So how do they get spin into the lasers?
Do the sharks need special training?
Now, how to boost fibre throughput to a stonking 240Gbps? With frikkin' spin-lasers, of course
Researchers at the Ruhr-Universität Bochum in Germany have said they have developed a novel method of encoding information with lasers that could boost the amount of bandwidth sent down a strand of fibre to 240 gigabits per second. In contrast, conventional optical communications currently max out at around 40 to 50Gbps. The …
-
-
-
-
-
Wednesday 17th April 2019 21:15 GMT The Oncoming Scorn
Re: So how do they get spin into the lasers?
There used to be a pet store in Calgary that had a small shark* in a very large tank no idea as to the sub-specie), but if you looked at it, it certainly resembled a great white.
*At that size & within a thick glass tank was is as close as I ever want to be to one.
-
-
-
-
-
Wednesday 17th April 2019 09:56 GMT I.S.
Bollocks.
"In contrast, conventional optical communications currently max out at around 40 to 50Gbps."
Bullshit.
400Gbps Ethernet (standardized IEEE802.3 bs) has been shipping to data centre customers for the past year.
More complex versions of 400G have been deployed in DWDM backbone networks since 2014.
This allows in excess of 30Tbps per fibre!
The race has been to develop plug-able optics modules to fit the form factor demanded by data centre customers. Power consumption and heat dissipation are the biggest hurdles, but they've all been cleared.
-
Wednesday 17th April 2019 10:12 GMT Anonymous Coward
Re: Bollocks.
The point of this is that you can increase the bit rate *per transmitted frequency*. Yes you can fire multiple frequencies to up the overall bandwidth, but this allows each frequency itself to be increased too.
But you go ahead and assume you know best. Let me plump up that armchair for you.
-
Wednesday 17th April 2019 11:26 GMT Jellied Eel
Re: Bollocks.
Actually, I think the point is-
The paper claims to demonstrate that spin-lasers can enable room-temperature modulation frequencies above 200GHz
The article's perhaps not clear about the benefits, ie DWDM systems can push Tbps down a single fibre. So current DWDM is generally based on G.694.1 grids/frequency spacing with 25Ghz or 50Ghz channels, giving around 160 channels to play with.
So depending on how easy this is to integrate, could mean 260Gbps per channel, so maybe 40Tbps per fibre.. But that also requires the lasers to be stable, reliable, and of course.. Cheap. Especially if it means having to use a pair of lasers per channel to benefit from the 200Ghz+ modulation. And then there's power requirements and cooling.
Or it could mean 260Gbps 'SFP's for squeezing 2x100Gbps Ethernets down a single DC patch fibre.
-
Wednesday 17th April 2019 21:41 GMT Kernel
Re: Bollocks.
"The point of this is that you can increase the bit rate *per transmitted frequency*. Yes you can fire multiple frequencies to up the overall bandwidth, but this allows each frequency itself to be increased too.
But you go ahead and assume you know best. Let me plump up that armchair for you."
I think you may be the one assuming that you know best.
Current DWDM systems transmit at up to 500Gbs using paired frequencies or 260Gbs per single frequency - the systems I work with transmit up to 99 different frequencies (we usually call them 'lambdas' in the trade), with 260Gbs per lambda currently being the more common.
The company I work for is actually in the process of phasing out some of its older 260Gbs transponders and replacing them with improved items - longer reach, better error correction. So yeah, 260Gbs per lambda, over distances measured in 100's of km without regeneration and with remotely controllable routing of individual lambdas along the way, is old tech.
Such products have been available from companies such as Cisco, Cienna, Nokia (Alcatel-Lucent) and Huawei for several years.
-
-
-
Wednesday 17th April 2019 12:34 GMT devTrail
Instead of boosting the bit rate
Since the technology does not rely on the intensity of the laser pulse I wonder whether it works with dim pulses. Instead of concentrating the effort on boosting the throughput could it be used to tolerate signal attenuation and reduce the number of repeaters required along the line? I guess that in this way the cost of the network could be reduced quite a lot.
-
Wednesday 17th April 2019 14:29 GMT Jellied Eel
Re: Instead of boosting the bit rate
I guess it depends. So spin messes with polarisation, which means polarisation mode dispersion (PMD) seems critical with the fibre. Attenuation is a limiting factor, but also chromatic dispersion. So the trick is to produce lasers that are compatible with current ITU grids, and also plays nicely with amps/repeaters. Cost of replacing those elements is often a lot less than having to lay new fibre. And for existing networks, there's already existing repeater sites. So squeezing say, another 10km out of a span may mean having to find new sites. And then shifting customer circuits, interconnects etc without extended service outages.
Plus side, that can reduce the need to squeeze new kit into existing, often cramped locations.
-
Wednesday 17th April 2019 16:20 GMT devTrail
Re: Instead of boosting the bit rate
Cost of replacing those elements is often a lot less than having to lay new fibre. And for existing networks, there's already existing repeater sites. So squeezing say, another 10km out of a span may mean having to find new sites. And then shifting customer circuits, interconnects etc without extended service outages.
Don't forget those area with low population density that that still don't have a network because the cost per customer is too high.
-
Thursday 18th April 2019 07:52 GMT Jellied Eel
Re: Instead of boosting the bit rate
Don't forget those area with low population density that that still don't have a network because the cost per customer is too high.
Yup. Some of this stuff is potentially 'backbone' technology, ie ULH kit to extend fibre spans. But that tends to be city-city, or landing station-city. So in the UK, there's a lot of fibre capacity running through Cornwall, but often heading straight for London non-stop. Challenge like you say is creating the service PoPs that mux down to consumer (or business) speeds.
New developments like this may help, if they can be used to cost-effectively aggregate low speed connections to feed into backbone networks. In that space, distance can be less critical, ie <=40km on SMF, but that can still be problematic with rural areas, especially given cable routing is rarely straight line. Developments in PON (Passive Optical Networking) have helped drive costs down though.
-
-
-
-
-
Thursday 18th April 2019 05:10 GMT Schultz
Hot air
The photon spin is just another nomenclature for the photon polarization. You can potentially distinguish 3 polarization (or spin) states: right-handed, left- handed and linearly polarized. Using this to triple data transmission through a fiber is tricky: does you fiber transmit those three without distortion?
I don't have access to this paper right now, but related papers discussed in the Register in the past made purely fictitious claims about creating many angular polarization states. This is nonsense and if you want to understand why, then read about "weak and strong measurements". You need to average multiple measurements to detect precise particle properties (e. g., A definite photon polarization) beyond Heisenberg's uncertainty principle. The time this takes will make up for the extra information you gained - Heisenberg won't allow you to cheat.
Biting the hand that feeds IT
