A new architecture for high speed core-selective switch for multicore fibers

Crist\'obal Melo; Diego Arroyo; Esteban S. G\'omez; Gabriel Saavedra; Gustavo Lima; Jaime Cari\~ne; Mat\'ias Reyes. F.; Miguel Figueroa; Stephen P. Walborn

REVIEW 5 major objections 4 minor 22 references

A new architecture for high speed core-selective switch for multicore fibers

T0 review · 5 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A four-path interference switch built from multicore fiber beam splitters routes optical signals between cores of a multicore fiber in under 0.7 microseconds, and the paper demonstrates it error-free on a field-installed multicore fiber…

desk verdict A genuinely faster MCF core-selective switch with a real field demo; the evidence supports the architecture, but the abstract's crosstalk number and the field 'error-free' claim are ahead of the data shown. read the letter →

arxiv 2411.17641 v1 pith:ARB237Q3 submitted 2024-11-26 quant-ph eess.SP

Cristóbal Melo , Matías Reyes. F. , Diego Arroyo , Esteban S. Gómez , Stephen P. Walborn , Gustavo Lima , Miguel Figueroa , Jaime Cariñe

show 1 more author

Gabriel Saavedra

This is my paper · ORCID

classification quant-pheess.SP

keywords multicorefibercore-selectiveswitchspace-divisionmultiplexingfour-pathMach-Zehnderinterferometerphasestabilizationopticalswitchingfield-installednetworkWDMcompatibility

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper is trying to establish that a core-selective switch for multicore fibers can be built from a four-path Mach–Zehnder interferometer that uses multicore fiber beam splitters to split and recombine light, with electro-optic phase modulators choosing which output core receives the signal. The claimed result is a switching time below 0.7 µs, roughly three orders of magnitude faster than existing multicore fiber switches, with average inter-core crosstalk around −16 dB and insertion loss of 7.7 dB. This matters because high-speed core switching is a missing piece for space-division-multiplexed networks; the paper further reports routing a 1 Gbps telecom signal error-free through the switch in a back-to-back test and over an actual campus-installed multicore fiber network. A sympathetic reader would take the central claim to be that MCF switching can be made fast enough for burst switching and data-center networking while remaining compatible with standard WDM.

What carries the argument

The central object is a four-path Mach–Zehnder interferometer formed by two four-core multicore fiber beam splitters (MCF-BS), with three electro-optic phase modulators and one reference path in between. A digital FPGA control system applies phases via the modulators to select the output core and runs a perturb-and-observe algorithm to cancel environmental phase fluctuations; the unitary matrix description in Eqs. (1)–(2) treats the beam splitter as a fixed unitary $M_{\mathrm{BS}}^{4\times4}$ and the phase shifts as a diagonal matrix $M_\theta$, so output core selection reduces to choosing relative phases $\theta_i$.

What would settle it

Leave the switch running for an hour in an environment with ordinary temperature drift and monitor per-core output power at a fixed phase setting; if the feedback loop cannot hold the four-path interference visibility near its measured 0.97–0.98 level and power leaks into non-selected cores beyond the reported −12 to −22 dB range, the balancing premise fails in practice.

Watch

Extended reading notes

Core claim

The central discovery is that a four-core multicore fiber beam splitter can serve as the splitting and recombining element of a fast optical switch: the first MCF beam splitter divides the input into four paths, three of which contain GHz-bandwidth lithium-niobate phase modulators, and the second recombines them so that interference directs all power to the chosen core. With an FPGA feedback loop applying perturb-and-observe phase corrections, the interferometer remains stabilized and can hop among all four output cores every 2.5 µs with a rise time under 0.7 µs. The authors report 19.8 dB extinction, −16.25 dB average inter-core crosstalk (range −12 to −22 dB), and 0.97–0.98 interference visibility, and they show the device works with a commercial SFP+ transceiver at 1 Gbps in the lab and over 170 m and 1,305 m field-installed multicore links with error-free transmission. To the authors' knowledge this is the first demonstration of a high-speed MCF switch operating in real-world conditions.

Load-bearing premise

The four interferometer paths must stay balanced (equal optical path length, equal loss, equal polarization response) so the second multicore beam splitter combines them coherently; if drift outruns the feedback loop, interference visibility, crosstalk, and core selection all degrade.

