Toward SDM-Based Submarine Optical Networks: A Review of Their Evolution and Upcoming Trends

Abstract

Submarine networks have evolved alongside terrestrial ones over the past several decades. Although there are similarities between these two network categories (e.g., the need to cover ultra-long-haul distances and transport huge amounts of data), there are also important differences that have dictated their different evolutionary paths. Space division multiplexing (SDM) promises to be the ultimate solution to cover future capacity needs and overcome problems of both networks. In this work, we review recent and future submarine technologies, focusing on all critical sectors: cable systems, amplifiers' technology, submarine network architectures, electrical power-feeding issues, economics, and security. Such an analysis, with the level of detail provided in this manuscript, is not available in the literature so far. We first overview all recently announced SDM-based submarine cable systems, compare their performance (capacity-distance product), and analyze the reasons that led to the first SDM submarine deployment. Also, we report up-to-date experimental results of submarine transmission demonstrations and perform a qualitative catego-rization that relies on their features. Moreover, based on all latest advances and our study findings, we try to predict the future of SDM submarine optical networks mainly in the fields of fiber types, fiber counts per cable, fiber-coating variants, modulation formats, as well as the type and layout structure of optical amplifiers. More specifically, results show that SDM can offer higher capacities (in order of Pb/s) compared to its counterparts, supported by novel network technologies: pump-farming amplification schemes, high counts up to 50 parallel fiber pairs, thinner fiber coating variants (200 μm), and optimum spectral efficiency (2-3 b/s/Hz). Finally, we conclude that tradeoffs between capacity and implementation complexity and cost will have to be carefully considered for future deployments of submarine cable systems.

Full text



Citation: Papapavlou, C.;
Paximadis, K.; Uzunidis, D.;
Tomkos, I. Toward SDM-Based
Submarine Optical Networks: A
Review of Their Evolution and
Upcoming Trends. Telecom 2022,3,
234–280. https://doi.org/
10.3390/telecom3020015
Academic Editors: Carlos Marques
and Sotirios K. Goudos
Received: 8 March 2022
Accepted: 5 April 2022
Published: 11 April 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Toward SDM-Based Submarine Optical Networks: A Review of
Their Evolution and Upcoming Trends
Charalampos Papapavlou 1, * , Konstantinos Paximadis 2, Dimitrios Uzunidis 1and Ioannis Tomkos 1
1Department of Electrical and Computer Engineering, University of Patras, 26504 Patras, Greece;
duzunidis@hotmail.com (D.U.); itom@ece.upatras.gr (I.T.)
2Department of Science & Technology, Hellenic Open University, 26335 Patras, Greece;
paximadis.konstantinos@ac.eap.gr
*Correspondence: c.papapavlou@upatras.gr
Abstract:
Submarine networks have evolved alongside terrestrial ones over the past several decades.
Although there are similarities between these two network categories (e.g., the need to cover ultra-
long-haul distances and transport huge amounts of data), there are also important differences that
have dictated their different evolutionary paths. Space division multiplexing (SDM) promises to be
the ultimate solution to cover future capacity needs and overcome problems of both networks. In
this work, we review recent and future submarine technologies, focusing on all critical sectors: cable
systems, amplifiers’ technology, submarine network architectures, electrical power- feeding issues,
economics, and security. Such an analysis, with the level of detail provided in this manuscript, is not
available in the literature so far. We first overview all recently announced SDM-based submarine
cable systems, compare their performance (capacity-distance product), and analyze the reasons
that led to the first SDM submarine deployment. Also, we report up-to-date experimental results
of submarine transmission demonstrations and perform a qualitative categorization that relies on
their features. Moreover, based on all latest advances and our study findings, we try to predict
the future of SDM submarine optical networks mainly in the fields of fiber types, fiber counts per
cable, fiber-coating variants, modulation formats, as well as the type and layout structure of optical
amplifiers. More specifically, results show that SDM can offer higher capacities (in order of Pb/s)
compared to its counterparts, supported by novel network technologies: pump-farming amplification
schemes, high counts up to 50 parallel fiber pairs, thinner fiber coating variants (200
µ
m), and
optimum spectral efficiency (2–3 b/s/Hz). Finally, we conclude that tradeoffs between capacity and
implementation complexity and cost will have to be carefully considered for future deployments of
submarine cable systems.
Keywords:
fiber pairs (FPs); multi-core erbium-doped fiber amplifiers (MC-EDFA); pump-farming
technology; power-feeding equipment (PFE); colorless directionless contentionless reconfigurable
optical add/drop multiplexers (CDC-ROADMs)
1. Introduction
Approximately one and a half centuries ago, the first submarine cables were sub-
merged to transmit telegraphy data [
1
]. Currently, with millions of people and applications
communicating in real time with each other across all continents, submarine cable systems
have become not only the most crowded, yet isolated, of deep-water networks, but they also
serve as a basic component of the whole global backbone network infrastructure [
2
,
3
]. Re-
cently, the traffic generated by end-users has been boosted beyond the average 50% annual
growth rate [
4
], due to the COVID-19 pandemic. With so many people interacting through
the internet, the need for speed (and thus more bandwidth) is increasing exponentially.
Space division multiplexing (SDM) was introduced in 2009 (A. R. Chraplyvy, “The
coming capacity crunch”, Proc. ECOC, plenary talk, September 2009) as the way to address
Telecom 2022,3, 234–280. https://doi.org/10.3390/telecom3020015 https://www.mdpi.com/journal/telecom

Telecom 2022,3235
the anticipated capacity crunch of the coming decades. It makes use of multiple spatial
channels which can be physically placed in bundles of standard single-mode fibers (Bu-
SMFs), as well as in multi-core fibers (MCFs) and multi-mode fibers (MMFs) [
4
]. MCFs
and MMFs are currently under investigation for future use as they not only multiply the
capacity of optical networks but also achieve lower costs and power consumption (per
transmitted bit) [
3
]. From this point of view, SDM can provide modular capacity scalability,
energy reduction, and improved reliability in submarine cable systems [
5
]. One major
problem in long-haul submarine networks is power feeding, as power can be provided
only from coasts and in general is an expensive resource. SDM is a promising candidate to
efficiently solve the critical power-feeding problem.
The first generation of SDM-based optical submarine systems utilizes space parallelism
(parallel fibers in a cable) and specific high-power voltage (in the range of 18 kV end-to-
end) [
3
] in order to achieve an increased system capacity and power resource optimization.
SDM in principle is a system that incorporates at least one subsystem, being a transmission
fiber, an amplifier, and a switching node, or terminal equipment, which implements the
concept of “spatial integration of network elements” [
6
]. For example, for a Bu-SMF to be
considered as an SDM, it needs to incorporate at least a pump-sharing scheme in the optical
amplifiers/repeaters. Therefore, the first SDM generation is mainly based on novel pump-
sharing technology and a high count of FPs (>8) in the cable [
3
,
4
]. The FPs use optimized
fiber characteristics such as reduced coating fibers (RCF), (200
µ
m) [
4
], that allows for higher
fiber density within the cable. The use of multiple spatial channels and pump-sharing
schemes results in reduction of cost/bit and power/bit, while providing the obvious
modular capacity scaling. In the future, the expected second generation of SDM will use
multiple spatial channels physically placed in MCFs and/or MMFs, constituting Bu-SMFs,
and taking advantage of the formers’ impressive capabilities as already demonstrated in
experimental studies [3,4].
The contribution of this work is twofold. First, the time for SDM in submarine environ-
ments has come; therefore we review in detail all recent developments and announcements
about the forthcoming deployments and discuss the ways that SDM will make inroads
into the submarine networks’ space. More specifically, we focus on the most important
submarine networks’ issues: systems’ capacity, amplification, architecture, power, eco-
nomics, and security. In addition, we analytically explore all SDMs’ special features that
will enhance submarine networks’ transmission capacity and efficiency. Second, we focus
on the pros and cons of different possible alternatives and discover the exact capabilities
of all recently presented SDM-based technologies especially designed for the submarine
environment. We explain why these new capabilities have led to the first SDM deploy-
ment in submarine networks. Finally, based on all these findings and on our own study
findings, we present all possible future directions mainly in the fields of fiber types, fiber
counts per cable, fiber-coating variants, modulation formats, and the type and layout
structure of optical amplifiers that are expected to lead to the second generation of SDM
submarine deployments.
The article is organized as follows: Section 2discusses the basics of submarine optical
networks and reports some of the most important achievements. Section 3overviews
all recently announced submarine cable systems. Section 4addresses the advances in
the field of optical amplification combined with long-haul submarine experiments, and
Section 5refers to possible submarine network architectures. Section 6analyzes topics
regarding the power-feeding equipment (PFE). Section 7discusses the economic benefits
of applying SDM in submarine networks. Section 8covers the most important security
issues in submarine networks. Finally, thoughts on possible guidelines for the future era of
submarine networks are presented in Section 9.

Telecom 2022,3236
2. Submarine Optical Networks Basics and Demonstrations Technologies on
Submarine Cable Systems
2.1. The Past Evolution of Submarine Transmission Systems
The submarine cables’ history (pre-optical era) started in Malaya, when by the late
1840s a newly found natural polymer (“the gutta percha”) provided the basic means
to successfully insulate cables for undersea use. A relatively short-distance cable that
connected the UK (Dover) to France (Calais) was submerged in the summer of 1850 [
1
].
Unfortunately, it lasted only for a few messages and was replaced approximately a year
later by a more durable successor. Shortly after that, the first cable connecting Europe
to the US was deployed and became operational on 16 August 1858 only for telegraphy
purposes [
5
]. Deployment of cables across the Pacific Ocean turned out to be more difficult
and therefore transpacific telegraph cables became operational about half a century later, in
1902, connecting the US to Hawaii [1,5].
Humanity had to wait approximately fifty more years for the deployment of newer
advanced cables to support voice and data communications. TAT-1 (Transatlantic No. 1)
was the first transatlantic cable that was used for telephony [
1
,
5
]. The cable that connected
Scotland to Clare Ville became operational on 25 September 1956, and initially supported
transmission of just 36 telephone channels [
1
]. The first coaxial cables that could transmit
frequency-multiplexed voice signals became operational in the 1960s. Figure 1depicts
some of this history, which led eventually to optical technology (1986, 1988) (optical era) by
using optical fiber links for high-speed/capacity submarine communications [1].
Figure 1. Submarine network elements’ evolution (1956–2022) [1–5].
Some significant (as characterized by the IEEE) milestones in the submarine era history
are [1] as follows.

Telecom 2022,3237
â1884:
The first submarine cable supporting phone data (from San Francisco
to Oakland
).
â1954:
The first submarine (high-voltage direct current) cable connected the island of
Gotland to mainland Sweden.
â1956:
The first deployment of repeaters (in the 1940s) boosted the TAT-1, which was
the first telephone cable crossing Atlantic.
â1964:
The first transpacific submarine coaxial telephone cable linking Japan, Hawaii,
and the US mainland.
â1986:
The first submerged international fiber-optic cable that connected Belgium to
the United Kingdom.
â1988:
The first submerged transoceanic fiber-optic cable, (named TAT-8), that con-
nected the USA to the United Kingdom and France.
Next, in Figure 2we illustrate a transatlantic cable capacity comparison beginning
with TAT-1 in 1956 and ending with TAT-14 in 2001.
Figure 2. Transatlantic cable capacity (circuits).
Figure 3shows the worldwide submarine cable map, including the majority of subma-
rine cables systems which transport approximately 99% of world’s intercontinental traffic
and $10 trillion of transactions daily. It comprises 487 cables, stretching over 1.3 million km
and 1245 landing stations that are currently in service or under construction.
Those networks were deployed by consortia of major telecom network operators.
However, during the current century, Google has become the most active investor in sub-
marine networks, alongside other major cloud service providers, such as Facebook and
Microsoft. From 2016 to 2018, Google invested approximately $47 billion dollars [
3
] in
capital expenditure (capex) to expand and upgrade Google Cloud infrastructure. Nowa-
days, Google counts 134 points of presence (PoPs) and 14 major subsea cable investments
globally [3] interconnecting all six continents, as shown in Figure 4.

Telecom 2022,3238
Figure 3. Submarine worldwide cable map [2].
Figure 4. Google cloud platform submarine network [3].
2.2. Important Milestones at Submarine Systems Evolution
Table 1shows a time-based taxonomy of all important milestones achieved in the
submarine networks’ history. For each milestone, we report the release year, the cable
system, the key technology, the total capacity achieved (and/or the channels deployed),
and finally the span length.
Table 1. Important milestones in submarine networks history [1–5].
Year Cable System Key Technology Capacity/Channels Length
1956–1978 TAT-1 First transatlantic * telephone cable,
electronic repeaters, hotline
36 (initial), 48 final
telephone channels 1942 km
1978–1994 TAT-7 Employs coaxial cable technology
4000 (initial), 10,500
final telephone
channels
8910 km
1988–2002 TAT-8 First *fiber-optic transatlantic cable 295.6 Mbit/s traffic or
40,000 phone circuits 6705 km
1991 TAT-9 First system to switch traffic on
demand between five landing points
560 Mbit/s traffic or
80,000 voice circuits 9305 km

Telecom 2022,3239
Table 1. Cont.
Year Cable System Key Technology Capacity/Channels Length
1996–2008 TAT-12/13 EDFA technology implemented 2×5Gb/s 12,307 km
2001 TAT-14
Employs 4fiber-pair (2FPs active + 2
FPs backup) WDM technology with
direct detection
3.2 Tb/s 15,428 km
2008 TPE Employs 10 G DWDM technology 5.12 Tb/s 17,968 km
2013 PC-1 Employs 100 G coherent technology 8.4 Tb/s 21,000 km
2018 PLCN First submarine cable employs C+L
band optical technology 144 Tb/s 12,971 km
2021 Dunant
First submarine cable employs 12 FPs
introduces first-generation SDM
technology, pump lasers,
pump-sharing technology
250 Tb/s 6600 km
2021 H2HE
The world’s first 16-fiber-pair
repeated submarine cable system. A
milestone in technological innovation
300 Tb/s 675 km
2022 Peace WSS ROADM BU, 200 G technology,
SDM repeater 90 Tb/s 15,000 km
2022 Equiano
First submarine cable employs
optical
switching at the fiber-pair level
(instead of
wavelength-level switching)
200 Tb/s 15,000 km
2023 Confluence-1
First submarine cable employs 24
fiber-pair SDM technology and it
will be the largest to be
recently installed
>500 Tb/s 2571 km
2023 Ellalink
First submarine system to incorporate
state-of-the-art ICE6 800 G coherent
technology
100 Tb/s 6200 km
2023 Firmina
To be the world’s longest undersea
cable capable of maintaining
operations with single-end feed power
by using 18-kV power technology
N.A 2500 km
2023–2024 2Africa
2Africa aims to be one of the largest
submarine cable systems (46 landing
stations, 45,000 km)
180 Tb/s 45,000 km
2025 Arctic. Connect
To be the first transarctic cable
system with new innovative cable
type
and will connect three continents
(85% of total world population)
200 Tb/s 14,000 km
2024–2025 Apricot
To incorporate 400 G technology, all
new submersible ROADM, flexible
bandwidth management based on
SDM-based design.
190 Tb/s 12,000 km
2025 SEA-ME-WE-6
Utilizes SDM cable, supporting up to
24 FPs and incorporates enhanced
branching units (eBUs) providing
flexible electrical power and optical
fiber routing with shore-based
telemetry control
126 Tb/s 19,200 km
* Characterized as IEEE milestone.

