One interesting way that fibers are not immune to RF noise is in the presence of lightning strikes [1]. The strong electromagnetic fields can cause a polarization rotation inside the fiber, which the DSP at the receiving end of coherent links has to track.
Spectral efficiency is also a nifty metric because, given some bandwidth and signal-to-noise ratio, there is an upper bound to which you can compare a result [0].
Frequency spacing makes sense as namibj points out. Most long-distance telecom links operate in the optical C-band, which is roughly 5 THz wide. (A wavelength of 1525nm has an optical frequency of 196.5 THz, and a wavelength of 1565nm has an optical frequency of 191.5 THz). You can select optical frequencies to modulate within this optical bandwidth. Given a certain modulation rate (>>GHz), separating the channels in units of 1 GHz is reasonable.
Figure 3 in this open-access paper [0] provides historical scaling trends in optical fiber communication (transport) compared to generation and processing. Depending on the time period under study, bandwidth increases at between 20% and 100% per year.
Though the improvement in transistor economics has definitely benefited transport, the large bulk of improvement over time is due to breakthroughs in manufacturing, materials science, semiconductor optics, and signal processing.
Howdy, I design optical integrated circuits for coherent communication. mmmBacon is correct. See here[0] or [1] for examples of the types of forward error correction used in coherent links. Pre-FEC bit error rate thresholds of 1e-3 or 2e-2 for <1e-15 post-FEC errors are fairly standard.
About using light: I don't believe light will replace the transistors performing logic except in a few niche applications (e.g. [1]). Light is physically constrained by it's wavelength. It's difficult to build interacting structures smaller than a few hundred nm and it's difficult to build light-generating elements with even shorter wavelengths--you're approaching the Deep UV and X-rays. Maybe you can get to the tens to low hundreds of nm with plasmonics, but this is still far from the realm in which it makes sense to replace an electronic transistor with an optical transistor. Furthermore, it's also difficult to achieve strong nonlinearities in optical systems, especially silicon. You need some sort of nonlinear element for switching.
Light-based communication probably will replace certain I/O blocks on chip. These tend to be quite large in terms of area after considering power and ESD constraints.
I think it's great that a data center operator is willing to relax their requirements. For too long we've been designing against telecom specs and operating environments.
I think in order to bring more OEM vendors in we need to see the other big players to also accept the relaxed specs.
Hopefully, we don't end up with another dozen different 100G or 200G MSAs that work from 15-55C.
I'm also curious what the pricing difference is for a CWDM4 transceiver and the OCP version.
I would guess the NRE to develop either is similar and that the design for either is almost the same. Perhaps Facebook is just trying to get the optics cost down by negotiating discounts on the non-yielding MSA parts that would have otherwise had to get thrown out?
Put it the other way: what would be the benefit of publicly disclosing their secret sauce? Perhaps to help fuel their recruiting pipeline, but I can't think of any other good reasons to do so.
They don't have a product out, their secret sauce might change, Google/MSR might be working on something similar and just need a few hints in the right direction to get there, they get to keep a big first-mover advantage in whatever area their tech enables them to move into, and there's no opportunity for possible early adopters like yourself to get in on the action yet.
I'm guessing their tech isn't so much "critically secret" as there just isn't a benefit to their business case for publicizing what they are doing
The Watts group at MIT has done some fantastic work into rare-earth doped silicon waveguides to produce lasers on a silicon platform [0]. However, I believe this work is still very much in the research stage. I'm not convinced their method is scalable to production for this $10 & million-unit-per-year LIDAR application since Erbium-doped waveguide lasers still require an off-chip pump laser source.
I believe that LIDAR will be to silicon photonics what the accelerometer was to MEMS.
The real challenge that the article only touched upon is to get lasers into the same package and keep the costs down. This is still an active area of pursuit in both research and industry--though, there are several promising methods emerging. The $10 cost used in the article is likely closer to the cost of the bare silicon die. Packaging is always the expensive part of optics (doubly so if the lasers are not monolithically or heterogeneously integrated onto the same die). That being said, even with today's technology, integrating a laser chip and a silicon photonics chip into a package is easily south of $1k, which is what they quoted competing technologies costing.
