By Geoff Bennett
Director, Solutions & Technology
ICE6 Helps Extract Every Bit of Bandwidth from Optical Networks
Optical technology is pushing the limits of capacity and reach, meaning every new generation of optical coherent transmitters and receivers delivers marginally better capacity and reach performance compared to the last. The reason for this is purely a matter of physics – we are approaching the non-linear Shannon limit in long-haul and subsea cable systems, so each innovation in coherent processing or optical components is focused on pulling out incremental fractions of a decibel in transmission budget to increase capacity-reach over a given fiber pair.
The Shannon limit describes the maximum theoretical capacity for a given optical cable.
Optical fiber is a non-linear medium and typically becomes the limiting term in modern coherent transmission.
Shannon limit aside, Infinera’s fourth-generation Infinite Capacity Engine (ICE4) continues to demonstrate record-breaking performance in multiple network trials and deployments. Check out the Light Reading webinar where Steve Grubb, Global Optical Architect at Facebook, and Infinera explain how ICE4 squeezes 26.2 terabits per second of capacity out of the 6,644-kilometer trans-Atlantic MAREA cable.
Infinera’s sixth-generation Infinite Capacity Engine (ICE6) enhances this experience with three key advances: Nyquist subcarriers, probabilistic constellation shaping (PCS) and vertical integration. Let’s take a closer look at how ICE6 overcomes the challenge of capacity-reach.
Second-generation Nyquist Subcarriers
Infinera’s ICE4 pioneered the use of Nyquist subcarriers, which you can read about in a recent Fibre Systems article. Unlike traditional optical signals, which experience higher optical penalties as baud rates increase, Nyquist subcarriers split the signal into multiple carriers with lower baud rates using the coherent processor in the transmitter. Lower-baud-rate subcarriers experience lower penalties than a high-baud-rate signal, meaning ICE4 and ICE6 transmission deliver higher optical performance, even when very tight channel spacing is used.
ICE6’s FlexCoherent® processor takes this to the next logical stage in order to increase the effective baud rate without incurring higher penalties. For example, an 88 gigabaud (Gbaud) system would implement eight subcarriers operating at ~11 Gbaud each. One challenge with high-speed transmission is that it’s not uncommon for the outer edges of a signal to face a more difficult optical environment that can limit the performance of single-carrier solutions. To address this challenge ICE6 uses dynamic bandwidth allocation (DBA) to rebalance each subcarrier payload independently based on its optimal tolerance to fiber impairments.
Subcarrier mastery demonstrates a clear advantage of Infinera’s lead in coherent optical implementation, and you can expect even more innovative uses for subcarriers in future ICE variants.
Highly Effective Probabilistic Constellation Shaping
PCS is a technique for optimizing modulation format. Instead of selecting a specific modulation constellation, such as quadrature phase-shift keying (QPSK), 16 quadrature amplitude modulation (16QAM), 32QAM, etc., PCS starts with a high-order constellation and dynamically adjusts how many bits per modulation symbol can be transmitted over a single wavelength to optimize capacity-reach performance. In order to accomplish this, there is a great deal of signal processing that must occur to adapt the right modulation rate to each wavelength.
PCS works by adjusting the probability of a given modulation symbol being transmitted, because not all symbol locations can be received with the same level of sensitivity. However, to be effective, a PCS implementation needs a significant level of processing power, which is determined by the size of the processing nodes on the chip itself – the smaller the better. So a 7 nanometer (nm) application-specific integrated circuit (ASIC) will automatically deliver higher-efficiency PCS than a 16 nm ASIC.
But PCS is just one way to use ASIC processing power to enhance coherent performance. Today, ICE4 is delivering around 30 percent better spectral efficiency than leading competitive platforms due to a combination of an advanced photonic integrated circuit (PIC) implementation and the use of Nyquist subcarriers. This combination was discussed in detail in the subsea Light Reading webinar. In fact, ICE4 performance is so good, it’s delivering about the same capacity-reach over modern subsea cables as the initial predictions for first-generation PCS implementations that are scheduled to begin shipping later this year.
The key point here is that getting the right balance of ASIC processing power is essential. The availability of 7 nm ASICs later this year will allow a far more effective implementation of PCS – one that will build on the value of ICE4’s Nyquist subcarrier-based efficiency to deliver even better optical performance.
Two additional enhancements will be included in ICE6’s PCS. First is the ability to adapt baud rates to be compatible with open line systems, which can operate over flexible grid or standard 50 gigahertz grids or aggressive reconfigurable optical add/drop multiplexer (ROADM) filters. Second is the simple ability to be able to turn off PCS because it’s not always the best approach to achieving the right combination of optical reach, capacity and power consumption.
ICE6 will build on the success of ICE4, adding highly effective PCS technology to supplement our high-performance subcarrier technology, which remains unique and differentiated in the market. Early indicators are already hinting that ICE6 performance will set a new level of excellence for coherent transmission.
Vertical Integration Is Essential for High Baud Rates
It is vital that coherent implementations deliver high baud rate operation and, more importantly, ensure seamless operation of the optics and digital signal processor (DSP) of an optical engine. Furthermore, understanding the nuance of the optical design coupling the hardware can be just as challenging as with software.
The use of Nyquist subcarriers means that a high baud rate signal experiences a lower non-linear penalty once it is launched onto the fiber. The optical transmitter, however, must operate at the full baud rate of the aggregated carrier, which in the case of ICE6 is 88 to 100 GBaud, delivering up to 800 gigabits per second per carrier.
For the past 20 years, Infinera has been an industry leader in the development and production of indium phosphide (InP) PIC technology. InP is the perfect choice for high-performance C- and L-Band lasers – something that is not possible using silicon photonics. But InP is also the best choice for high-performance modulators at very high baud rates because InP has far superior carrier mobility compared to silicon.
Taking these high-performance optics and combining them with leading-edge electronics must be done in the optical packaging. At very high baud rates, the timing tolerances between these components is measured in femtoseconds (10-15 seconds). So there is a clear advantage for a company to be able to design and develop all three of these technologies in house: optics, coherent electronics and leading-edge module packaging.
Infinera’s end result – ICE6 – represents a new generation of the cross-platform digital coherent optics (DCO) module that will be rolled out across all appropriate platforms in our portfolio.
Welcome to The Infinite Network
ICE4 and ICE6 represent cutting-edge optical innovation that leverages the power of Nyquist subcarriers, highly effective PCS and vertical integration. While our ICE4 and ICE6 optical engines are fundamental to Infinera, they are just one piece of the innovation that is part of the The Infinite Network, a vision to deliver networking solutions that can handle the insatiable need for bandwidth, deliver it instantly and ensure it is ubiquitous – all while leveraging the science of simplicity.
Why Choose Infinera?
Infinera is a unique network system vendor that leverages the advantages of vertical integration to deliver powerful, easy-to-use network solutions. We are backed by a global infrastructure that delivers end-to-end networking solutions to many of the world’s largest operators. We are recognized for delivering a network experience like no other, called the Infinera Experience, and we would love to share it with you. If you are a network operator looking for innovation, improved network economics and a world-class customer experience, contact us.