Telecom 2022,3240
2.3. Submarine Systems’ Performance Metrics
Submarine cable systems are more challenging compared to terrestrial ones. As a con-
sequence, they have major differences in the ways information about service performance
can be determined and measured.
Some major differences are the following.
â
In submarine networks the service performance can be determined by information de-
scribing the health status of basic network components (BUs, intermediate repeaters).
This information is obtained by coherent transponders which are placed at the ends
of a submarine cable. Their terrestrial counterparts are by far more easy to monitor.
Terrestrial networks can process more data with regard to each unit’s contribution to
the whole system’s performance.
â
Total output power (TOP) constraint is another key difference between terrestrial
and submarine systems as it changes the way that total SNR is calculated. The TOP
constraint in submarine amplifiers results in signal depletion whereas amplifier noise
is accumulated because the total channel power (S+N) remains fixed with distance.
New coherent modems used on D
+
submarine optical cables change the parameters
that define total system capacity and so ITU updated the commissioning process (thus the
final tests before going commercial) in a G.977.1 recommendation [7].
Optical signal-to-noise ratio (OSNR) refers to the ratio of service signal power to noise
power for a valid bandwidth (0.1 nm, so ~12.5 GHz at 1550 nm). OSNR is used to quantify
the linear noise from amplified spontaneous emission (ASE). However, if two systems are
using different baud rates, the OSNR of the higher baud rate system will be reported as
higher (although the two systems may have the same noise level). So, OSNR has to be
defined without reference to channel spacing and symbol rate.
Signal-to-noise ratio ASE (SNR
ASE
) is similar to OSNR except that the noise bandwidth
is equivalent to the signal bandwidth. This leads to a measurement not dependent on
symbol rate and which accounts for all noise detected on the receiver side. Therefore, by
implementing this approach it is easier to compare the SNR metric for signals with different
baud rates.
Similar to the baud independence of SNR, a metric that provides a line rate-independent
measurement of performance would be useful for evaluating potential system capacities.
Effective SNR (ESNR) measures linear and non-linear noises and reports their impact to
the signal performance. As a result, ESNR will not be changed for the same noise levels, no
matter the signal’s line rate. This is an improvement over Q factor, which varies for signals
experiencing the same total noise on different line rates. In this way, ESNR can be used as a
future teller (i.e., if measured at one line rate, it can then be used to predict performance of
higher data transmission rates for the same cable system).
As performance of upcoming systems depends also on optical nonlinearity (SNRNL),
it would be convenient to measure both linear and non-linear performance. Generalized
OSNR (GOSNR) sums the non-linear and the linear noise of the wet plant optical systems.
GOSNR’s updated baud-independent version is GSNR. Other effects are both guided
acoustic–optic wave Brillouin scattering (GAWBS) and signal droop. GAWBS is an effect
which leads to a penalty for a given wet plant design. This effect is caused by the interaction
between light and the acoustic modes that occur in the optical fiber.
GSNR is evaluated directly through simulations or analytic models and indirectly
through experiments. Figure 5summarizes all existing and new metrics and indicates at
which exact point each of them is measured.

Telecom 2022,3241
Figure 5.
Summation of existing and new cable performance metrics and indication of the exact
location where each one of them is measured [7].
2.4. SDM Transmission Technologies in Long Haul Transoceanic Systems
2.4.1. Features of SDM-Based Submarine Cable Systems
The evolution of SDM in submarine networks started with Suboptic’16 (powered by
Alcatel-Lucent Submarine Networks (ASN) Ltd., founded in 1983) which was the first
16-FPs submarine cable system [
8
]. Until recently, in all networks, the initial goal was
twofold: to increase the total cable capacity (up to 70% with regard to traditional cable) and
to decrease the required cost and power per transmitted bit.
The innovative features that characterize the first generation of SDM submarine
networks are [8] as follows.
â
A relatively high count of FPs (in the same cable) in order to increase the t
ransported capacity
.
â
The deployment of lower effective area fibers in order to optimize cost through the
use of a smaller number of regenerators.
â
The implementation of the novel “pump farming” repeaters’ technology. Pump
farming means that a set of pump lasers isused to amplify a set of FPs. Reliability,
redundancy, and better power management are the main advantages. In particular,
reliability can be a cost-reduction factor as submarine cables’ failures and repairs
(bringing downtime in provided services) are very costly.
â
SDM aims to achieve higher capacities by using the same amount of used power
through a more efficient power management. The key concept is to reduce the optical
power provided to each FP as a way to decrease the nonlinearities as implementing
high count of FPs in the same cable.
However, in the event we want to cover small-distance undersea links (e.g., unrepeated-
festoon networks) or if we want to increase the capacity of an existing traditional submarine
cable system (consisting of a limited number of FPs), multiband transmission is a more
effective solution compared with SDM, which is mostly preferable for long-haul distance
links. More specifically, in multiband transmission there is no need to change the existing
wet plant infrastructure during the upgrade process and this can result in both an increase
(by double) of the capacity/FPs and in cost savings.
Table 2presents the pros and cons of using SDM over a single band as opposed to
multiple bands. The options presented are doubling the number of fibers at C band only
and using the same number of FPs over the C + L bands. Note however that the C + L
transmission is less efficient because C + L has to be separated and recombined (mux/de-
mux) in the repeater for each span. This extra multiplexing/de-multiplexing leads to an
extra loss per span of about a few dBs, which is contrary to the “optimizing efficiency”
basic SDM concept.

Telecom 2022,3242
Table 2. Compare SDM with other non-SDM solutions [8].
PROS CONS
Lower capacity per FP, as FP
becomes consequently the
new granularity
Higher quantity of FP used
C-Band FP switching used to drop a whole
FP in a branch
Bigger cable needed to contain all
the FPs
Easier to sell FP -
Easier to swap FP -
-Less efficient as MUX/DMUX
should be used (~1 dB losses)
-Attenuation is slightly higher in
L band (beyond 1600 nm)
(C + L) Band
Limits the number of FP in the cable
Interband effects between C and
L bands
No need to develop big cable to
contain all the FPs
Spectrum sharing needed to
sell/swap a portion of a FP
-Higher cost for BU with WSS to
manage spectrum sharing
- L band amplifier needed
Figure 6shows the different types of submarine cables and the various types of fibers,
respectively. The selection of the optimal cable type depends on the depth at which each
cable is sinked. For example, double-armored (DA) submarine cable is used at the shore
end, terminated at the beach manhole of the cable landing site, and interconnected with a
much lighter land cable (LWA) moving toward the cable landing station [3].
Figure 6. Different types of submarine cables.

Telecom 2022,3243
2.4.2. Multiple Spatial Channels in SDM: MCF (Multi-Core Fibers)-MMF (Multi-Mode
Fibers)-Bundles of Single-Mode Fibers (Bu-SMFs)
Regarding the number of optical spatial channels in a cable system, three options are
considered for submarine SDM deployments:
â
Multiplying the number of conventional fibers (thus implementing a parallelism that
consists of single-core/single-mode fibers), considering the existence of at least one
element that performs spatial integration, e.g., an amplifier with sharing pumps, a
switching node, or terminal equipment.
âMultiplying the number of cores in MCF fibers.
âMultiplying the number of modes in MMF fibers.
A novel fiber option for adoption in the subsea environment are MMFs, which have
more friendly production compared to MCFs but are expected to induce higher attenuation
compared to their SMF counterparts. MMFs can be used in SDM terrestrial networks,
supporting 3, 10, 15, or even more guided modes [
9
]. MMFs’ major disadvantages are
modal dispersion, interference, and high values of differential mode group delay (DMGD)
which leads to poorer physical layer performance compared with SMFs. A major challenge
for their use in transoceanic networks is to solve their mode-coupling complications. On
the other hand, MCFs, which consist of multiple cores in one cladding, can be used in SDM
terrestrial networks, thereby achieving transmission speeds at the order of petabits/second.
MCFs suffer from extra loss from the use of fan in–out devices and possibly higher typical
attenuation (compared to SMFs). MCFs and MMFs were not mature enough in 2016/2017
when the SDM submarine solution was first deployed [8].
Researchers expect that MCFs could be mature in the medium term. So, the first SDM-
based system proposed by ASN uses multiple conventional fibers (bundles of parallel SMFs)
which, as already mentioned, is considered to be the first generation of SDM submarine
technology. Although SDM deployment in terrestrial systems is strongly favored with
MMF and MCF systems, in the submarine era the requisite stronger, waterproof coatings
limit the number of fibers that can be packed in a single cable.
Figure 7pictorially describes various types of fibers. Bu-SMFs incorporate a number of
SMFs within a single cable. Furthermore, MMFs can support tens of guided modes whereas
FMFs are manufactured to consider propagation of less guided modes. Finally, coupled
core (CC) fibers consider strong mode coupling between the different cores, allowing them
to attain a shorter core-to-core distance and higher spatial density compared with the
uncoupled MCFs.
Figure 7. Different types of fibers.

Telecom 2022,3244
Recently, thinner coatings have been tested in an effort to leave more space for fibers.
Most commercial submarine cable systems up to now use bundles of SMFs which are a
reliable and tested platform. In the near future, we expect that SMFs can operate with a
less than 200-
µ
m coating diameter [
10
], which technically can be achieved either with a
thicker cladding diameter (<125
µ
m) and the same coating thickness or by using 125-
µ
m
cladding diameter and reduced coating thickness in ranges less than 200
µ
m. For this, the
mechanical reliability of new fiber design has to be guaranteed. In both cases, additional
tests for bend performance have to be conducted.
2.5. Basic Segments (“Plants”) of a Submarine System
Submarine networks consist of the dry plant (the coastal supporting equipment) and
the wet plant (the submerged part).
Figure 8shows the basic components of both plants, focusing on SDM technology.
Figure 8.
A typical submarine cable system takes into account all network elements and state-of-the-
art SDM repeaters (analytically described below).
The dry part includes the transponders (which transmit and receive the optical data
carried by the submerged FPs), the PFE, which as its name suggests supplies power (in the
range of 15–20 kV) to the wet plant, and the monitor and control equipment which perform
network management tasks.
The wet plant includes the cable (containing the actual optical data carriers, the FPs),
the conductor (which manages the high-voltage supplied by the PFE), and the neces-
sary repeaters (installed every 75–100 km), which provide the necessary amplification for
optical signals.
Figure 9shows all types of a wet plant submarine cable system, with time and
technology-based classification. The upper left part of Figure 9depicts the dispersion-
managed cable used until 2010, and the lower left part depicts uncompensated schemes
launched with positive dispersion after 2014, which resulted in SDM technology technique
cable systems that are used up to the present and seen at lower right. The festoon and
unrepeated cable systems (upper right part of Figure 9) form an alternative form of the dry
plant which is used only in short-distance submarine cable systems. Festoon networks con-
nect coastline locations by using unrepeated (passive) submarine cables in order to avoid
terrestrial routes for many reasons, such as the presence of challenging terrain (mountains),
dense cities, etc.

Telecom 2022,3245
Figure 9. Types of submarine cables (e.g., SDM, festoon, uncompensated).
2.6. Cable-Installing Ships
An important part of submarine networks is the cable ship whose work is to install,
bury, repair, and service cables. Figure 10 illustrates a typical cable ship profile and describes
in detail all the basic elements and the use of each element [
11
]. Specially designed ships
are used for sinking the wet plant cables and devices.
Figure 10. A typical cable ship profile.
As cable feeding from one ship to another is not an easy job to perform in the middle of
an ocean, most ships are able to carry enough cable to cross the Atlantic, thus approximately
6500 km. The whole long cable is stored in cable tanks and sinked into the ocean by using
specialized equipment. As cable damage most often occurs near shores due to the presence
of other ships, there are existing options to bury the cable if needed. Moreover, modern
GPSs can ensure the exact position of sinking, considering possible underwater current
and winds in an effort to keep up with the pre-selected, shortest possible path route.
3. Recently Announced Submarine Cable Systems
3.1. A Detailed Overview of SDM-Based Technology Cable Systems
In this subsection, we refer to all known submarine cable systems [
1
–
5
]. Table 3
summarizes their main characteristics: capacity, fiber pairs (FPs), length, and year of release.
As SDM seems to be the most promising multiplexing technology for the high-speed
bandwidth submarine era, we focus on SDM-based cable systems. As expected, a common
characteristic of these systems is the use of an increased number of parallel FPs inside the
submarine cable.