Editorial extensions

If this is right

MCF core selection can be carried out at burst-switching timescales, making multicore fiber viable for data-center and high-capacity metro networks that require sub-10 µs reconfiguration.
Because the switch operates on all wavelengths together and maintains >0.99 visibility over roughly 1540–1560 nm, it can be inserted into WDM systems without per-wavelength switching.
Error-free 1 Gbps routing over installed campus MCF links with crosstalk of −18 to −22 dB indicates the switch does not add a penalty beyond expected attenuation.
The architecture uses standard fiber components and an FPGA control loop, so it can be packaged in a standard 19-inch rack shelf compatible with existing telecom infrastructure.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The insertion-loss bottleneck is the commercial lithium-niobate modulators (about 3.3 dB); replacing them with lower-loss phase shifters could bring total loss closer to the 2.2 dB level of the passive splitter/recombiner stages.
The 0.8 MSps feedback update rate sets an upper bound on how fast environmental perturbations can be tracked; faster feedback or feed-forward phase control would be needed for outdoor or vibration-heavy deployments.
The same four-path interferometer could be repurposed for quantum information tasks, such as actively routing single photons between multicore fiber cores, since interference visibility of 0.97–0.98 is already suitable for quantum interference.
Combining this switch with few-mode multicore fibers may extend core-selective routing into spatial-mode-selective routing, going from one dimension of SDM switching to two.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Read it. The device is real and the core result stands: a four-path Mach-Zehnder built from multicore-fiber beam splitters and LiNbO3 phase modulators, switched in under 0.7 us and routed over an installed campus MCF link. That speed is three orders faster than the LPG, beam-steering, and mechanical alternatives they cite, and the field test is legitimately new.

Credit where due: the paper shows interference visibilities of 0.97 and 0.98, BER curves for all four output cores with 0.4 to 0.9 dB penalty, crosstalk between -12 and -22 dB, and a 15-second active-stabilization trace. The architecture has no hidden free parameters: the phase settings are measured, and Eq. (1) takes the MCF-BS matrix from the earlier component characterization. The use of internal polarizers and per-path polarization control is a sensible way to make a coherent interferometer out of telecom parts.

Soft spots, in proportion. First, the abstract says average crosstalk is below -18 dB, but the body reports -16.25 dB. That is a small but real inconsistency. Second, the field "error-free" claim is not supported by the evidence shown. Figure 4b is two averaged photodiode traces at 150 MHz bandwidth; there is no BER curve, no PRBS error count, and no indication the switch was toggled at speed during the 15-second window. The stress-test note lands here. This is the biggest gap. Third, WDM compatibility is shown only as visibility versus wavelength. There is no crosstalk or insertion-loss data across the band and no test with a modulated WDM comb, so "fully compatible" is too strong. The visibility also dips to about 0.90 at 1527 nm, so the range is narrower than implied. Fourth, the 0.7 us rise time comes from a single trace in Fig. 2c; a second measurement or an error bar would help. Fifth, the stability demonstration is only 15 seconds; long-term thermal drift under field conditions is not quantified. That is not fatal, because the feedback loop is the right instrument, but an hour-long trace would make the claim solid.

The math and the citation pattern look clean. Equations (1) and (2) are standard unitary interferometer algebra, and the MCF-BS matrix is borrowed from a prior characterization, which is fine. The load-bearing premise is that the four paths stay balanced over time, and the measured visibilities support that premise for the windows shown.

Who gets value from this: people working on MCF networking, SDM switching, and field trials of spatial-division devices. It deserves a serious referee. I would send it out, with a request that the field BER evidence and the WDM crosstalk-versus-wavelength data be added before acceptance.

Referee Report

5 major / 4 minor

Summary. The paper presents a multicore-fiber core-selective switch built as a four-path Mach-Zehnder interferometer, with two multicore-fiber beam splitters, lithium-niobate phase modulators, and an FPGA-based stabilization loop. The authors report a switching rise time below 0.7 microseconds, an average inter-core crosstalk of -16.25 dB, an insertion loss of 7.7 dB, and back-to-back transmission of a 1 Gbps PRBS signal at BER 1e-9 with about 0.6 dB sensitivity penalty. They also report routing a 1 Gbps signal over a field-installed multicore fiber network on a university campus and claim this is the first demonstration of a high-speed MCF switch operating under real-world field conditions, with speeds three orders of magnitude faster than prior MCF switches. The main results are experimental measurements rather than a new theoretical derivation, with the interferometer model in Eq. (1) borrowed from earlier MCF beam-splitter characterization.

