lie at the extremes in size and scale: large data centres and personal devices. Poon will detail how advanced integrated photonics can transform computing, from making large-scale AI systems more efficient to enabling wearables and human brain-computer interfaces.
Rump Session
This year’s Rump Session is titled: Quantum Technologies: Research hype or on track to commercial success?
The Rump Session will look at the funding Quantum players are attracting and address what financial return, if any, has been achieved. The topics to be tackled include quantum secure communications and quantum computing.
“Quantum companies are mushrooming everywhere, with Europe maybe an extreme case,” says Elbers. “It is interesting to explore what is behind this and what happens if the funding dries up?”
The session will address what is real and ready for near-term commercial deployment and what remains longer term.
Individual interests
Each Chair has their own ECOC interests regarding the latest research.
“My interest is to see how the energy bottleneck in the data centre will be solved concerning the demands of the AI processors,” says Freund. “One interesting approach is to move forward with integration, for instance, co-packaged optics and whether there are other possibilities.”
Professor Mas Machuca is interested in network security. “How can we use all these entanglement properties [to detect eavesdropping] so they can be applied in real networks?” she says. “For now, they are far from deployment, and it will be interesting to see how far it can go.”
The challenges with entanglement are the limited distances and durations achieved, measured in milliseconds, says Professor Mas Machuca.
Elbers’s hot topic is photonic integration: “The progress here has been tremendous.” What interests him is how to bring the functional blocks closer and building new functions using chiplets, for example.
Another of his interests at the show is how generative AI can be ued to enable customers to operate their networks more efficiently.
Industry issues
What are the Chairs’ concerns and observations regarding the photonics industry?
For Europe, there is a gap between research and forming of companies, says Freund. There is progress but more momentum is needed for start-ups to transfer innovation from research into applications so that smaller companies can provide solutions for larger ones.
“There’s a lot of money around, but there also has to be the people – entrepreneurs – to do that,” says Freund.
Professor Mas Machuca highlights sovereignty issues such that telecom operators in Europe have the suppliers to guarantee working networks. “How to design your network so that if one manufacturer is banned from your country, you still have an operating network,” says Professor Mas Machuca.
It requires standardisation efforts that the equipment makers must follow to ensure interoperability. “This is always a challenge,” she says.
Elbers overall is positive but does highlight challenges. The fundamental drivers remain: traffic growth is continuing while there are government investments to promote broadband in rural areas.
But there are also geopolitical issues that are forcing the telecom operators to chose carefully their suppliers. These geopolitical issues are also forcing equipment vendors to focus on particular markets.
There is also equipment-maker consolidation taking place, the recent Nokia-Infinera announcement being an example.
“That certainly brings challenges for some people, but I think it also brings opportunities for others,” says Elbers.
Industry issues
Carmen Mas Machuca is the Professor of Communication Networks at the Department of Electrical Engineering and Information Technology at the University of the Bundeswehr, Munich. Her interests include network planning, resource allocation, and resiliency of telecom networks, not just core but also access networks.
Jörg-Peter Elbers is Senior VP, Advanced Technology, Standards and IPR, at Adtran Networks. His responsibilities cover optical packet networks, optical access, and residential gateways.
As passive optical network speeds rise to 50 gigabit-per-second and beyond, there is a blurring between access, aggregation, and metro regional networks, says Elbers. There is also the growing importance of photonic integration as part of system design and this is also impacting his work interests.
Ronald Freund heads the telecommunications systems group at the Fraunhofer Heinrich Hertz Institute (HHI), Germany. The HHI is the largest applied research organization in Europe.
Freund’s group research interests span the full spectrum of networks – from in-house, access, and metro to wide area networks. A key focus of research is fibre-based communications to advance reach, capacity, and data rates, as well as network security and power consumption issues. The group has also started to address optical satellite and optical free-space communications.