Telecom 2022,3246
Table 3.
A taxonomy of SDM-based submarine cable systems [
1
–
5
]. RFS: Ready for Service year,
SDM 1: First Generation of SDM-based system introduced by (ASN/HMN) supplier.
Submarine Cable
System RFS Cable Length
(Km) Capacity (Tb/s) Technology Fiber Pairs
(FPs)
SDM
Nature/Info
Dunant 2021 6400 250 SDM 1ASN 12 Pump Sharing Repeater/(12 FPs)
Malbec 2021 2600 108 SDM 8 Fiber Count ≥* 8FPs (8 FPs)
Hainan to Hong
Kong Express
(H2HE)
2021 675 307 SDM 1HMN 16 High Fiber Count (16 FPs)
Peace 2022 15,000 90 SDM 1ASN †16 SDM Repeater/(16 FPs)
Equiano 2022 12,000 200 SDM 12 Fiber Count ≥* 8FPs (12 FPs)
Grace Hopper 2022 7191 352 SDM 16 High Fiber Count (16 FPs)
Amitie 2022 6792 320 SDM 16 High Fiber Count (16 FPs)
2Africa 2023 45,000 180 SDM 1ASN 16 Fiber Count ≥* 8FPs/(12FPs)
ECHO 2023 17,184 144 SDM 12 Fiber Count ≥* 8FPs (12 FPs)
IAX 2023 5791 200 SDM †12 Fiber Count ≥* 8FPs (12 FPs)
Confluence-1 2023 2571 ≥500 †SDM 24 Ultrahigh Fiber Count (24 FPs)
Firmina 2023 2500 N.A SDM 12 Fiber Count ≥* 8FPs (12 FPs)
Bifrost 2024 15,000 180 SDM 1ASN 12 Fiber Count ≥* 8FPs (12 FPs)
Apricot 2024 12,000 190 SDM †16–20 Fiber Count ≥(12 FPs), †
(16–20 FPs)
IEX 2024 9775 200 SDM †16 High Fiber Count (16 FPs)
Medusa 2024 8760 480 SDM 24 Ultrahigh Fiber Count (24 FPs)
Blue Raman 2024 7500 400 †SDM 16 High Fiber Count (16 FPs)
Caribbean Express
(CX) 2024 3472 280 SDM 18 Very High Fiber Count (18 FPs)
Hawaiki Nui 2025 22,000 240 SDM 12 Fiber Count ≥* 8FPs (12 FPs)
Sea-We-Me 6 2025 19,200 126 SDM 10 Fiber Count ≥* 8FPs (10 FPs)
* 8FPs is the limit of previous technologies cable systems, †expected/predicted.
The Dunant submarine cable system connects Virginia Beach (US) to France (French
Atlantic coast), via a 6600-km cable. It is the first submarine system over Atlantic Ocean
that employs a 12 FPs SDM-based design, featuring a capacity of 25 Tb/s/FP and thus a
total of 300 Tb/s. Compared to previous subsea cable technologies, there are based on a
dedicated set of pump lasers which are used to amplify each fiber pair, the Dunant cable
system implements a common sharing of pump lasers and associated optical components,
among multiple fiber pairs. This system is built by Google. Google and Orange are the
managers of the landing sites in US and France respectively. Dunant has been in use since
19 January 2021.
Equiano is submerged along the west coast of Africa, connecting Portugal and South
Africa, and is expected to serve several countries along its route. It is the first subsea cable
system performing optical switching at the fiber-pair level (instead of wavelength-level
switching). Alcatel Submarine Networks is building the cable, which will travel for more
than 12,000 km and deliver a capacity of 200 Tb/s to its users with approximately 20 times
more network capacity than its predecessor. Equiano is expected to be ready for service in
late 2022.
Peace is a 25,000-km-long (15,000 km at the begging and 25,000-km with the exten-
sions), privately owned cable system which uses the shortest direct route from China to
Africa and Europe with main landing points in France, Egypt, Kenya, Pakistan, Singa-
pore, Maldives and (recent update March 2022) indicates that the system will extend to
Seychelles. The submarine system will deploy 200-G WDM technology interfacing with
HMN’s SDM repeater, which provides capabilities to transmit up to 25 Tb/s/FP, highly
optimized algorithms, and latest-generation Nyquist subcarriers. PEACE Mediterranean
segment, with landing points in Marseille, Cyprus, Malta and Egypt, has already been put
into commercial use since early 2022. The Peace extension to Singaporeis expected to be
ready for service in late 2023.
The Hainan to Hong Kong Express (H2HE) is a state-of-the-art SDM technology cable
system, consisting of a 675-km-long cable, with landing points at Hong Kong SAR and
Hainan Province, Mainland China, with a branch to Guangdong Province. The H2HE cable
system is the first 16 fiber-pair repeated submarine cable system in the world, achieving

Telecom 2022,3247
19.2 Tb/s per fiber and delivering a capacity of 307.2 Tb/s to its users. The H2HE cable
system is invested and owned by China Mobile, supplied by HMN Tech, and has been in
service since September 2021.
The Malbec cable system is a new investment owned by Facebook and Globenet. It
consists of a cable 2600 km in length and connects the cities of Rio de Janeiro, Sao Paulo, and
Buenos Aires, with a future connection to Porto Alegre. It uses the latest SDM technology
with low latency and 400 Gb/s optical channels. The initial available capacity of the Malbec
system is 108 Tb/s, approximately 18
+
Tb/s/FP, and has been in service since 10 June 2021.
The Grace Hopper cable system (another Google investment like Curie, Dunant, and
Equiano) connects the US with the UK and Spain. It travels for approximately 6250 km
from New York to Bude (UK), and another 6300 km from New York to Bilbao (Spain). It
is a private subsea cable and the first-ever route to Spain. It incorporates 16 FPs and it
transports 22 Tb/s/FP, thus delivering a total of 352 Tb/s, which is higher than the capacity
of current internet infrastructure from the US to Europe. SumComand and Telxius are the
supplier and Spanish landing partner, respectively. Grace Hopper is expected to be ready
for service in 2022.
The Amitie submarine cable system will interconnect Massachusetts (US), Le Porge
(France), and Bude (UK). The signal will travel for 6600 km through the Atlantic Ocean.
Facebook, Microsoft, Aqua Comms, and Vodafone (through Cable & Wireless Americas
Systems, Inc.) are the consortium partners. It will give over 320 Tb/s of capacity divided in
three segments by using 12 and 16 FPs. ASN will provide the system, which is expected to
go commercial in the first quarter of 2022.
Confluence-1 will connect Miami and New York. It will travel for 2571 km and will
use 24 FPs (ultra-high fiber count) achieving a 50% improvement in fiber count compared
to the previous 16 FPs cable systems. Confluence-1 highlights another possible use of
the submarine networks, which is to provide a more direct and/or a backup alternative
to terrestrial ones. Note however that this possible use can be applied only along long
coastlines (like the Miami–New York coastline) or in closed seas (like the Aegean Sea in
Greece or the Mediterranean Sea). Confluence-1 hopes to be ready for service in 2023.
The Echo submarine system will connect Eureka (USA), Guam, Indonesia, Singapore,
and other countries. Echo will be the first subsea direct connection from the US to Singapore.
It will travel 16,200 km from Singapore to Eureka and will use 12 FPs. Providing 12 Tb/s/FP,
Echo aims at a total system capacity of 144 Tb/s. Google and Facebook are the joint builders.
Indonesian Telco, Google Singapore PTE Ltd., and Facebook Edge USA deal with the dry
plant in Indonesia, Singapore, and the US (& Guam), respectively. NEC will supply the
system, which is expected to be ready in 2023.
The Firmina cable system is a new investment of Google. It will travel for more than
3500 km and will have four landing points in Las Toninas (Argentina), Praia Grande (Brazil),
Punta del Este (Uruguay), and the East Coast of the United States. It will consist of 12 FPs,
and it will be the world’s longest undersea cable, capable of maintaining operations with
single-end feed power by using 18-kV power technology. The cable system is expected to
be ready for service by the end of 2023.
Topaz cable system is the first-ever submarine optic-cable to connect Canada and Asia
invested by Google. It will travel from Vancouver to the small town of Port Alberni on the
west coast of Vancouver Island in British Columbia, and across the Pacific Ocean to the
prefectures of Mie and Ibaraki in Japan. It will use state-of-the-art Wavelength Selective
Switches (WSSs) to provide flexible routing and resilience. It will consist of 16 FPs giving a
total capacity of 240 Tb/s and is expected to be ready for service in 2023.
The India-Asia-Xpress (IAX) and India-Europe-Xpress (IEX) cable system will connect
Mumbai and Chennai to Singapore and to Europe, respectively. IAX will also connect
several Far East countries to the US West Coast, whereas IEX will connect to the East Coast
of the US. The IAX and IEX cable systems will adopt open cable system technology (which
means that it can be connected to any type of technology or brand) and the innovative
wavelength-switched ROADM/branching units, providing rapid upgrade deployment

Telecom 2022,3248
and flexibility to add/drop wavelengths across multiple stations. Both the IAX and IEX
cable systems are built by Reliance Jio Infocomm and are expected to be ready for service
in mid-2023 and early 2024, respectively.
The Sea-Me-We 6 (SMW6) cable system will be the latest project of the Southeast
Asia-Middle East-West Europe (SEA-ME-WE) series and it will connect Singapore and
Marseille via Egypt. It will travel 19,200 km, and it will consist of up to 10 fiber pairs,
with 12.6 Tb/s/FP and a total system capacity of 126 Tb/s. It will adopt the latest SDM
technology. The SMW6 consortium comprises Bharti Airtel, Bangladesh Submarine Cable
Company Limited (BSCCL), China Telecom, China Mobile, China Unicom, Djibouti Tele-
com, Orange, Singtel, Sri Lanka Telecom, Telecom Italia, and TSA. It is expected to be ready
for commercial usein early 2025.
2Africa (initially announced in May 2020) aims to connect 46 cable landing stations in
33 countries in Africa, Asia, and Europe, and will travel for 45,000 km (Figure 11). Recent
updates (August 2021, September 2021) indicate that the system will include the Seychelles,
the Comoros Islands, and Angola, a new landing point, to southeast Nigeria and extend to
the Persian Gulf, Pakistan, and India. With its recent extension, 2Africa is now the longest
subsea cable in the world. 2Africa will use SDM powered by ASN with up to 16 FPs and
will incorporate optical switching for flexible bandwidth management. It is also the first
system of its size that deploys an innovative aluminum conductor specially constructed for
submarine cable systems. The project will use WSS-ROADMs (used for the first time in
Africa) and is expected to go commercial in 2023–2024.
Figure 11. 2Africa submarine cable system [3].
The Caribbean Express (CX) will connect Florida with Balboa (Panama) and will branch
to Mexico and Colombia. Future extensions include Cuba, Grand Cayman, Guatemala,
Jamaica, Honduras, Nicaragua, and Costa Rica. Ocean networks will provide the SDM-
based system, which will deploy 18 FPs, set at a minimum of 18 Tb/s/FP. It will be the first
and only cable system in the Caribbean area that will offer full fiber pairs to customers. CX
is expected to be ready for service in the first quarter of 2024.
The Bifrost cable system will connect Singapore, Indonesia, the Philippines, Guam,
and the US West Coast. It will travel for more than 15,000 km. When fully deployed, it will
be the largest-capacity high-speed transmission cable across the Pacific Ocean. Facebook,
PT. Telekomunikasi Indonesia International (Telin), and Keppel Telecommunications &
Transportation Limited (Keppel T&T) comprise the consortium. ASN will supply the
system, which is planned to use SDM technology with 12 FPs. Bifrost is expected to go
commercial in 2024.
The Apricot submarine cable system will connect Japan, Taiwan, Guam, the Philip-
pines, Indonesia, and Singapore, travelling approximately 12,000 km. It will use SDM,

Telecom 2022,3249
including 400-G transmission, multiple pairs of high-capacity optical fibers (16–20 FPs
expected), and will incorporate a state-of-the-art submersible ROADM employing WSSs.
The consortium comprises Facebook, Google, NTT, Chunghwa Telecom (CHT), and PLDT.
NTT will operate the system and manage the three cable landing stations in Japan, Tuas,
Singapore, and Indonesia. Apricot will be ready for service in late 2024.
The Medusa submarine cable system will be a state-of-the-art fiber-optic system
providing open cable features. It consists of 24 FPs, with 20 Tb/s/FP and a total system
capacity of 480 Tb/s. The Medusa cable system will interconnect Southern European
countries—Portugal, Spain, France, Italy and Greece—with North African countries—
Morocco, Algeria, Tunisia and Egypt. It will travel more than 8760 km. The Medusa cable
project is expected to cost
€
326 million (US $374 million), and be partially financed by the
European Investment Bank. Medusa will be ready for service in third quarter of 2024.
The Hawaiki Nui submarine cable system will interconnect Southeast Asia, Australia,
and North America. PT Mora Telematika Indonesia (Moratelindo) is the Indonesian
strategic partner of the project. It consists of 12 FPs and has a total system capacity of
240 Tb/s. The Nui cable project is expected to be ready for service in 2025.
Table 3presents a taxonomy of SDM-based submarine cable systems; each of them
deploys parallel fibers and/or novel pump-farming technology [1–5].
From an industrial point of view, some of the biggest submarine services providers,
such as Alcatel Submarine Networks and HMN point out the main characteristics of an
SDM-based cable system:
The SDM
1
system powered by ASN [
8
] is trying to revolutionize the well-known
submarine designs by:
âpushing the limits of theoretical design capacity;
âminimizing nonlinear effects to reduce needed equipment, cost, and complexity;
â
designing an efficient optical and electrical network based on repeater pump farming,
low Aeff submarine fibers, and higher fiber counts; and
â
working in the optimum spectral efficiency of submarine line terminal equipment
(SLTE): 2–3 b/s/Hz and lower chromatic dispersion compensation.
The SDM system powered by HMN [
12
] Tech involves the high fiber count solution
and includes:
â
a high fiber count submarine repeater which broke through the fiber count limitation of
existing products and can support up to 16 fiber pairs which can double
the capacity
;
â
an industry-leading 39.5 nm ultrawide bandwidth, which covers C-band and extended
C-band, to maximize the capacity of one fiber pair; and
âa significantly reduced cost/bit.
Table 4presents a comparison between traditional and SDM submarine cables and
summarizes the main features of each cable’s technology.
Table 4. A comparison between Traditional and SDM submarine cables [13].
Subsea Component SDM Cable Traditional Cable
Submarine Cable High Count of FPs
(12, 16 FPs and more in future)
Limited number of FPs
(6 FPs and maximum 8 FPs)
Fiber Effective Area (Aeff) Low effective area,
Aeff 110–80 µm2, att. 0.155 dB/km
High effective area,
Aeff 150–125 µm2, att. 0.15 dB/km
Repeater Low power repeaters: (+14 to 20 dBm) Very high power repeaters: (>+20 dBm)
Repeater Type Repeater pump farming Each fiber has own laser pumps
Branching Unit ROADMs Fiber pair switching in (BUs) No fiber pair switching in (BUs)
OSNR Low OSNR High OSNR
Modulation Formats PCS (probabilistic constellation shaping) BPSK, QPSK, 8-QAM and 16-QAM
C + L Band Technology Currently only C-Band
C + L Band supported up to 144 channels
per FP
PFE Same PFE, capacity can be increased Same PFE

Telecom 2022,3250
Figure 12 depicts the new FP granularity of SDM (16 FPs) vs. traditional cables
(
<8 FPs
). In this conceptualization, the potential capacity per FP is decreased. For example,
in the past,
200 Tb/s
could be transmitted by using 8 FPs (25 Tb/s/FP). Now the same
200 Tb/s
is achieved by using 16 FPs (12.5 Tb/s/FP), thus bringing technological and
economic benefits (easier FP-switching, selling a whole FP to a customer instead of portions
of it).
Figure 12. Traditional vs. SDM cable granularity.
Table 5summarizes the main characteristics of current submarine cable systems: type
of topology, cable length, number of landing points, and number of operators of the major
cable systems around the world. Note the large number of operators that are involved
in many systems. This is expected, as these cables interconnect a number of countries;
as a consequence, a number of operators exploit these cables to transmit data from one
continent/country to another.
3.2. Attainable Capacity of Submarine Cable Systems
As already pointed out, all technologies (including SDM) target an increase in the
transmission bandwidth. If we use both C + L bands for transmission (instead of only using
the C-band), the attainable transported capacity is higher by approximately a factor of 2. In
other words, we can achieve higher system capacity either by doubling the number of FPs
(with SDM) or by using the C + L bands and not changing the number of FPs. However,
C+L
transmission (e.g., PLCN cable system in Figure 13) is less efficient compared with
the doubling of FPs because C + L bands must be separated and recombined (mux/de-mux)
in the repeater for each span.