Significance. If fully substantiated, the result is significant for space-division-multiplexed networks: a core-selective switch with sub-microsecond switching would be a large practical improvement over millisecond-class mechanical and beam-steering switches. The paper's strengths are its direct experimental approach, the use of commercial components, the back-to-back BER characterization, and the ecological validity of the field-installed MCF test. The reported interference visibilities of 0.97-0.98 support the core unitary-combiner assumption. However, several headline claims are currently backed by incomplete evidence, so the significance is conditional on the missing measurements being supplied.

major comments (5)

[Abstract / 'Core switching' Results] The abstract states 'less than -18 dB of average inter-core crosstalk,' while the Core switching section reports an average IC-XT of -16.25 dB with a range of -12 to -22 dB. These numbers are mutually inconsistent; since average crosstalk is a headline specification, the abstract must be corrected to match the measured value or the measurement must be re-derived.
[Operation in optical networks, Fig. 4] The field-test claim of 'error-free transmission' is not supported by the evidence shown. Fig. 4b presents only low-bandwidth photodiode traces of the received signal; there is no BER curve, no PRBS error count, no statement about clock/data recovery, and no indication of whether the switch was actively toggled during the 15 s window. Without a quantified error measurement and a description of the switching pattern during the field test, the central claim of a field demonstration with error-free 1 Gbps routing is not established.
[Core switching / Methods, WDM compatibility] The WDM compatibility claim rests on a single visibility scan over wavelengths from 1527 to 1569 nm, with no crosstalk, insertion loss, or BER data reported for detuned wavelengths and no test with a modulated WDM comb. The abstract's statement that the design is 'fully compatible with standard multiplexing techniques' therefore goes beyond the presented evidence; either add a WDM transmission test or soften the claim.
[Core switching, Fig. 2c] The 0.7 microsecond rise time appears to be extracted from a single oscilloscope trace (Fig. 2c shows 0.68125 us) with no definition of the rise-time threshold, no repeated switching events, and no statistics across the four output cores. Because the three-orders-of-magnitude speed advantage is a core claim, the rise time should be reported as a distribution over many switching events with a clear 10-90% definition.
[Active phase stabilization, Fig. 2a] The active-stabilization demonstration in Fig. 2a covers only 15 s, while the switching and field experiments do not report how long the feedback loop maintained the selected output state. Since the device requires balanced optical paths and the control loop operates at 0.8 MSps, the absence of a stability timescale leaves the long-term field operability claim under-supported.

minor comments (4)

[Main text, Abstract and Introduction] There are several typographical and terminology issues: 'wave-division multiplexing' should be 'wavelength-division multiplexing', '1GBs' should be '1 Gbps', and the sentence 'this is the first time a a signal is directed...' contains a duplicated article.
[Fig. 5c and Core switching text] The figure axis for the C-band visibility plot shows 1530-1560 nm, while the text states wavelengths between 1527 and 1569 nm; please make the figure and text consistent.
[Methods, Active phase stabilization] The phase-stabilization description would benefit from a circuit-level block diagram or a more explicit timing analysis; the stated 0.8 MSps conversion rate and the mention of RF filters are not sufficient to assess the loop bandwidth and settling behavior.
[Equation (1) and (2)] Eq. (1) uses the same symbol MBS4x4 for the first and second multicore beam splitter without noting whether they are nominally identical devices; a sentence confirming the assumption would clarify the model.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the reported switch performance is directly measured, and the cited prior MCF-BS characterization is independent component evidence rather than a fitted or self-defined premise.