I look forward to seeing these sensors integrated into my self-driving car in 5-10 years.
Efficiently directing light into a fiber optic cable is a difficult problem, in general. The luminescence is their proposed solution to this problem and one of the key results in their journal article.
You have to shine light end-on into a fiber--you can't couple light into the fiber by shining on it at the side. Even for large, visible-light scintillating fiber used in this paper, you only have a few degrees of possible incidence angles that direct the light into the fiber. See for example, (one of the first results from Google), the figure on page 4 of this datasheet [0].
Yes, an interferometer is certainly a very common method to perform switching. What ultimately gets used will depend on the technology/material system. Silica-on-silicon and silicon photonics-based switches will likely use Mach-Zehnder interferometers. MEMS switches currently use movable mirrors or gratings.
Generally speaking, being able to split based on wavelength lets you transmit data on multiple wavelengths to increase your bandwidth. The flow is to modulate each wavelength individually, mux them together, send them through a single fiber or waveguide, and then demux them on the other side. In a switch, you could imagine switching each wavelength individually and optionally combining them into a single waveguide out of each port of the switch.
This particular device could not be used to make an interferometer. The device has 1 input (call it port 1) and 2 outputs (call them ports 2 and 3). If you input 1550 nm light to port 1, most of it goes to port 2. If you input 1310 nm light to port 1, most of it goes to port 3. This also works backwards: if you input 1550 nm light to port 2 most of it goes to port 1. If you input 1550 nm light to port 3, 10% of it goes to port 1 and the other 90% gets radiated outward as loss (crosstalk is -10 dB). So if you tried to input 1550 nm light to both ports 2 and 3 there won't be much interference at port 1 unless there is a large power imbalance between your two input beams.
I did my PhD work in silicon photonics (in a different lab group and not associated with the authors in the paper) and thought I could chime in with some extra background and why this result is interesting to the silicon photonics community.
First off, silicon photonics has already made its way into several products, mostly active optical cables (a device that directly converts an electrical signal to an optical signal). See, for example, Luxtera/Molex, Acacia, and Kotura/Mellanox. Additionally, many other companies have demoed interesting things at trade shows (e.g. Cisco, Intel, Fujitsu, and others).
In general, the appeal of silicon photonics is that we can fabricate almost all of the components of an optical link on a single chip using the same fabrication tools as what you might find in a standard CMOS fab. Modulators, detectors, switches, filters, and other devices have been demonstrated on a single wafer. Many organizations (ePIXfab, IBM, IME A*STAR, Intel, Freescale, and others) have fabrication processes that have all of these devices right next to each other on a wafer and are capable of 25+ Gb/s data transmit and receive.
Others in the comments have mentioned the lack of switches in the article. Making optical switches in silicon has been demonstrated before, usually with either a Mach-Zehnder interferometer or resonant structure. The phase of light or resonance are most commonly adjusted through the thermo-optic or plasma dispersion effect. I'm at work now, but I can dig up references if anyone is interested later.
This result by Piggot, et. al., is most interesting because it is a unique device geometry for performing a wavelength splitting function. The performance of the device itself isn't particularly impressive relative to other devices with similar functionality that have already been demonstrated [1]. Additionally, the use of an MMI structure for wavelength multiplexing is also not novel [2].
So how does this relate to "light-based computers?" The vision that places like IBM research try to sell is that we will eventually integrate photonics (either monolithically, or flipped in some form) onto our processors and memory chips to enable high-throughput on- and off-chip I/O. This is still likely 10 years away from commercial products. Near-term, look for silicon photonics in your data centers and fiber-optic regional, metro, and long-haul networks. (FTTx one day, but silicon photonics currently can't compete in economics with a DML shoved into a TO can.)
[1] PDF Warning: https://www.ofcconference.org/getattachment/d0ec1565-ce81-48...