Telecom 2022,3251
Table 5. A taxonomy of existing submarine cable systems based on their topology [1–5].
Topology Submarine Cable
System
Cable Length
(Km)
Landing
Points
Number of
Operators
SeaMeWe-3 39,000 39 52
FLAG Europe-Asia (FEA) 28,000 17 Global Cloud
Xchange
AsiaAfrica Europe-1 (AAE-1) 25,000 20 18
SeaMeWe-4 20,000 16 16
SeaMeWe-5 20,000 18 18
Africa Coast to Europe (ACE) 17,000 22 20
** SDM Repeater Peace 15,000 14
Peace Cable
International Network
Co., Ltd.
EuropeIndia Gateway (EIG) 15,000 12 16
Southern Cross NEXT 13,700 7 Southern Cross Cable
Polar Express 12,650 10 Russian Government
BRUSA 11,000 4 Telxius
Africa-1 10,000 10 Etisalat UAE
South Atlantic Express (SAEx) 10,000 5 SAEx International
Trunk&Branch Amitie 6792 5 3
TE North/TGN-Eurasia/ . . . 3634 4 6
Malbec 2600 3 Facebook, GlobeNet
* SDM Technology 2Africa 45,000 29 8
* SDM Technology Hawaiki Nui 22,000 13 Hawaiki Submarine
Cable
* SDM Technology ECHO 17,184 6 2
* SDM Technology Bifrost 15,000 5 3
* SDM Technology Equiano 12,000 7 Google
* SDM Technology Apricot 12,000 4 7
* SDM Technology Medusa 8760 16 AFRIX Telecom
* SDM Technology Grace Hopper 7191 3 Google
* SDM Technology Amitie 6792 3 5
* SDM Technology Caribbean Express (CX) 3472 11 OceanNetworks (ONI)
* SDM Technology Confluence-1 2571 5 Confluence Networks
EAC-C2C 36,500 16 Telstra
Mesh Trans-Pacific Express (TPE) 17,000 6 7
MedNautilus Submarine System 7000 7 Telecom Italia Sparkle
Apollo 13,000 2 Vodafone
CAP-1 12,000 2 Amazon Web Services,
Facebook
Seabras-1 10,800 2 Seaborn Networks,
Telecom Italia Sparkle
Point-to-Point MAREA 6605 2 Facebook, Microsoft,
Telxius
INDIGO-Central 4850 2 5
JGA-N 2600 2 RTI
BlueMed 1000 2 Telecom Italia Sparkle
* SDM Technology Dunant 6400 2 Google
Japan-U.S. Cable Network (JUS) 22,682 6 24
Pacific Crossing-1 (PC-1) 22,900 4 NTT
Ring TPC-5 22,560 6 13
Atlantic Crossing-1 (AC-1) 14,301 4 Lumen
* The cable system incorporates the SDM technology; ** The cable system incorporates at least one SDM-based
device (e.g., SDM repeater).

Telecom 2022,3252
Figure 13.
Evolution of submarine cable systems’ capacity through time (pre-SDM era vs. first-
generation SDM era).
Figure 13 presents the evolution of the submarine systems’ capacity through time.
Blue and red dots represent non-SDM and SDM cable systems, respectively. Orange dots
represent multiband (C + L) cable systems. SDM systems are divided into two main
categories: the SDM systems using a high-count of parallel FPs in the same cable achieving
a reduced cost/bit (first generation of submarine SDM by ASN), and SDM systems by
using the novel pump farming technique at the repeater’s side. SDM cable systems of the
first category are furthermore subdivided based on the number of FPs.
Cable systems using both techniques (high count of FPs and pump farming) are
represented by colorful dots. Figure 13 shows that SDM-based cable systems can achieve
higher system capacities. As expected, a higher number of FPs achieve a higher Tb/s of
total capacity.
Because the most crucial factor is often not the maximum capacity but the distance over
which this maximum capacity is achieved, Figure 14 presents the evolution of the product
of the submarine cable systems’ capacity multiplied by the associated length through time.
Again, blue and red dots represent non-SDM and SDM cable systems, respectively. Orange
dots represent multiband (C + L) cable systems. SDM cable systems of the first category are
furthermore subdivided based on the number of FPs. We observe that recently announced
SDM cable systems can achieve significantly higher values of the capacity
×
length product
over non-SDM systems.

Telecom 2022,3253
Figure 14.
Submarine cable system capacity
×
cable length through time for 70 different cable systems.
3.3. Experimental Demonstrations That Show a Glimpse to Possible Future Evolution
Many important submarine cable experiments have been performed so far. In [
14
],
Nakamura, K. et al. demonstrated a transmission of 200-G carriers with 60
+
Gbaud each
over 17,680 km in distance. Such a cable aims to operate in the linear regime of an SDM-
optimized span length. The spectral efficiency (SE) was in the order of 2.91 b/s/Hz resulting
at a rate of 2.67 b/s/Hz with adequate safety margins. This demonstration seems to win
the gold medal for the longest baud-rate (>60 Gbaud) channel transmission.
An 8.12 Tb/s long-haul transmission over a transoceanic distance of 9750 km using
100 Gb/s 8D coded modulation, minimizing nonlinear transmission penalty, and resulting
in a high-power efficiency of 54.8 Pb/s ×km/W was demonstrated in [15].
Also, we have another gold medal for reaching a capacity of 24 Tb/s over 6644 km
transatlantic cable system. The system used 16QAM synthesized subcarriers, forward error
correction (FEC) gain sharing and multi-carrier wave locking [16].
The work in [
17
] surveyed the evolution and the future of submarine technology. The
authors concluded that the maximum number of fiber pairs (FPs) that can be efficiently
positioned in a submarine cable is approximately 30. As current WDM systems may face
the capacity crunch soon, it is rather obvious that new multiplexing technologies must
be adopted. Enriching the previous findings, authors in [
18
] reported a 38% capacity
increment by using 24 FPs placed in a 4-core MCF.
The work in [
19
] spotted the design issues and construction processes of submarine
cable systems. Authors discussed all steps for construction of a submarine optical network:
proper route selection, branching units’ selection, topology design, type of cables, defining
undersea housing and proper spacing for the repeaters selecting the proper PFE’ architec-
ture, organizing ship operations, and finally selection of the proper submarine terminal
equipment (SLTE).
The latest advances include specialized WSSs for submarine applications (recently
announced for commercial use). These WSSs can distinguish every wavelength of each
incoming fiber and then switch it to the appropriate output port of add–drop multiplexer
(OADM) subchannels at a granularity of 3.125 GHz [
20
]. The use of WSSs in terrestrial
ROADMs [
21
] and submarine OADMs [
22
] solves several design issues. Other basic com-
ponents of submarine OADMs are the branch unit (BU) and the wavelength management
unit [13].
Optical amplifiers are of extreme importance in the submarine environment. Topics
that are under investigation are the efficiency of cladding pumping EDFAs and the ability

Telecom 2022,3254
of PFEs to provide power to the submerged equipment over ultra-long distances (in the
order of thousands of kilometers). Amplification efficiency of cladding pumping depends
on core density (thus the ratio of the number of cores by the cladding area) whereas core
pumping will improve by using the pump light to support multiple optical fiber cores.
All recently announced pumping schemes for SDM submarine networks were described
in [13,23].
A new and promising technique is the pump recycling [
24
], explained below (in
Section 4.7). A detailed view of amplifiers technology follows in Section 4.
Benchmarking of submarine architectures are of a great interest during the design
phase of a submarine network. The layout of submarine architectures and topologies can
be designed considering three categories: point-to-point, trunk and branch, and connecting
ring [
13
]. The basic design decision, about which option to adopt, is the distance between
land stations. The use of BUs may be rather limited if the distances between candidate land
stations are relatively short (i.e., closed seas, such as the Mediterranean Sea). In such cases,
as noted in [
25
], we can have no presence of switching elements under water as optical
paths can be switched or/and reconfigured from the land-based ROADMs. Submarine
architectures incorporating BUs and MCFs, routing cases, and others issues of the BUs
deployment were presented in [
25
]. Submarine architectures are discussed in detail in
Section 5.
Finally, Table 6makes a comparison of transmission experiments at transoceanic
distances where C is the capacity in Tb/s, L is the transmission distance in km, SE is the
spectral efficiency in (b/s/Hz), and finally the relative technology, on which they are based.
From the values of SE, we can conclude that SDM systems are working in the optimum
spectral efficiency (~2–3 b/s/Hz) compared to non-SDM systems that are working at
higher levels (up to 7.36 b/s/Hz). Finally, it is worth mentioning that [
16
] reviewed the
deployment of the MAREA submarine cable system.
Table 6. Comparison of transmission experiments at transoceanic distances.
Reference C (Tb/s) L (Km) ↓SE (b/s/Hz) Technology
[26] 71.65 6970 7.36 non-SDM
[26] 70.46 7600 7.23 non-SDM
[16] * 24.6 6664 6.21 non-SDM
[27] 30.58 6630 6.10 non-SDM
[28] 105.1 14,350 3.20 SDM
[15] 8.12 9750 3.20 SDM
[14] 12 17,680 2.91 SDM
[29] 9.0 15,050 2.00 SDM
* Review the deployment of MAREA cable system.
4. Submarine Amplifiers
4.1. Overview and History Evolution of Submarine Amplifiers
Signal amplification is one of the most important key parameters in telecommunication
networks. In submarine networks, power stations and power amount are limited. Distances
are in the order of thousands of km and so amplification plays the most crucial role in
signal transmission. There has been a lot of progress on submarine amplifiers’ evolution
from the first generation (TAT-1 electric repeater which used only electric parts, most of
them presented and pictorially described in Figures 15 and 16), through the fifth generation
32 FP petabit-level optical amplifier (repeater, Figure 17), which is almost ready to release
to service [30]. The submarine amplifiers evolution history starts with the first generation
of submerged repeater as a main part of the cable TAT-1, the first transatlantic telephone
cable (already characterized as an IEEE milestone).

Telecom 2022,3255
Figure 15.
All element parts of the first-generation sectioned submerged repeater for TAT-1 the first
transatlantic telephone cable [1].
Figure 16.
First-generation sectioned submerged repeater for TAT-1 the first transatlantic telephone
cable [1].
Figure 17. Fifth-generation remote optical pumping amplifier [12].
Nowadays, for undersea operations there are two main types of submarine amplifiers.
The first type is named the remote optical pump amplifier (ROPA) and is presented in
Figure 17, and it is applied in unrepeated systems.
ROPA is capable of amplifying the optical signal efficiently, increasing both transmis-
sion distance and cable system capacity and also compensating the optical signal. This
type of amplifier can support signal amplification for different bands (C-Band, etc.) with
up to 32 fiber pairs, which is the commercial systems’ limit up until the present. Finally, it
features a pressure-resistant, anti-corrosion, high-strength housing, and it is able to support
underwater operations under big depths.
A second type of fifth-generation submarine amplifier is presented in Figure 18. It is
capable of supporting up to a 32 FPs system (not commercially existing at present), and
was first applied in a 16 FPs system presented in Figure 19, aiming to satisfy the latest huge
transmission demands.
Figure 18.
Fifth-generation HMN 1660-32 FP petabit-level optical amplifier (repeater) prototype with
a commercial release expected in 2022 [30].
This type of amplifier embedded new innovative features such as a new electronic
design and an innovative pump sharing scheme aiming to reduce voltage drop by 30%.

Telecom 2022,3256
Pump sharing is a key concept for signal transmission over long transoceanic distances.
It has the capability of providing 39.5 nm bandwidth, and so can enhance the maximum
achievable capacity of existing systems. It can support up to a 160-km transmission of
each segment in repeated submarine cable systems and up to a 16,000-km transmission
in transcontinental systems [
30
]. Finally, this type of amplifier has an innovative design,
ensuring stable operation at high water depths with an approximately 25-year lifetime.
Figure 19.
The Hainan to Hong Kong Express (H2HE) cable system will be a milestone in tech-
nological innovation as the world’s first 16 fiber pair repeated system, incorporating the HMN
fifth-generation space-division multiplexing (SDM) technology [3,12].
4.2. Differences between Terrestrials and Submarine Amplifiers
The wet plant (compared to the dry plant) has higher demands and standards regard-
ing the deployed undersea network infrastructure. This is partly due to the high cost of
maintenance and failover for undersea components. Consequently, terrestrial and subma-
rine network amplifiers are rather different. The key concept for a submarine amplifier
is the high reliability. This is one of the most important differences between its terrestrial
counterpart, as a point failure in a submarine amplifier means both long downtime in
provided services and expensive repairs.
A second difference is the amplifier gain. As initial and maintenance costs are high, the
construction of submarine amplifiers should be efficiently designed. Submarine amplifiers
should provide spectral and power-efficient amplification and at the same time keep the
costs down. In the upcoming SDM era, features such as the repeater pumping schemes and
the use of high count of FPs have to be implemented.
Submarine amplifiers must also have low NF and high OSNR, due to the submarine
systems’ longest links and much more expensive infrastructure. In the next few years, we
will have to face the capacity crunch, therefore, the capacity requirements for submarine
transmission have to overcome the power efficiency limits which a submarine amplifier
repeater can support at the present time.
4.3. Multiband Amplification Technology
Multiband (C + L) amplification schemes (presented in Figure 20) is a complementary
way to achieve the SDM plans. More specifically C + L repeaters can increase the cable
capacity for systems with a limited number of FPs. Furthermore, a multiband C + L
design solution (compared to C-band only) seems a cost-effective solution, as the wet plant
cost remains the same (compared to C-Band only). Significant progress and research on
multiband technology has been made at present. In [
26
] authors presented two 70.46 Tb/s
and 71.65 Tb/s transmissions over 7600 km and 6970 km respectively with a C + L band

Telecom 2022,3257
EDFA by using a multidimensional coded modulation format with hybrid probabilistic
and geometric constellation shaping.
Figure 20. ×
Tb/s capacity at the cost of 4 fiber pairs (C Band only) on top,
×
Tb/s capacity at the
cost of 2 fiber pairs (C + L Band) band bottom part of the figure.
A later experiment improved upon the previous record by employing a C + L trans-
mission achieving a capacity of 74.38 Tb/s over a transoceanic distance of 6300 km, by
using SMF and implemented hybrid Raman/EDF amplifiers, as was demonstrated in [
31
].
A novel demonstration of a dual C + L band transmission with an in-line dual-band
6-mode-EDFA achieving a total capacity of 266.1-Tbit/s over 90.4-km was proposed in [
32
].
Ionescu, M. et al. in [
33
] addressed the different design challenges and applications
of machine learning (ML) in modeling optical amplifiers. The authors concluded that the
optimization criteria of neural networks (NNs) depend on its various applications.
Therefore performing an NN architecture search is vital in determining the best-fit
ML model of an optical amplifier. Therefore, they proposed consideration of the NNs’
properties in an effort to simplify the network implementation and to achieve more accurate
future predictions.
Another experiment using the previous ML technique for optimization of a C + L band
hybrid EDFA-Raman amplifier was presented in [
34
]. Finally, two experiments achieving
4.88-b/s/Hz and 10.3-b/s/Hz net-spectral efficiency by using C + L band EDFAs were
demonstrated in [35,36], respectively.
4.4. Pump Farming (SDM) Technology
A straightforward approach to face increasing capacity needs is to increase the capacity
per fiber by boosting the output power of submarine repeaters (over 20 dBm) and/or by
using larger effective area (150
µ
m
2
) fibers. However, this approach is constrained by
current PFE (which can currently deliver power at around ~18 kW) and thus may not be cost
efficient. Pecci et al. in [
23
] introduced the optical and electrical formulas resulting that the
PFE and OSNR are strongly linked. In addition, cost savings can be made on the resistivity
of the conductor. In the upcoming SDM submarine era, to cope with future needs, new
motivations especially in the field of optical amplification, are ready for implementation.
Pump lasers or any other possible optical network component will have to be shared among
high count of FPs in the submarine cable (i.e., SDM systems in Table 3) both for fail-over
and economic reasons.