full rationale

The derivation chain is not circular. Eq. (1) uses MBS4x4 from Ref. 17, but that matrix is a prior experimental characterization of the multicore-fiber beam splitter, not a parameter fitted to the present switch outputs; the citation is independent support for a component property. The control phases are obtained online by a perturb-and-observe search, not by fitting the model to the reported BER, crosstalk, or rise time. All central claims—less than 0.7 µs rise time, -16.25 dB average inter-core crosstalk, 19.8 dB extinction ratio, 7.7 dB insertion loss, and 10^-9 BER—are direct measurements from photodiodes, an oscilloscope, power meters, and the FPGA bit-error counter. The 0.97-0.98 visibility and C-band visibility curve are separate interferometer characterizations, not quantities forced by the switching data. The field 'error-free' claim would benefit from a displayed BER curve, but that is a reporting/evidence gap, not a circular reduction: the claim is neither defined in terms of the measured output powers nor derived from a fitted parameter. No load-bearing step reduces by construction to its own input.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central claims are experimental, so there are no model fits. The key dependencies are the unitary model of the MCF beam splitter (from ref 17), the assumption of balanced paths, and the manufacturer-specified phase modulator bandwidth. No new physical entities are introduced.

free parameters (1)

Core-selection phase shifts theta_PM,i = Not reported; determined empirically by perturb-and-observe
The switch's ability to route to each output core depends on a set of phase shifts applied by the three phase modulators. The paper does not list these values; they are found by a feedback algorithm, so a reproduction must re-derive them. This is a hand-chosen input, not a fitted model parameter.

assumptions (4)

domain assumption The multicore fiber beam splitter behaves as a unitary 4x4 device with matrix M_BS4x4 as characterized in ref 17.
Eq. (1) in Methods models the four-core beam splitter with a unitary 4x4 matrix taken from the authors' prior work. If the splitter has excess loss or non-unitary mode coupling, the interference model and the resulting output routing are invalid.
domain assumption The four optical paths have equal optical path length, equal loss, and equal polarization response.
The Design section requires these equalities for perfect interference, and Eq. (2) assumes equal loss and polarization on each path. Delay lines and polarization controllers approximate this, and measured visibility 0.97-0.98 partially supports it.
domain assumption The phase modulators have GHz bandwidth and can apply the required phase shifts quickly enough for 0.7 microsecond switching.
The switching speed claim depends on the phase modulators' bandwidth, assumed from the manufacturer datasheet (ref 18) rather than independently measured in this paper.
domain assumption The optical path length differences between interferometer paths are smaller than the duration of the 1 Gbps information pulses.
The back-to-back section states this condition explicitly as necessary for interference with the modulated signal; it is assumed but not directly measured.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A new architecture for high speed core-selective switch for multicore fibers." pith.science (2026). https://pith.science/paper/ARB237Q3

@misc{pith2026241117641,
  author       = {Pith},
  title        = {Pith review of: A new architecture for high speed core-selective switch for multicore fibers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ARB237Q3}},
  note         = {Machine review of arXiv:2411.17641}
}

read the original abstract

The use of multicore optical fibers is now recognized as one of the most promising methods to implement the space-division multiplexing techniques required to overcome the impending capacity limit of conventional single-mode optical fibers. Nonetheless, new devices for networking operations compatible with these fibers will be required in order to implement the next-generation high-capacity optical networks. In this work, we develop a new architecture to build a high-speed core-selective switch, critical for efficiently distributing signals over the network. The device relies on multicore interference, and can change among outputs in less than 0.7 us, while achieving less than -18 dB of average inter-core crosstalk, making it compatible with a wide range of network switching tasks. The functionality of the device was demonstrated by routing a 1GBs optical signal and by successfully switching signals over a field-installed multicore fiber network. Our results demonstrate for the first time the operation of a multicore optical fiber switch functioning under real-world conditions, with switching speeds that are three orders of magnitude faster than current commercial devices. This new optical switch design is also fully compatible with standard multiplexing techniques and, thus, represents an important achievement towards the integration of high-capacity multicore telecommunication networks.

Figures

Figures reproduced from arXiv: 2411.17641 by the authors.

**Figure 1.** Implementation and Operation of the Multicore Fiber Switch. (a), Switch concept between 2 multicore fibers. (b), Photo of MCF switch implemented in the laboratory. (c), Conceptual design of the MCF switch. 2/8 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 2.** Output power for each core as a function of time. (a), Output power for every core with: no phase stabilization (0-7 s), active phase stabilization (7-15 s). (b), Core switching at 2.5 µs. (c), Rise time of the switching operation. core j 20. An average IC-XT value of -16.25 dB was observed, with a minimum of -22 dB and a maximum of -12 dB. Finally, a second tuneable CW laser was used to evaluate WDM compatibility. … view at source ↗

