This interview is an excerpt from our report titled “Optical transceivers: The new gold standard in data communications by 2030.”
With an increasing demand for optical transceivers, the landscape of data transmission is undergoing significant transformation. The next three to five years hold promise for substantial growth in transceiver utilization, driven by advancements in speed, efficiency, and technology.
Prescouter talked with Misuk Saha, a Photonics Engineer III at Benchmark Electronics. Misuk is leading the way in delivering optical packaging for transceivers and various photonic-based applications.
We interviewed Misuk about the opportunities and challenges posed by increased data transmission demand, focusing on the role of optical transceivers as a key technology.
This interview covers the following topics:
- Role and demand of optical transceivers
- Key drivers for transceiver development
- Design challenges and innovative solutions
- Potential trade-offs design
- Technological advancements and promising technologies
- Innovation and competition in the industry
- Technology roadblocks and challenges
- Return on investment and market outlook
- Differentiation and competitive strategies
Role and demand of optical transceivers:
Q: How do you see the role of optical transceivers in data transmission in the next three to five years?
A: Actually, there is a major growth in the demand for optical transceivers. If you compare 2020 or maybe three years later or at the end of 2030, the demand will be increased by at least three to four times.
However, it will depend on the application you are using. Basic transceivers such as CFP, CFF, or SFF were very low-speed transceivers between 10 GBPS and 40 GBPS. Then, newer developments facilitate transceiver manufacturing with higher speeds reaching 100 GBPS.
Based on the current market, the major transceiver manufacturing companies are trying to make 400 GBPS and 800 GBPS. So, the demand is growing. Moreover, due to recent advances, the 100 GBPS optical transceiver market is over. So, I think 800 GBPS or more than 1 terabyte is very much in demand in the next few years.
Finally, other criteria include the operating wavelength and the transmission length that you want the data to be transferred, over 1 km, 10 km, or 100 km.
Key drivers for transceiver development:
Q: What are the key drivers for the role of optical transceivers in data transmission?
A: The key drivers are the speed of the transceiver and power consumption. So, one transceiver can transmit 100 GBPS, but the power consumption is probably 3.5 watts. But now, newer developments are addressing the reduction of the amount of energy required to power them from 3.5 watts to 2 watts or 2.5 watts. The modulation techniques in the transceiver used different types for NDR, PAM4, or QAM16.
So, what is the benefit of an optical transceiver? It is that you can transmit 16 different signals at the same time, but those are very complex in nature. From my perspective, the main goal is to transfer more data with less power consumption.
Q: How important are form factors or other protocols in the context of data transfers, speed, and distance?
A: The form factor is very important because, in the transceiver market, there is a multi-source agreement (MSA). So, those compliances have some rules, like if you want to make a 100 GBPS transceiver, you can use four channels. Each channel has to carry 25 GBPS.
Design challenges and innovative solutions:
Q: What must be done to design transceivers to achieve speeds beyond 100 GBPS?
A: Because the 100 GBPS market is over, if you want to make a transceiver with 100 GBPS, I don’t think you can make money. But if you plan to make anything beyond 400 GBPS or 800 GBPS, maybe in the next few years, the demand will be much higher for those products.
Q: What needs to be done to achieve these high speeds?
A: In the optical transceiver, there are two major components: the laser source and the modulator. If you want to increase your power and speed, you need high speed & high modulation techniques.
Another key factor is polarization. So, some companies are currently trying to use polarization for data transfer.
It depends on what kind of modulation technique you want to use and what design you want. For this, you need to determine what kind of expertise a design engineer has.
In the current market, I would say there are 5-10 companies trying to make high-speed transceivers, but all of them have different platforms.
Potential trade-offs in design:
Q: How important is the trade-off between speed and distance when transmitting data?
A: The speed and distance are correlated. If you have to transmit one single data in 10 meters is not the same as wanting to transmit it 100 kilometers. For this, you need a powerful transceiver that can allow you to transfer your data at higher distances.
This is also influenced by your connector type and the technique you are using. Like for QSFP-DD, which is for high-distance transmission of more than 10 kilometers because when you increase the distance, there is a term called intersymbol interference that will cause your signal deterioration.
So, that’s why you need a powerful data transmitter when transmitting long-distance data.
Right now, based on the demand for an increase in speed, there is another form factor called double density. The space which was used for 100 GBPS, now you need to make 400GBPS with the same space. So, that form factor will define your criteria for making such a product.
Promising technologies:
Q: What technologies do you think are promising? Some are coherent transmission, indium phosphide-based heterojunction bipolar transistors, advanced digital signal processing, optical frequency combs, and external cavity lasers.
A: So, right now, I would say silicon photonics is difficult because, in my last job in New Mexico, the company tried using a hybrid approach between Silicon and Indium Phosphide. Silicon is a group 4 material, and Indium Phosphide, if you consider the periodic scale, it’s a III-V material. So, when you integrate silicon with a III-V material, efficiency is lost.
So, I would say if you use one platform, either silicon or indium phosphide, and don’t mess up with those things on one platform, that will create more difficulties. So, I think Indium Phosphide is a very good platform for high-speed transmission, and it will depend on the type of modulation technique you implement. You need your DSP, which means digital signal processing to match your goal.