Telecom 2022,3258
A new innovation toward sharing submarine optical network equipment is repeater
pump farming (RPF). The RPF structure (presented in Figure 21) is based on a group (called
farm) of repeaters which are cross connected to each other. Both RPF and FPs groups
support each other and therefore, a group of optical pumps are used to support a group
of FPs. Moreover, RPF technology enhances the reliability of the submarine system even
if more than one pump failures occur, as it will continue to pump the FPs. Therefore, it
is a promising solution aimed at increasing the flexibility and reliability of a submarine
cable system. RPF technology consists of sharing x repeater pumps (x > 2) with other y
FPs (
y>1
), cross-connected via two-stage optical fiber couplers. On the other hand, in
terrestrial systems each pump is dedicated to one single FP.
Figure 21.
Shows a diagram of RPF systems for future subsea cable systems’ deployments. Further-
more, more complex designs implementing not only pump lasers but also EDFAs have already been
announced [13].
A variety of RPF schemes are proposed in [23], including:
â4 Pumps/2 Fiber Pairs, 4 Pumps/4 Fiber Pairs;
â8 Pumps/4 Fiber Pairs, 8 Pumps/8 Fiber Pairs; and
â16 Pumps/8 Fiber Pairs, 16 Pumps/16 Fiber Pairs.
4.5. Core Pumping (EDFA) Combined with SDM Technology
Depending on how amplifiers are being pumped, MC-EDFAs can be mainly divided
into two main categories, the core-pumped and cladding-pumped EDFAs. As noted earlier,
the implementation of these technologies is a key concept for the introduction of SDM. The
amplification schemes of first category core-pumped MC-EDFAs are based on sharing the
pump light of LDs among the cores inside the amplifying medium. A basic scheme for
MC-EDFA technology which presents core-pumping methods is shown in Figure 22a.
Table 7compares the main characteristics of core- and cladding-pumping MC-EDFAs.
Features such as mode, pump power,
λ
direction, number of LD’s, and use of cooler are
compared for both techniques.
Table 7. Cladding and Core-Pumping features.
Core Pumping Cladding Pumping
Mode Single-Mode Multi-Mode
Pump Power (W) Low (−5) High (−30)
Wavelength(λ) (nm)
Direction
1480
Backward
980
Forward
Number of Pump LD’s N-Cores 1
Presence of Cooling subsystem Yes No

Telecom 2022,3259
Figure 22.
Pumping methods for MC-EDFA. (
a
) Core pumping. (
b
) Cladding pumping. (
c
) Hybrid pumping.
Finally, authors in [
37
] proposed a multicore erbium-doped fiber (MC-EDFA) amplifier
which achieved simultaneous amplification for 7-cores composed of two tapered fibers
bundled (TFB) couplers with low insertion loss. The gain achieved in the MC-EDF fiber
was 30 dB and NF less than 4 dB. Finally, a record of 1.03-Exabit/s
×
km at 7.326 km
transmission by using a 7-Core-pumped EDFA amplifier was demonstrated in [38].
4.6. SDM Technology-Cladding Pumping EDFA (MC-EDFA)
SDM technology dictates hardware integration and energy saving especially in under-
sea environments where submarine amplifiers are deployed. The development of SDM-
based amplifiers is mandatory for the further deployment of SDM. A cladding-pumping
MC-EDFA can successfully meet these requirements and reduce the number of the pump
lasers connected with the spatial channels, compared to individually core-pumped am-
plifiers. Many demonstrations and experiments have already been done, boosting the
development of cladding-pumping amplifiers.
In [
39
] an optical gain greater than 15 dB with a noise figure smaller than 5.5 dB was
achieved with the implementation of a double-cladding MC-EDFA. In [
40
], the authors
presented both a C-band and an L-band 7-core EDFA with gains of 18.0 dB and 13.1 dB and
NF of 6.1 dB and 5.8 dB, respectively, showing that the core-absorption enhancement is
effective to increase output power of cladding-pumped multi-core EDFAs.

Telecom 2022,3260
Rahman, T. et al. in [
41
] demonstrated a 108 Tb/s transmission over 120 km by em-
ploying a 7-core cladding-pumping EDFA using hybrid modulation, bringing optimization
to the data throughput and achieving a net spectral efficiency of 39.27 bit/s/Hz. A 19-core
cladding pumped EDFA has also been reported for first time, achieving optical signal
amplification over 1500 km in [
42
]. Authors in [
43
] demonstrated a 0.715 Pb/s transmission
over a distance of 2009 km, equivalent to a throughput-distance product of
1.44 Eb/s ×km
which is the largest throughput
×
distance product achieved for demonstrations using
SDM-based amplifiers.
In [
44
], a 32-core amplifier has been used to demonstrate 1 Pb/s transmission over
205 km, achieving high-aggregate spectral efficiency of 217.6 b/s/Hz. Another 32-core
demonstration counted a greater than 17-dB gain and a 6.5 dB NF by using MC-EYDFA. A
distance of over 1850 km resulted in power consumption benefits as the number of cores
are scaled in MC-EYDFAs, as presented in [45].
A schematic of an SDM EDFA depicting all sections of amplification process is pre-
sented in Figure 23.
Figure 23. ×Tb/s schematic of a fully integrated SDM EDFA amplifier.
4.7. Pump Recycling (SDM) Technology
Pump-recycling technology is a promising candidate to enhance optical amplifica-
tion efficiency of cladding-pumped MC-EDFAs. This technique reuses part of the pump-
amplification light (after being separated from the signal light) and, by doing so, achieves
higher amplification efficiency. As a higher volume of pump optical power inserts into
the outer cladding of the double-cladding MC-EDFA, it is egressed, and thus it does
not take part in SDM signal amplification. More recently, a variety of schemes based on
pump-recycling technology have been proposed and experimentally studied supporting
7-core [24,46], and 19-core MC-EDFAs, respectively.
An MC-EDFA compared to a single-core EDFA (SC-EDFA) implements a pump splitter,
and there is a multiplex between the re-injected recycled pump in the spatial dimension
with the original signal. The pump splitter has a major role for the pump-recycling process
due to the fact that the optical power of the recycled pump light relies on its efficiency.
A variety of both fiber-type [
46
,
47
] and free space optics-type [
24
] pump splitters have
already been evaluated. Another difference concerns the pump combiners of SC-EDFA
and MC-EDFA because coupling loss due to the pump combiner for the cladding-pumped
EDFA is much smaller than its counterpart, the SC-EDFA.
A new investigation in pump-recycling technology, called the turbo cladding pumping
scheme (TCP) is presented in [
24
]. It improves the optical amplification performance by
3.5 dB
of cladding-pumping MC-EDFAs by implementing a paired free-optics combiner
and splitters [
17
]. Finally, an improvement of the pump recycling ratio of the turbo cladding-
pumped MC-EDFA from 12% to 42% by using paired prototypes of a spatial pump combiner
and splitter achieved a 1-dB greater optical gain, as studied in [48].

Telecom 2022,3261
4.8. Hybrid Core and Cladding Pump-Sharing EDFA (SDM) Technology
There are two basic techniques for implementing hybrid-pumped MC-EDFAs [
49
].
The first is a single-stage amplifier, and the second is a multi-stage amplifier. An important
characteristic of cladding-pumped EDFAs is their ability to realize automatic gain control
of the individual cores (an important aspect for an SDM MCFs-based network) by using
a WDM signal. This happens because optical power for each core in a cladding-pumped
MC-EDFA cannot be controlled independently between cores. In others words, the optical
gain cannot be controlled by adjusting the cladding-pump power. In its counterpart,
the core-pumped MC-EDFA, it is possible to adjust the gain of each core independently,
without making power consumption or network elements savings (a key concept for the
first generation of SDM technology).
A new and promising candidate to solve this issue is a hybrid structure which is com-
posed of both elements presented in Figure 24a,b. An MC-EDFA that employs both cladding
and core pumping and further introduces the gain control in a multi-core erbium/ytterbium-
doped fiber amplifier (MC-EYDFA) was extensively studied in [
50
]. Moreover, K. Abedin
reviewed in [
51
] the recent development of MC-EDFAs and presented an experimental
demonstration of core- and cladding-pumped MC-EFFAs amplifiers. The author concluded
that core pumping (besides being more expensive) allows an independent gain control
in each core. On the other side, cladding pumping requires fewer optical components by
using low-cost, energy-efficient multimode diodes. A small signal gain of >20 dB could be
achieved by employing a side-coupled pumping technique throughout the C-band, and
furthermore by extending the doped region beyond the core of few-mode doped fiber.
Figure 24.
(
a
) Hybrid + core pump-sharing. (
b
) Hybrid + variable core pump-sharing methods
for MC-EDFA.
Authors in [
52
] demonstrated a 256-Gb/s transmission by using a 16-QAM modulation
format for a distance of 404 km. The hybrid and core-pumping scheme MC-EDFA controlled
based on monitored temperature, achieving a reduction in power consumption of up to 38%.
Finally, a recent work [
49
] investigated the static and dynamic gain control characteristics
of hybrid-pumped single-stage and double-stage MC-EDFAs. Experiments in [
49
] studied

Telecom 2022,3262
the hybrid-pumped single-stage MC-EDFA with an average flattened gain of 21 dB and the
hybrid-pumped double-stage MC-EDFA with a maximum gain of 32 dB.
Table 8illustrates all available pumping schemes for MC-EDFA amplification.
Table 8. Available Pumping Schemes [13].
Pumping Scheme Acronym Core Pumping Clad Pumping Comments
Individual Core Pumping ICP Yes No Reference
Shared Core Pumping SCP Yes No With 3 dB coupler
Variable Shared Core Pumping VSCP Yes No With tunable coupler
Common Cladding Pumping CCP No Yes Need of core
attenuation
Hybrid with Individual
Core Pumping HICP Yes Yes -
Hybrid with Shared Core
Pumping HSCP Yes Yes With 3 dB coupler
Hybrid with Variable Shared
Core Pumping HVSCP Yes Yes With tunable coupler
4.9. SDM Technology-Multi-Mode EDFA (MM-EDFA)
Until now, there has been significant progress in multi-mode EDFA amplifiers which
succeed in amplifying signals propagating on multiple modes in a multi-mode fiber. A
novel proposal includes a few-mode gain-flattening filter (FM-GFF) based on long-period
fiber gratings (LPFG) in double-cladding few-mode fibers (DC-FMF). This proposed FM-
GFF scheme, which is profitable for the practical design of FM-EDFAs, was presented
in [53].
In [
54
], Sleiffer, V. et al. demonstrated a 73.7-Tb/s few-mode transmission by using
DP-16QAM modulated signals over 119 km cooperating with an inline multi-mode EDFA.
Authors concluded that by reducing the mode coupling losses, a further increase in trans-
mission distance may be achieved. In [
55
], authors measured the modal gain and noise
figure characteristics of a few-mode EDFA supporting LP01 and LP11 propagation in the
C-band. Experimental results showed that using a ring-doped EDFA enables modal gain to
be equalized to ~11 dB at a useful gain value of ~12 dB per mode.
A technique implementing a refractive index profile of EDFA’s core by using a 3-mode
ring-shaped core profile with a low DMG (1 dB) was studied in [
56
]. Another experiment
in [
57
] achieved a DMG of 3 dB by optimizing the LP11 mode pump profile by using
a phase mask. Finally, in [
58
] a trench-assisted 6-mode L-band EDFA with a very low
differential modal gain compared with cladding pumping was showed to achieve a gain
efficiency increase of 23.1% and a smaller differential modal gain when both average gains
reached 20 dB.
A latest proposed design of ring-core few-mode fiber (RC-FMF) with the erbium
doping at the cladding region near the outer ring edge (demonstrated in [
59
]) showed
that intensity overlapping difference between LP 01 and LP 11 modes can be diminished,
leading to a DMG of 0.22 dB. Moreover, the saturated input power was enhanced from
−17.7 dBm for LP01 mode and −16 dBm for LP11 mode to −8.5 dBm simultaneously.
4.10. Experimental Demonstrations of Submarine SDM-Based Amplifiers at
Transoceanic Distances
Most long-haul transmission experiments that adopt optical amplifiers can be divided
into three main categories: circulating loops, test beds, and finally free trial or specific
measurements performed on deployed systems.
The first category’s evaluation was held in 1991 to examine the feasibility of optical
amplifier transmission systems. On the other hand, loop transmission experiments are
a useful tool for optical amplifier feasibility demonstrations, measuring the BER of long,
pseudo-random data patterns introduced by Bergano et al. in [
60
,
61
]. Later (1992 to 1993),
long amplifier chains were developed for laboratory use as a test bed for establishing design
parameters and feasibility of monitoring systems concepts for submarine components.