**Figure 3.** Transmission of 1 Gbps telecommunication signal. (a), Transmission setup. (b), Received bit error rate as a function of the received power. 4/8 [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Transmission of 1 Gbps telecommunication signal over MCF network. (a), Overview of the installed MCF links at Universidad de Concepción. (b), Received optical power after propagation for both transmission links. Given the favorable characteristics shown by MCF switch i…

**Figure 5.** Figure 5: Characterization of interference in a four-path Mach Zehnder interferometer. (a), Interference patterns between two paths. (b), Stabilized interference between all 4 paths. (c), Visibility across the C-band. Active phase stabilization As the output of the device depend…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 15 canonical work pages

[1]

Richardson, D. J. Filling the light pipe. Science 330, 327–328, DOI: 10.1126/science.1191708 (2010). https://science. sciencemag.org/content/330/6002/327.full.pdf

work page doi:10.1126/science.1191708 2010
[2]

J., Fini, J

Richardson, D. J., Fini, J. M. & Nelson, L. E. Space-division multiplexing in optical fibres. Nat. Photonics 7, 354–362, DOI: 10.1038/nphoton.2013.94 (2013)

doi:10.1038/nphoton.2013.94 2013
[3]

J., Rademacher, G

Puttnam, B. J., Rademacher, G. & Luís, R. S. Space-division multiplexing for optical fiber communications. Optica 8, 1186–1203, DOI: 10.1364/OPTICA.427631 (2021)

work page doi:10.1364/optica.427631 2021
[4]

Winzer, P. J. & Neilson, D. T. From scaling disparities to integrated parallelism: A decathlon for a decade. J. Light. Technol. 35, 1099–1115 (2017)

work page 2017
[5]

Mendinueta, J. M. D., Shinada, S., Hirota, Y ., Furukawa, H. & Wada, N. High-capacity super-channel-enabled multi-core fiber optical switching system for converged inter/intra data center and edge optical networks. IEEE J. Sel. Top. Quantum Electron. 26, 1–13, DOI: 10.1109/JSTQE.2020.2969558 (2020)

arXiv 2020
[6]

Rademacher, G. et al. 10.66 peta-bit/s transmission over a 38-core-three-mode fiber. In2020 Optical Fiber Communications Conference and Exhibition (OFC), 1–3 (2020)

work page 2020
[7]

Soma, D. et al. 10.16-peta-b/s dense sdm/wdm transmission over 6-mode 19-core fiber across the c +l band. J. Light. Technol. 36, 1362–1368 (2018)

work page 2018
[8]

Ryf, R. et al. Transmission over randomly-coupled 4-core fiber in field-deployed multi-core fiber cable. In 2020 European Conference on Optical Communications (ECOC) , 1–4, DOI: 10.1109/ECOC48923.2020.9333277 (2020)

arXiv 2020

Show all 22 references

[9]

Takeshita, H. et al. Demonstration of uncoupled 4-core multicore fiber in submarine cable prototype with integrated multicore edfa. J. Light. Technol. 41, 980–988, DOI: 10.1109/JLT.2022.3195190 (2023)

2023
[10]

& Suzuki, M

Morita, I., Igarashi, K., Takahashi, H., Tsuritani, T. & Suzuki, M. Trans-oceanic class ultra-long-haul transmission using multi-core fiber. Opt. Express 22, 31761–31773, DOI: 10.1364/OE.22.031761 (2014)

2014 doi
[11]

Marom, D. M. & Blau, M. Switching solutions for wdm-sdm optical networks. IEEE Commun. Mag. 53, 60–68, DOI: 10.1109/MCOM.2015.7045392 (2015)

2015
[12]

Huo, L. et al. Reconfigurable inter-core signal switching within multicore fibers based on long-period gratings. J. Light. Technol. 37, 6025–6032, DOI: 10.1109/JLT.2019.2945178 (2019). 7/8

2019
[13]

& Zervas, G

Deakin, C., Enrico, M., Parsons, N. & Zervas, G. Design and analysis of beam steering multicore fiber optical switches. J. Light. Technol. 37, 1954–1963, DOI: 10.1109/JLT.2019.2896318 (2019)

2019
[14]

Mulvad, H. C. H. et al. Beam-steering all-optical switch for multi-core fibers. In 2017 Optical Fiber Communications Conference and Exhibition (OFC), 1–3 (2017)

2017
[15]