Innovation and competition in the industry:
Q: How will companies innovate to keep up with the demand for increasing data communications in the future?
A: You need to invest in research and development in your photonic chip. So, as I said already, different companies use different types of platforms. Some companies use Silicon Photonics and Indium Phosphide platforms.
Based on the platform and the modulation technique you use, the modulation technique means you can use NRZ. You can use PAM4, QAM16, or QAM64. So, based on those combinations of the platform you use to make your silicon photonic chip and the modulation technique that will relate to the speed of your transceiver.
Q: Which companies, in your opinion, are innovating in this space, and how are they innovating?
A: So, I would say a few big companies are trying to make some very, very good cool products. So, I would say Arista Networks and Broadcom are investing huge.
Recently, Cisco bought Acacia Communications, a New Jersey-based transceiver company, for $2 billion. Acacia had a product in the market.
Finisar, Lumentum, and Intel are trying to make something different. So, these companies are now major leaders for transceivers, at least in the United States.
Q: What makes them leaders? How are they innovating so well?
A: Based on my understanding, how cheap can you make your product, and how efficient and reliable? So, those are the key points. Like how much investment do you need to make one transceiver, or how efficient or how reliable is your transceiver? These are very important factors to take into account.
So, I would say Cisco has a good product, but they don’t have their own product. The product comes from its recent acquisition.
Arista has some good products. But it all comes down to investment, how cheap you can make your product, that’s very important, and how efficient.
Q: Do you think that they’re all innovating in the same way, or they’re doing things drastically different from each other?
A: I would say they are different from each other, totally different. But the basics are the same. But if you consider the platform, the DSP, and the electronics, they are different, for sure.
Q: Can you elaborate on that a little more, maybe some of the differences in some of them? I know that you don’t work for all of them or anything like that. We’re just really interested in this aspect.
A: For each company, they have their own platform, like how they will make their photonics cheap. Some companies use silicon photonics, and some companies just use III-V material. So, based on that platform, it will define which platform each company is using. That’s one type.
Another type is the modulation technique. The modulation technique means your light needs to be modulated with your signal. Some companies are using PAM4. That means they are sending two bits in one signal.
Some companies are using QAM16. They are sending four data in one signal. But the more complex you make your modulation technique, the more difficult it is to make that product. But the good thing is there will be more speed. So, you have to trade off based on your design and technique.
Technology roadblocks and challenges:
Q: What are the potential technology roadblocks for developing novel transceivers to achieve those ultra-fast speeds and high distances?
A: So, one of the major roadblocks right now is optical packaging, I would say. Optical packaging means you have the photonic chip and the fiber in your transceiver, but your fiber needs to connect with your chip. That means you need an optical alignment, and these are the basics of optical packaging. So, I would say optical packaging still needs to be much more developed because, with the current technology we have, it’s very expensive.
Photonic chip fabrication is not that expensive because we use normal microelectronics equipment, so we don’t need to buy big equipment or expensive equipment to make this photonic chip.
But the challenge is that for this optical packaging, you need more manpower, experienced technicians, and engineers, which will make your transceiver cost higher than expected. So, I would say one of the roadblocks is the optical packaging. You need too much money for optical packaging.
Then there are some other roadblocks, like in your chip application technique, where there is some problem in the wet etching. So, that also needs more research and development work in the fabrication. Hence, with this kind of thing right now, the current roadblock is making more high-speed fiber optic transceivers with minimum cost.
Return on investment and market outlook:
Q: In your view, if a player is investing in research and development of ultra-fast optical transceivers, could they see a potential return on investment in the near future, say five to seven years?
A: I would definitely say yes. Yet you need to think about where you are investing your money because right now in the market, there are many startup companies. But the problem is most of them don’t have good expert engineers, so they’re working based on trial and error. So, that’s why you need to think before you invest in the company and how they are trying to make it. You need to understand.
Differentiation and competitive strategies:
Q: Can companies differentiate themselves to stay ahead of their competitors in the silicon photonics market?
A: So, actually, this is a very tough question because every company is trying to develop their own idea on their own platform, but you don’t know what the other companies are doing. So, maybe they are doing something very cool rather than you. So, it’s very hard to tell what the other companies are doing.
But lots of research work right now is going on. R&D work is going on. That’s for sure. But it is hard to understand the other companies because most of the companies are very IP protected. So, you don’t know what they are actually doing.
But if you follow the company, you can understand what they are trying to make and what they are investing in. Some companies are investing in 400 GBPS. Some companies are investing in 800 GBPS. So, based on their platform and technology, it depends. It’s hard to tell which company is better and which company is not.
Conclusion:
The interview with Misuk Saha highlights the ongoing evolution of optical transceiver technologies in response to escalating data demands. As the demand for higher speeds and longer distances intensifies, the race to innovate in optical transceiver technology becomes more critical.
Companies investing in research and development are set to utilize opportunities in the dynamic data transmission market. By navigating through challenges and employing unique strategies, they are ready to achieve significant gains.
Disclaimer: Comments and opinions expressed by interviewees are their own and do not represent or reflect the opinions, policies, or positions of PreScouter or have its endorsement.