Telecom 2022,3263
Lastly, specific measurements or today’s named free trials occurred on the first installed
amplifier systems (1994 to 1995) in order to determine the feasibility of upgrading a system
before or after its installation.
In 1995, Neal S. Bergano et al. [
62
] performed a 100-Gb/s transmission experiment
in which 20
×
5-Gb/s NRZ data channels were transmitted over 6300 km in 11.4 nm of
optical bandwidth by using a gain-flattened EDFA chain. In this experiment, the usable
EDFA bandwidth was increased by a factor of 3 by using long-period fiber grating filters
(Vengsarkar, Lemaire, Jacobvitz, et al. in 1995) as gain equalizers. Afterward, the 100-Gb/s
WDM experiment was eventually extended to 9300 km, adopting a low noise, 980-nm
pumped amplifier chain. NRZ was the transmission format selected, combined with a
synchronous polarization modulation and amplitude modulation (Bergano et al. in 1996).
Authors in [
38
] demonstrated a 140.7-Tbit/s transmission consists of 7
×
201-channel
super-Nyquist 100-Gbit/s WDM signals over 7326 km by using 7-core EDFAs, achieving
a record capacity distance product of 1 Exabit/s
×
km. Maxim A. Bolshtyansky in [
63
]
overviewed past and present similarities and differences in the requirements of submarine
and terrestrial amplifiers. He studied the impact of these requirements in the design of
submarine amplifiers and finally discussed future possibilities. In [
64
], Oleg V. Sinkin et al.
experimentally found the optimal SE for a 12-core fiber, SDM-based transmission using
EDFAs. Their findings showed the optimal SNR and SE values that can maximize the power
efficiency of a power limited SDM-based transmission system. Also, their experimental
results showed that the optimum SNR and SE do not depend on system parameters (i.e.,
the transmission distance) and a system operating at the Shannon limit is most power
efficient at SNR ≈1.72 and SE ≈2.89 b/s/Hz.
A long-haul MCF transmission of 2520 km using cladding-pumped 7-core EDFAs was
experimentally demonstrated in [
65
]. Authors confirmed that 73
×
128-Gbit/s Nyquist-pulse-
shaped dual-polarization QPSK signals were successfully transmitted in the above distance.
Benjamin et al. in [
43
] demonstrated the potential of high core-count EDFAs to support
high-capacity transmission over transoceanic distances with SDM integration. Their results
showed that the transmission throughput achieved by 2 cores using PDM and QPSK
modulation for the distance of 8007 km was 18.6 Tb/s and 14.5 Tb/s, respectively.
Another recent circulating loop experiment using innovative SDM-based ampli-
fiers was demonstrated and experimentally studied in [
66
]. This study demonstrated
a
130.8 Tb/s
SDM transmission in full C-band over transoceanic distance of 12,700 km
using 12-core MCFs (110
µ
m
2
) (Figure 25). Regarding power efficiency, both a ~70 mW of
pump power per path EDFA and a new 8D-QPSK modulation format with SE of 2.2 b/s/Hz
were implemented. However, the fundamental network elements of this structure were
hybrid micro-assembly-based EDFAs, which are being developed for applications requiring
a small form factor such as plug-in modules (showed in Figure 25). These micro-assembly
EDFAs seem to be an important candidate for supporting the future era of SDM amplifiers,
considering the limitations in repeater housings.
The experimental results showed that the use of compact amplifiers and the improve-
ments in power efficiency can effectively provide a significant increase in the number of
transmission paths, which is a crucial factor to achieve higher capacities in the future.

Telecom 2022,3264
Figure 25.
Schematic structure of a transmission test bed based on 12-core fiber and hybrid
microassembly-based amplifiers with fan in–out devices, wavelength selective switch (WSS), and a
loop synchronous polarization controller (LSPC) [66].
Pascal Pecci et al. in [
23
] studied various SDM designs for different target performances
(OSNR, GOSNR, capacity, etc.) at transpacific and transatlantic lengths. They pointed out
that long-haul systems with classical repeaters without pump farming have no chance
to increase submarine systems’ capacity, and so, the introduction of the pump farming
concept seems mandatory.
In [
28
], Turukhin, A. et al. made a circulating loop experiment demonstrating a
105.1-Tb/s
submarine transmission in a 12-core fiber over ~14.350 km, using a power-
efficient 8 DAPSK modulation format and optical EDFA amplifiers. To simulate a submarine
system, they limited the total pump power used by all 12 EDFAs (Figure 26) that may reside
in one repeater to that of a single, standard LD with an ~800 mW rating. Their results
confirmed another record capacity-distance product of 1.51 Pb/s
×
km compared to the
previously reported world record distance product of 1.03 Eb/s ×km [38].
Figure 26. Schematic structure of circulating loop experiment composed of 12 EDFAs [28].
Figure 27 demonstrates the most recent long-haul transmission (at transoceanic dis-
tances) experiments implementing different types of EDFA amplifiers [38,41–45,54,65,67].

Telecom 2022,3265
Figure 27.
Up-to-date long-haul transmission experiments at transoceanic distances implementing
different types of EDFA amplifiers.
5. Internal Architectures of Submarine Cable Systems and BUs
Regarding submarine architectures and topologies, we have four options which were
shown in [
9
,
25
]. If the distance between two stations is small or there is not a nearby third
land station that has to be connected, a point-to-point topology is used (Figure 28a). The
trunk-and-branch topology is used when we split a point-to-point link to support other
neighboring land nodes (Figure 28b). The ring topology (Figure 28c), as its name suggests,
makes use of a ring and serves the obvious: flexibility, connectivity, and redundancy. For
example, in closed seas like the Mediterranean Sea or the Aegean Sea (in Greece) [
13
], the
relatively short distances between land stations not only favor point-to-point topology but
also can easily use rings needed for redundancy.
On the other hand, transatlantic connections (i.e., from the UK to the US), with no
intermediate stations to connect, there is no other choice than the point-to-point topology.
In cases where unfriendly neighboring countries must be connected to a submarine cable
(and do not want to use direct links between each other), the trunk-and-branch option is
the most appropriate. Finally, the mesh topology (Figure 28d) can be applied in closed seas
or nearby islands.
Although CDC-ROADMs are already used in terrestrial optical networks, their use in
submarine systems was not (until recently) straightforward due to demanding underwa-
ter specifications. However, ROADMs can be used only in submarine systems in closed
seas, where they can be placed and operated from land. Recently announcements [
68
] in-
clude submarine ROADMs and reconfigurable BUs (RBUs) based on passive mux/de-mux,
and submarine specs WSSs. The former suffers from limited reconfiguration capabilities
whereas the latter questions system reliability and increases in cost and complexity. More-
over, as RBUs will use additional filtering devices, they will probably need more amplifiers
and power, which may limit their applications.
To cope with all these problems, Md. Nooruzzaman et al. in [
69
] studied the possible
savings from the use of recently proposed filterless optical network architecture. “Filterless”
refers to submerged components. Its main idea was to avoid active switching and filtering
in submerged devices in an effort to avoid the above-mentioned issues. Therefore, the
architecture is based on agile edge nodes and coherent transponders in the terminal dry
plant and passive colorless combiners and splitters in the submerged BUs. The main
objective of this architecture is to perform the light path reconfiguration from the land
nodes. This architecture was simulated under time-varying offered traffic and proved to be
a resource and cost-saving alternative.

Telecom 2022,3266
Figure 28.
Underwater connecting options. (
a
) Point to Point topology. (
b
) Trunk and Branch
topology. (c) Ring topology. (d) Mesh topology.
As expected, the branching architecture and the BU’s layout affect the number and use
of other submarine network components. Md. Nooruzzaman et al. in [
25
] studied various
Bus’ potential architectures and investigated the number of network components required
for an MCF-based submarine network. Moreover, they focused on the crosstalk effects in
long-haul submarine systems and proposed a trunk-and-branch architecture using passive
OADM BUs for an ultra-long-haul submarine network. However, the authors noted that
this architecture did not support submerged BUs’ reconfiguration. Finally, they studied
the effect of crosstalk as is seems that it can affect the benefits of MCFs. After conducting
various experiments, they concluded that selecting the proper number of FPs in an MCF
cable is crucial to keep crosstalk effects at low levels.
D. Kovsh in [
70
] reported on various branching architectures depending on a network’s
geographical limits. Regarding transoceanic networks the usual is to use one MCF per
connection and if needed split it to several branches near the shore. PFE is provided by
stations on each side (in the order of 15–20 kVs) and may limit the total capacity due to
long distances.
For regional networks, the options are a trunk to many branches, a mesh topology
with nested branches, and the connection of a submarine network to terrestrial bypasses
(like some roads in northern Norway heading for the Nord Cape). Power can be supplied
from many spots and must be carefully designed to provide the necessary redundancy. As
already mentioned, the only options for connecting relatively small distances are point-to-

Telecom 2022,3267
point links. In this case, power is coming from land stations and there is no need for tens
of kVs as the distances are not demanding. For special-purpose submarine networks (like
oil/gas mining companies’ networks), a trunk to many connections’ architecture is suitable.
Power feeding is straightforward for the stations and has to be easy to reconfigure in order
to cope with failures or programmed maintenance.
The author of [
70
] summarized the main designing decisions of a submarine cable
system: electrical (PFEs), optical (fibers, repeaters, amplifiers, span length), architecture
(type of branching), route selection, and economics.
As the wet plant of submarine networks is usually owned by several companies, L.
Garrett in [
19
] noted all system design issues that affect the total cost. A general rule
regarding route selection (for transoceanic systems) is to follow a direct, great-circle path.
Of course, many other constraints may apply. Ocean trenches, countries specific restric-
tions, military prohibitions, and/or danger areas also affect the design. As noted earlier,
sometimes the power scheme needs to be reconfigured for various reasons. A solution to
that is the power-switched bus, which can select the power providing the line based on
the distance from PFEs, network needs, possible failures, or scheduled maintenance. An
example of such power-switched BUs is shown in Figure 29.
In Figure 29a, all BUs are powered by both the land trunk stations’ PFEs, and branch
cables are powered by branches’ PFEs. In Figure 29b, Trunk A provides power to BU1 and
Branch station 2 (through BU2). Trunk B provides power to BU3. The power of Branches
1 and 3’s power scheme remains as in Figure 29a. This feature of reconfiguring power
schemes can be utilized to provide power (in case power is lost in a network section) or
cut off power (in case of scheduled hardware maintenance or upgrades). Although BUs
contribute to the formulation of the submerged architecture, their “static connection” nature
does not provide any flexibility. Reconfiguration of submarine networks is needed not only
when traffic patterns change but also when programmed maintenance or upgrades need to
deactivate specific devices.
As noted in [
70
], two products can enable submarine network reconfiguration: the
enhanced BUs (eBUs) and the recently announced submarine ROADMs.
Figure 29.
Reconfigurable power switched schemes for BUs. (
a
) Dual-end power feeding on trunk
route (
b
) Reconfiguration of dual-end power feeding from one trunk station to the Branch 2 station.

Telecom 2022,3268
The eBUs can support both dynamic routing (by implementing optical switching in the
FPs), and selective power feeding to temporarily enable device deactivation. After subma-
rine specs, WSSs were announced in 2015 [
19
], and WSS ROADMs became waterproof and
were submerged, usually in the same housing, with BUs in 2019. A submerged ROADM
was deployed by SubComin 2019 [
70
] and can support up to four trunk FPs. ROADMs and
eBUs can achieve bidirectional, unidirectional, or 3D shared branches depending on the
traffic patterns and the clients or places they have to serve (Figure 30).
From the past to the future era of submarine terminal equipment (SLTE/CLS), there
is a lot of progress on transponders and power management of the cable systems as
demonstrated in Figure 31. The first generation of CLS, up until 2010, achieved data
rates of up to 40 G and baud rates of ~11 Gbaud. It used QPSK modulation formats and
implemented technologies such as coherent detection and PMD compensation.
The second generation (2012–2015), when compared to the features of the first genera-
tion, added higher data rates up to 200 G, baud rates of 28–35 Gbaud, enhanced 16QAM
modulation formats, SD FEC, and CD pre-dispersion. The third generation (2016–2019)
achieved data rates up to 400 G, modulation formats up to 32 QAM, modulation of 2D, 4D,
Nyquist shaping, and improved FEC.
Figure 30. Branch types implemented by ROADMs–eBUs.
Figure 31. Evolution of submarine line terminating equipment.

Telecom 2022,3269
Finally, in the fourth generation (2020+), the data rates increased up to 800 G, the
baud rates up to 95 Gbaud, new modulation formats up to 64 QAM, and features such as
const. shaping (PCS), improved FEC and nonlinear comp added to the newest generation
of this evolution.
Although the commissioning process (the final testing before going commercial) is
always performed by the cable provider, recently emerged devices for the dry plant, led
to the open cables design model [
19
]. In this model, the cable system is deployed by one
provider and several others can utilize it with their own SLTE.
6. Submarine Power Feeding
Submarine cable systems extend to ranges from hundreds to thousands of kilometers.
Power feeding in short extended systems is straightforward, but for long-haul systems
it is complicated. Long-haul systems incorporate large numbers of repeaters (one every
75–100 km [
3
]) and must ensure that all of them obtain power in a certain range of volts.
Unfortunately, power experiences losses over long distances. The use of very high voltage
power systems (to compensate for the losses) is not an option as such voltage may exceed
the maximum voltage specifications of the equipment (cables, repeaters/amplifiers) and
damage it. Typical voltages of PFE are in the range of 12–18 kVs and more recently
20 kVs
[
3
]. S. Desbruslais in [
71
] presented design methods that maximize the submarine
systems’ capacity considering PFE limitations. Optimization of the optical submarine
system based on the Gaussian noise (GN) model is accomplished either by seeking for
the optimum span loss that minimizes the PFE voltage or by seeking the span loss that
maximizes the capacity for a given PFE and optical to signal-to-noise-ratio (OSNR).
Pilipetskii et al. in [
22
] and Turukhin et al. in [
72
] initially reported some experiments
made for long-haul SMF experiments. These experiments showed that certain amount
of power can be spread over a wider range of bandwidth (or over a wider range of
space dimension), and can increase a system’s capacity. Moreover, to optimize power
efficiency an appropriate modulation format must be selected. Moreover, the authors
noted that the optimal system capacity is not achieved at the optimal efficient power
utilization region. Later experiments were based on a 12-core MCF, reaching 105.1 TB/s
over a
14,350-km, 22-nm
bandwidth EDFA, and 8D-APSK modulation operating at SE of
32 bit/s/Hz were also incorporated. Results revealed that SDM-based MCFs managed to
achieve higher capacities than those achieved by SMFs without the need for extra power. So,
as SDM seems promising, the next dilemma is how to allocate the PFEs’ power to several
SDM dimensions and find the range of SE values that can lead to maximum capacities.
Experiments showed that maximum capacities can be achieved in specific values of SE.
The authors concluded that SDM can provide the increased capacities and at the same
time manage the power efficiently. These two SDM characteristics are especially crucial in
long-haul submarine systems.
SDM in submarine networks is evolving not only because of the upcoming deployment
of MCFs/MMFs but also because of its ability to handle power more efficiently. The
Shannon formula dictates that cables’ capacity increases logarithmically with the signal-to-
noise (SNR) ratio. On the other hand, incorporating multiple numbers of fibers (either in an
MCF or an MMF or bundles of SMFs layout), capacity increases linearly by the number of
fibers. In long-haul submarine systems (where power is not and cannot become unlimited),
SDM can lower the power demands on each FP and feed power to more FPs. In this way,
we may achieve higher capacities with the same amount of power. Optical amplifiers in
submarine systems use an amount of power for their own processes and then convert
power to optical energy to boost fiber-optic signals. The optical power that amplifiers
feed fiber-optics defines the electrical to optical power conversion efficiency (EOPCE). As
EOPCE determines the number of FPs that can be powered, it also determines the total
capacity of a cable system. Moreover, as pool pumps are evolving in the submarine era,
amplifying several FPs get more complicated, and so the exact definition and attributes of
the EOPCE must be investigated.