& Katayama, K

Fukai, C., Abe, Y ., Uematsu, T., Ogushi, I. & Katayama, K. Multi-core fiber rotated optical switch. Opt. Fiber Technol. 81, 103470, DOI: https://doi.org/10.1016/j.yofte.2023.103470 (2023)

2023
[16]

Ben Yoo, S. J. Prospects and challenges of photonic switching in data centers and computing systems. J. Light. Technol. 40, 2214–2243, DOI: 10.1109/JLT.2021.3136570 (2022)

2022
[17]

Cariñe, J. et al. Multi-core fiber integrated multi-port beam splitters for quantum information processing. Optica 7, 542–550, DOI: 10.1364/OPTICA.388912 (2020)

2020 doi
[18]

LN65S-FC - Lithium Niobate Modulator

Thorlabs. LN65S-FC - Lithium Niobate Modulator. https://www.thorlabs.com/thorproduct.cfm?partnumber=LN65S-FC
[19]

& Krishnamurthy, A

Zhang, Q., Liu, V ., Zeng, H. & Krishnamurthy, A. High-resolution measurement of data center microbursts. InProceedings of the 2017 Internet Measurement Conference , IMC ’17, 78–85, DOI: 10.1145/3131365.3131375 (Association for Computing Machinery, New York, NY , USA, 2017)

2017
[20]

& Matsuo, S

Saitoh, K. & Matsuo, S. Multicore fiber technology. J. Light. Technol. 34, 55–66, DOI: 10.1109/JLT.2015.2466444 (2016)

2016
[21]

T., Jusoh A

Alik R, S. T., Jusoh A. A review on perturb and observe maximum power point tracking in photovoltaic system.Telecommun Comput. Electron Control. 13, 745–751, DOI: 10.12928/telkomnika.v13i3.1439 (2016)

2016 doi
[22]

OpenStreetMap

OpenStreetMap contributors. OpenStreetMap. https://www.openstreetmap.org/copyright (2024). Acknowledgements This work was supported by Fondo Nacional de Desarrollo Científico y Tecnológico (ANID) (Grants No. 1200266, 1240746, 1240843, 1231826, 1220960, 1231940), ANID – Millenn...

2024

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Richardson, D. J. Filling the light pipe. Science 330, 327–328, DOI: 10.1126/science.1191708 (2010). https://science. sciencemag.org/content/330/6002/327.full.pdf

work page doi:10.1126/science.1191708 2010

[2] [2]

J., Fini, J

Richardson, D. J., Fini, J. M. & Nelson, L. E. Space-division multiplexing in optical fibres. Nat. Photonics 7, 354–362, DOI: 10.1038/nphoton.2013.94 (2013)

doi:10.1038/nphoton.2013.94 2013

[3] [3]

J., Rademacher, G

Puttnam, B. J., Rademacher, G. & Luís, R. S. Space-division multiplexing for optical fiber communications. Optica 8, 1186–1203, DOI: 10.1364/OPTICA.427631 (2021)

work page doi:10.1364/optica.427631 2021

[4] [4]

Winzer, P. J. & Neilson, D. T. From scaling disparities to integrated parallelism: A decathlon for a decade. J. Light. Technol. 35, 1099–1115 (2017)

work page 2017

[5] [5]

Mendinueta, J. M. D., Shinada, S., Hirota, Y ., Furukawa, H. & Wada, N. High-capacity super-channel-enabled multi-core fiber optical switching system for converged inter/intra data center and edge optical networks. IEEE J. Sel. Top. Quantum Electron. 26, 1–13, DOI: 10.1109/JSTQE.2020.2969558 (2020)

arXiv 2020

[6] [6]

Rademacher, G. et al. 10.66 peta-bit/s transmission over a 38-core-three-mode fiber. In2020 Optical Fiber Communications Conference and Exhibition (OFC), 1–3 (2020)

work page 2020

[7] [7]

Soma, D. et al. 10.16-peta-b/s dense sdm/wdm transmission over 6-mode 19-core fiber across the c +l band. J. Light. Technol. 36, 1362–1368 (2018)

work page 2018

[8] [8]

Ryf, R. et al. Transmission over randomly-coupled 4-core fiber in field-deployed multi-core fiber cable. In 2020 European Conference on Optical Communications (ECOC) , 1–4, DOI: 10.1109/ECOC48923.2020.9333277 (2020)

arXiv 2020