Telecom 2022,3270
Liang, X. et al. in [
73
] analyzed EOPCE in SDM submarine cable systems through a
model. The model considered the repeaters topology, the power transfer model, several
measures of pump laser efficiency, and simulation results of amplifiers conversion efficiency.
Results showed that maximum pump EOPCE can be achieved from a specific range of
input pump power. Moreover, there exists an optimal combination of the number of pump
amplifiers and the power from their FPs that results in maximum EOPCE. The derived
model enables the analysis of the cable systems’ efficiency by calculating the total capacity
achieved for a given amount of electrical power.
Shrinivas, H. et al. in [
74
] studied power efficiency of SDM-based submarine systems
calculated from a total system’s capacity per unit of power consumption. The authors
reported on numerous experiments considering various values of system power and length
spans. They comment on potential performance gains that may occur by optimizing the
cable system design and they estimated the optimal number of FPs and the system’s total
capacity for specific submarine system’s length span.
Finally, Figure 32 demonstrates a comparison concerning power feeding versus system
capacity for two cable system links, one transpacific (10,000 km) and one transatlantic
(
6000 km
). In this scenario, the power-feeding values vary from 15 kV (for legacy systems)
to 18 kV (for the latest-generation systems) compared to system capacity, increasing from
16 to 32 FPs.
Figure 32.
Comparison of two cable systems for different power (from legacy to the latest generation)
and capacity values [75].
The results show that a 20% improvement of power output translates into an over
50% increase in system capacity by increasing the fiber pairs in a long-haul transoceanic
system, lowering the cost per Tb/s. Another important aspect is the enhanced reliability
resulting from the extension of single-end system power-feeding distance which is a crucial
parameter under fault conditions or during system-maintenance operations.
7. Economic Aspects of Submarine Networks
SDM is evolving in submarine networks as it promises increased capacities. Appar-
ently, these increased capacities will come along with a significant cost. Having in mind
that the submarine environment is quite different from the terrestrial one, the economics
of SDM-based submarine systems need to be investigated. Common sense dictates that
we tend to seek increased capacities up to the point we can afford them or they are no
longer worth the associated cost. There have been various economic aspects, models, and
proposals for defining and limiting the cost of an SDM-based submarine cable system. We
briefly discuss most of them in this section.
In Figure 33a there is a demonstration of how global cable construction costs during
the period from 2010 to 2020 derived a total investment cost of $7.6 billion for all the regions
(Figure 33b).
Dar, R. et al. in [
76
,
77
] explored cost saving cases in submarine systems. They first
noticed that cost/bit-optimized systems follow different design rules than those followed by
capacity-optimized ones. The authors focused on cost-optimized systems and showed that

Telecom 2022,3271
SDM-based submarine systems with 50 FPs, long spans, and MCFs can achieve lower costs
compared to the use of several single-fiber cables. They concluded that using SDM with
45 FPs could achieve a 44% cost saving compared with the current Pacific system, which
incorporates 8 FPs of PSCF110 and C-band optical amplifiers. Moreover, the integrations of
optical amplifiers (i.e., in pump-farming arrays) and transponders result in extra significant
cost savings. An interesting finding regarding power supply is that even if we could supply
increased power beyond known levels and achieve higher capacities, we would not result
in a reduced cost/bit. This happens because the cable and deployment costs (which are
responsible for a large percentage of the total cost) will not be affected.
As SDM and MCFs are evolving and amplifiers represent a significant percentage of
the total cost (especially in long-haul networks), methods of reducing amplifiers’ cost are
investigated. Thouras, J. et al. in [
78
] investigated economic issues of the upcoming state of
the art MC-EYDFA. MC-EYDFA’s architecture is described in the amplifiers section. The
authors performed a techno-economic analysis by using a 12-core EYDFA over a 2280-km
terrestrial network in France. The network consisted of 7 links, has 8 ROADMs with a
nodal degree varying from 2 to 5, and used 29 EDFAs. Results showed that in the future,
the replacement of EDFAs by MC-EYDFAs will achieve significant savings in the range of
33–35% for amplifiers cost (by year 2038) and of 60% for energy consumption (by 2026).
These savings lie in the range of tens of millions of euros in absolute values.
In addition to the technological research and progress in optical hardware (single-
core fibers, MCFs, MMFs, amplifiers, transponders) research has also been conducted on
advanced architectures that will efficiently support the demanding SDM technology. Asano
and Jinno [
79
] performed a cost comparison of several proposed hierarchical optical cross-
connect (HOXC) architectures which can be embedded in spatial core networks (SCNs).
The HOXC architectures examined are the baseline-stacked wavelength cross connect
(WXC) architecture, the full-size matrix (FS-MS)-based architecture, the sub-matrix switch-
based (Sub MS) architecture, the full-size core-selective switch (FS-CSS)-based architecture,
and the sub core selective switch (Sub-CSS)-based architecture. Although today there is
an uncertainty regarding what exactly these OHXC architectures will cost, the authors
concluded that the sub-MSs and CSSs SXC architectures will be more cost-effective over
the well-known full-size matrix and the ROADM/WXC architectures.
J. Downie in [
80
] investigated maximum subsea cable capacities and the minimum
cost/capacity for SC fibers by using C and C + L bands and MCFs using C-band. Basic
assumptions are fixed power constraints, specific SNR requirements and similar fiber core
main characteristics. The length of the test bed links were 6600 km and 10,000 km (suitable
for transatlantic cables). Results revealed that the MCF and SC C + L minimum cost capacity
were 20% and 5% higher than that for the SC C-band only for the 6600-km link and 37%
and 12% higher for the 10,000-km link. The author concluded that cable systems using high
count of SC fibers achieve lower cost/capacity values than those achieved by MCFs.
Downie et al. in [
81
] investigated costs between SMFs and MCFs for submarine
systems. The authors compared costs based on a system model which counts for two
different priorities: maximum cable capacity and minimum overall cost/capacity. They
concluded that although MCF systems provide greater cable capacities for a small number
of fibers, the cost/capacity ratios are significantly higher. When they focused on minimum
cost/capacity the MCF systems get an easy victory over SC systems. It was also noted that
submarine-deployed MCFs may need increased electrical power requirements (over the
SMFs) for feeding the MC-EDFAs and may suffer from possible physical space limitations.
Electrical power and physical sizes affect not only the electrical and deployment hardware
issues but also the cost. Consequently, for a proper cost analysis, these factors should be
taken into consideration.
Bolshtyansky et al. in [
82
] presented a model that considers not only economics but
also physical factors that describe a submarine connection. A techno-economic model was
shown in Figure 5[
79
] and included all cost-affecting parameters/factors. By using a “least
amount of technological change” approach, the authors concluded that a 16-FPs SDM-

Telecom 2022,3272
based system is a feasible yet economical solution for near-future submarine cable systems,
approaching the cost (per unit) capacity limit. More FPs in a single cable seems to have
serious drawbacks both for technology issues and for cost. Moreover, results for the cost
ratio both for optimized SDM and non-SDM solutions were presented. Physical and initial
costs are similar for both solutions, except for the operating optical power requirements.
Results showed that SDM gives a minimal cost benefit at short distances but significant
benefits (in the range of 14%) for longer distances and higher capacities.
Paximadis and Papapavlou et al. in [
13
,
83
], performed an economic study including
almost all of the equipment needed for a closed-sea submarine network. An analytical cost
model was described which derives the total relative cost (TRC) of each ROADM design
and total relative network cost (TRNC); thus the network cost includes the cost of network
components plus site survey costs. The cost model took into account all three available
WSS ROADM switching strategies (independent switching, fraction joint switching, and
joint switching) [
84
], several common spatial dimensions (of 4, 8, 12, and 16), the number
of add/drop (A/D) capability nodes, and the in-line amplifiers (both for EDFAs and
for MC-EYDFAs). The model was applied to two new submarine backbone networks
of 29 and 10 nodes specially designed for the Mediterranean Sea and the Aegean Sea
(Greece), respectively. A special characteristic of these submarine networks is that they
can act as supplemental and backup networks to adjacent terrestrial ones. The results
of the cost analysis showed that fraction joint switching turns out to be more economic
than independent switching, (5% cost savings for all spatial dimensions S = 4, 8, 12, 16).
Joint switching is also more cost effective than independent switching, also outperforming
fraction joint switching in all spatial dimensions (for S > 4). Beyond the S = 4 threshold,
joint switching achieves significant costs savings, in the order of 8% for S = 12. Results for
the amplifiers’ cost comparison showed that the use of MC-EYDFAs (over the EDFAs) in
the 29-node, 16,400-km span Mediterranean network can achieve cost savings of around
37%.
Fiber coating will be a cost-reduction factor in the near future. Specifically, in order
to cope with future capacity needs, increasing the number of FPs and keeping the costs
down are of important interest. As the space housing in a submarine cable is limited, a
novel solution is to develop fibers with thinner coating variants of approximately 200
µ
m,
and also the fiber out diameter. Although during the first 30 years of single-mode fiber
manufacturing, a coating diameter of 245–250
µ
m was a standard in the industry, OFS in
2014 launched a 200-
µ
m fiber to deal with the need of the higher fiber density demands.
Submarine fibers with these characteristics can give up to a 50% higher fiber density by a
reduction on both the primary and secondary coatings and an existing fiber diameter at
125 µm, which guarantees that the existing fiber-processing procedures can be used.
In [
85
], John D. Downie, et al. made a techno-economic analysis for the application
of MCFs consisting of 2–4 weakly coupled cores and reduced diameter fibers (RDF) in
high-capacity submarine cables simulating the transatlantic link length. Their research
evaluated three system design methods in order to examine the use of MCF in submarine
cable systems. They compared two different diameter SCFs with MCFs exploring cable
capacity and cost/bit achieved. Their study’s results showed that MCFs can provide
higher cable capacity if FP limits are imposed, but likely not at lower cost/bit unless
optimistic and best-case assumptions are considered with respect to MCF relative fiber
cost. Moreover, their results showed that the reduced diameter SCFs can bring much of the
density and cable cost reduction that motivates interest in MCF without the challenges of a
new ecosystem as required by MCF. However, MCF may trigger the design of the highest
cable capacities, such as 1 Pb/s, which cannot be possible with SCFs unless important cable
changes are made.

Telecom 2022,3273
Figure 33.
(
a
) Global cable construction costs for the period ranging from 2010–2020. (
b
) Costs by
region [86].
From Figure 34 we can conclude that there is one lowest cost/bit “optimal” design
and capacity for each FP bringing out the SDM benefit. Also, an optimal design is below
maximum capacity/FP, achievable with the given line card by ~1 dB. In the case of a
power-limited submarine system (i.e., a very long distance requiring many FPs) “optimal”
design has to be more “linear” in order to be more power efficient.
Figure 34. Cable cost per bit for 8, 12 and 16 FPs [70].
8. Submarine Networks Security
As submarine networks are the main component of Earth’s global backbone network,
security issues are of critical importance. Submarine networks are “the world’s information
super-highways” and carry more than 99% of international traffic between continents.
Keeping in mind that the submarine environment is not as friendly, stable, and predictable
as the terrestrial one, all possible fail causes have to be explored and faced. Fishing
techniques with large ships, anchor dragging, and more rarely earthquakes are among the
causes for most damage done to cables [87].
In addition, major “digital threats” such as backdoors during the cable manufacturing
process, targeting onshore landing stations and facilities linking cables to networks on land,
or tapping the cables at sea are of important interest. The nightmare scenario of a hacker
who gains control, or administrative rights, of the submarine network management system
may result in the hacker discovering vulnerabilities, interrupting or diverting data traffic,

Telecom 2022,3274
or even executing a “kill click” by which he or she might delete the wavelengths which
transmit data has to be prevented. For all these reasons, live cable monitoring techniques
are proposed to predict, diagnose, and program the proper restore actions. Figure 35 shows
the percentages of various threats for a submarine cable system. As is evident, fishing and
anchorage are responsible for about the two thirds of cable damages.
Figure 35. Dangers of submarine cable systems (in percentages).
9. Can We Predict the Future in Submarine Networking?
As subsea traffic demands have been growing rapidly, SDM is becoming a de-facto
subsea technology promising to solve bandwidth-demand issues, avert the capacity crunch,
and reduce the cost/bit. In this review, we presented and extensively analyzed all the latest
technologies and innovations in the sector of submarine communications focusing on SDM
technology. We both presented and analyzed the majority of the latest submarine cable
systems (already existed or planned), examined their fundamental components’ technology
and combined this with up-to-date long-haul experiments (at transoceanic distances). As
for the new submarine era presented in Figure 36, (e.g., for the evolution of undersea
amplifiers), the history of which was full of innovations and milestones, we are expecting a
promising future speculating that we can benefit from SDM technology which focuses on
integrated elements with small undersea housing [
22
] and smaller output power, allowing
us to share pumps across multiple amplifiers through techniques such as pump-farming
amplification schemes.

Telecom 2022,3275
Figure 36. Evolution of submarine system’s technology.
As for submarine cables, higher fiber counts of up to 50 parallel fiber pairs per ca-
ble [
77
], combined with cost-effective aluminum conductors and smart environmental
sensors [
87
], will be a near-future concept achieving lower cost/bit compared to already
existing systems. The trend is that SDM systems with thinner coating variants (200
µ
m)
will enable up to 50% higher fiber density [
10
] within an existing cable design. Ultra-low
attenuations will play a crucial role, as every improvement (i.e., 0.001 dB/km) will lead to
cost/bit reduction. An alternative ultimate future solution will be based on very high fiber
count cables (enabled by the 200-
µ
m coating fiber presented in Figure 37) in combination
with C + L band technology and ultra-low attenuation, but to our knowledge with lower
possibilities than the SDM.
Figure 37. Comparison of recent and future fiber coating technology.
As for multicore fibers, additional losses from fan in–out devices and likely higher
typical attenuation compared to SC fibers, will reduce the achievable cable capacity as it is
likely that the cost/per core of an n-core MCF will be higher than the cost of “n” individual
fibers, but as new manufacturing techniques continue to evolve, MCFs can “cut the ribbon”
and win its SC counterparts. On the other hand, ROADMs and eBUs supporting optical
switching and flexible traffic routing of many trunks, combined with greater than
800 G
transponders, high order modulation formats (
≥
256 QAM) achieving high baud rates
(200 Gbaud) will be a future aspect too. Moreover, the optimum spectral efficiency range
(2–3 b/s/Hz) compared with traditional technology of (5 b/s/Hz) will be a future trend

Telecom 2022,3276
too. Furthermore, future demands dictate that more electrical power should be launched
into the submarine cables, by either increasing the voltage (more than 20 kV) or by using
lower resistivity cable because recent research show that a 20% improvement of power
output translates into over a 50% increase in system capacity by increasing the number
of fiber pairs. These steps will be mandatory to achieve 500-terabit/s cables or even 1-
petabit/s cables in the near future. Finally, security is a crucial parameter as vulnerabilities
of submarine cables could make the entire global telecommunication to break down so new
optical techniques are needed to solve present and future security issues [88,89].
Our study concludes that SDM technology will be the most important candidate as it
meets all key requirements to support the new submarine era.
Author Contributions:
Conceptualization, C.P.; writing—original draft preparation, C.P. and K.P.;
data curation, K.P. and D.U.; writing—review and editing, I.T. All authors have read and agreed to
the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The data used are contained in this article.
Conflicts of Interest: The authors of this article declare no conflict of interest.
References
1.
History of the Atlantic Cable & Undersea Communications. Available online: https://atlantic-cable.com/Cables/CableTimeLine/
(accessed on 5 March 2022).
2. Available online: https://www.submarinecablemap.com/ (accessed on 5 March 2022).
3. Available online: https://www.submarinenetworks.com/ (accessed on 7 March 2022).
4. Available online: https://subtelforum.com/ (accessed on 7 March 2022).
5.
List of International Submarine Communications Cables. Available online: https://en.wikipedia.org/wiki/List_of_international_
submarine_communications_cables (accessed on 4 February 2022).
6.
Shariati, B.; Klonidis, D.; Marom, D.M.; Comellas, J.; Velasco, L.; Tomkos, I. Spectrally and Spatially Flexible Optical Networks:
Recent Developments and Findings. In Proceedings of the 2018 20th International Conference on Transparent Optical Networks
(ICTON), Bucharest, Romania, 1–5 July 2018; pp. 1–4. [CrossRef]
7.
Ciena. New Metrics Use Cases and Considerations. Simplifying Open Submarine Cable Link Engineering. 2021. Available
online: https://www.ciena.com/insights/articles/simplifying-open- submarine-cable- link-engineering- blog.html (accessed
on 7 March 2022).
8. SDM1 BY ASN. Available online: https://web.asn.com/SDM1.html (accessed on 7 March 2022).
9. Papapavlou, C.; Paximadis, K.; Tzimas, G. Progress and Demonstrations on Space Division Multiplexing. In Proceedings of the
2020 11th International Conference on Information Intelligence, Systems and Applications, Piraeus, Greece, 15–17 July 2020;
pp. 1–8. [CrossRef]
10.
Makovejs, S.; Hedgpeth, J. Fiber Technology for Subsea Networks Today vs. Tomorrow. Submarine Telecoms Magazine, Issue 117.
26 March 2021.
11.
Markow, A. Summary of Undersea Fiber Optic Network Technology and Systems. Available online: https://studylib.net/doc/83
81708/summary-of-undersea-fiber-optic-network-technology/ (accessed on 4 February 2022).
12.
High Fiber Count Solution: Bring Larger Capacity to the World. 2021. Available online: https://www.hmntechnologies.com/
enjsrdB.jhtml (accessed on 8 February 2022).
13.
Papapavlou, C.; Paximadis, K.; Tzimas, G. Design and Analysis of a new SDM submarine optical network for Greece. In
Proceedings of the 12th International Conference on Information Intelligence, Systems and Applications (IISA2021), Piraeus,
Greece, 15–17 July 2021.
14.
Nakamura, K.; Inoue, T.; Matsuo, Y.; Inada, Y.; Mateo, E. 60+ Gbaud real-time transmission over a 17.680 km line designed for
GEN1-SDM submarine networks. In Proceedings of the 2020 Opto-Electronics and Communications Conference (OECC), Chiba,
Japan, 2–8 October 2020; pp. 1–3. [CrossRef]
15.
Zhang, H.; Turukhin, A.; Sinkin, O.V.; Patterson, W.; Batshon, H.G.; Sun, Y.; Davidson, C.R.; Mazurczyk, M.; Mohs, G.; Foursa,
D.G.; et al. Power-efficient 100 Gb/s transmission over transoceanic distance using 8-dimensional coded modulation. In
Proceedings of the 2015 European Conference on Optical Communication (ECOC), Valencia, Spain, 27 September–1 October 2015;
pp. 1–3. [CrossRef]

Telecom 2022,3277
16.
Mertz, P.; Grubb, S.; Rahn, J.; Sande, W.; Stephens, M.; O’Connor, J.; Mitchell, M.; Voll, S. Record Ultra-High Full-Fill Capacity
Trans-Atlantic Submarine Deployment Ushering in the SDM Era. In Proceedings of the 2020 Optical Fiber Communications
Conference and Exhibition (OFC), San Diego, CA, USA, 8–12 March 2020; pp. 1–3.
17.
Takeshita, H.; Sato, M.; Inada, Y. Past, Current and Future Technologies for Optical Submarine Cable. In Proceedings of
the IEEE/ACM Workshop on Photonics—Optics Technology Oriented Networking, Information and Computing Systems
(PHOTONICS), Denver, CO, USA, 18 November 2019.
18.
Takahashi, H.; Soma, D.; Tsuritani, T. Promising Technologies for Future Submarine Cable Systems. In Proceedings of the 2020
Opto-Electronics and Communications Conference (OECC), Taipei, Taiwan, 4–8 October 2020.
19. Garett, L. Design of Global Submarine Networks. J. Opt. Comm. Netw. 2018,10, A185–A195. [CrossRef]
20.
TE SubCom and Nistica Announce Availability of Undersea Qualified Wavelength Selective Switch Modules. Available
online: https://www.subcableworld.com/newsfeed/technology/te-subcom- and-nistica- announce-availability- of-undersea-
qualified-wavelength-selective-switch-modules (accessed on 10 February 2022).
21.
Papapavlou, C.; Paximadis, K.; Tzimas, G. Analyzing ROADM cost in SDM networks. In Proceedings of the 11th Int. Conf. on
Information Intelligence, Systems and Applications (IISA 2020), Piraeus, Greece, 15–17 July 2020.
22.
Pilipetskii, A.; Sinkin, O.; Turukhin, A.; Bolshtyansky, M.; Foursa, D. The Role of SDM in Future Transoceanic transmission
Systems (invited). In Proceedings of the European Conference on Optical Communication (ECOC), Gothenberg, Sweden,
17–21 September 2017. [CrossRef]
23.
Pecci, P.; Jovanovski, L.; Barezzani, M.; Kamalov, V.; Marcerou, J.; Cantono, M.; Gumier, M.; Courtois, O.; Vusirikala, V. Pump
Farming as Enabling Factor to Increase Subsea Cable; SubOptic: Nozay, France, 2019.
24.
Takeshita, H.; Matsumoto, K.; Hasegawa, H.; Sato, K.I.; de Gabory, E.L.T. Improved Optical Amplification Efficiency by using
Turbo Cladding Pumping Scheme for Multicore Fiver Optical networks. IEICE Trans. Commun.
2019
,102, 1579–1589. [CrossRef]
25.
Nooruzzaman, M.; Morioka, T. Multicore Fibers for High-Capacity Submarine Transmission Systems. J. Opt. Commun. Netw.
2018,10, A175–A184. [CrossRef]
26.
Cai, J.-X.; Bathson, H.G.; Mazurczyk, M.V.; Sinkin, O.G.; Wang, D.; Paskov, M.; Patterson, W.W.; Davidson, C.R.; Corbett, P.C.;
Wolter, G.M.; et al. 70.46 Tb/s Over 7600 km and 71.65 Tb/s Over 6970 km Transmission in C + L Band Using Coded Modulation
With Hybrid Constellation Shaping and Nonlinearity Compensation. J. Lightwave Technol. 2018,36, 114–121. [CrossRef]
27.
Mazurczyk, M.; Foursa, D.G.; Batshon, H.G.; Zhang, H.; Davidson, C.R.; Cai, J.-X.; Pilipetskii, A.; Mohs, G.; Bergano, N.S. 30 Tb/s
Transmission over 6630 km Using 16QAM Signals at 6.1 bits/s/Hz Spectral Efficiency. In European Conference and Exhibition on
Optical Communication, OSA Technical Digest (Online); Optica Publishing Group: Washington, DC, USA, 2012; Paper Th.3.C.2.
28.
Turukhin, A.; Sinkin, O.V.; Batshon, H.G.; Zhang, H.; Sun, Y.; Mazurczyk, M.; Davidson, C.R.; Cai, J.-X.; Bolshtyansky, M.A.;
Foursa, D.G.; et al. 105.1 Tbps power-efficient transmission over 14.350 km using a 12-core fiber. In Proceedings of the 2016
Optical Fiber Communications Conference and Exhibition (OFC), Anaheim, CA, USA, 20–24 March 2016; pp. 1–3.
29.
Cai, J.-X.; Vedala, G.; Hu, Y.; Sinkin, O.V.; Bolshtyansky, M.A.; Foursa, D.G.; Pilipetski, A.N. 9 Tb/s Transmission using 29
mW Optical Pump Power per EDFA with 1.24 Tb/s/W Optical Power Efficiency over 15,050 km. J. Lightwave Technol.
2022
,40,
1650–1657. [CrossRef]
30.
HMN Tech Launches 32FP Petabit-Level Repeater Prototype. 2021. Available online: https://subtelforum.com/hmn-tech-
launches-32fp-petabit-level-repeater-prototype (accessed on 10 February 2022).
31.
Ionescu, M.; Lavery, D.; Edwards, A.; Sillekens, E.; Semrau, D.; Galdino, L.; Killey, R.I.; Pelouch, W.; Barnes, S.; Bayvel, P. 74.38
Tb/s Transmission Over 6300 km Single Mode Fiber Enabled by C + L Amplification and Geometrically Shaped PDM-64QAM. J.
Lightwave Technol. 2020,38, 531–537. [CrossRef]
32.
Wakayama, Y.; Soma, D.; Beppu, S.; Sumita, S.; Igarashi, K.; Tsuritani, T. 266.1-Tbit/s Repeated Transmission over 90.4-km
6-Mode Fiber Using Dual C + L-Band 6-Mode EDFA. In Proceedings of the 2018 Optical Fiber Communications Conference and
Exposition (OFC), San Diego, CA, USA, 11–15 March 2018; pp. 1–3.
33.
Ionescu, M.; Ghazisaeidi, A.; Renaudier, J. Machine Learning Assisted Hybrid EDFA-Raman Amplifier Design for C + L Bands.
In Proceedings of the 2020 European Conference on Optical Communications (ECOC), Brusseles, Belgium, 6–11 December 2020;
pp. 1–3. [CrossRef]
34.
Da Ros, F.; de Moura, U.C.; Luis, R.S.; Rademacher, G.; Puttman, B.J.; Rosa, A.M.; Carena, A.; Awaji, Y.; Furukawa, A.; Zibar, D.
Optimization of a Hybrid EDFA-Raman C + L Band Amplifier through Neural-Network Models. In Proceedings of the 2021
Optical Fiber Communications Conference and Exhibition (OFC), Washington, DC, USA, 6–11 June 2021; pp. 1–3.
35.
Zhang, H.; Davidson, C.R.; Batshon, H.G.; Mazurczyk, M.; Bolshtyansky, M.; Foursa, D.G.; Pilipetskii, A. DP-16QAM based coded
modulation transmission in C + L band system at transoceanic distance. In Proceedings of the 2016 Optical Fiber Communications
Conference and Exhibition (OFC), Anaheim, CA, USA, 20–22 March 2016; pp. 1–3.
36.
Yu, Y.; Jin, L.; Xiao, Z.; Yu, Z.; Lu, Y.; Liu, L. 100.5 Tb/s MLC-CS-256QAM Transmission over 600-km Single Mode Fiber with C
+ L Band EDFA. In Proceedings of the 2018 Asia Communications and Photonics Conference (ACP), Hangzou, China, 26–29
October 2018; pp. 1–3. [CrossRef]
37. Abedin, K.S.; Taunay, T.F.; Fishteyn, M.; Yan, M.F.; Zhu, B.; Fini, J.M.; Monberg, E.M.; Dimarcello, F.V.; Wisk, P.W. Amplification
and noise properties of an erbium-doped multicore fiber amplifier. Opt. Express 2011,19, 16715–16721. [CrossRef]
38.
Igarashi, K.; Tsuritani, T.; Morita, I.; Tsuchida, Y.; Maeda, K.; Tadakuma, M.; Saito, T.; Watanabe, K.; Imamura, K.;
Sugikazi, R.; et al
.
1.03-Exabit/s
×
km Super-Nyquist-WDM transmission over 7326-km seven-core fiber. Opt. Express
2014
,22, 1220–1228. [CrossRef]

Telecom 2022,3278
39.
Mimura, Y.; Tsuchida, Y.; Maeda, K.; Miyabe, R.; Aiso, K.; Matsuura, H.; Sugizaki, R. Batch Multicore Amplification with
Cladding-Pumped Multicore EDF. In European Conference and Exhibition on Optical Communication, OSA Technical Digest (Online);
Optical Society of America: Washington, DC, USA, 2012.
40.
Tsuchida, Y.; Maeda, K.; Watanabe, K.; Takeshima, K.; Sasa, T.; Saito, T.; Takasaka, S.; Kawaguchi, Y.; Tsuritani, T.; Sugisaki, R.
Cladding Pumped Seven-Core EDFA Using an Absorption-Enhanced Erbium Doped Fiber. In Proceedings of the 42nd European
Conference on Optical Communications, Düsseldorf, Germany, 18–22 September 2016 . paper M.2.A.2.
41.
Rahman, T.; Spinnler, B.; Calabro, S.; de Man, E.; Pulverer, K.; Castro, C.; Mizuno, T.; Miyamoto, T.; Takenaga, K.; Jain, S.; et al.
108 Tb/s Transmission over 120 km of 7-Core Multicore Fiber with Integrated Cladding Pumped Multicore Amplifiers. In
Proceedings of the 2018 European Conference on Optical Communication (ECOC), Rome, Italy, 23–27 September 2018; pp. 1–3.
[CrossRef]
42.
Soma, D.; Beppu, S.; Maeda, K.; Takasaka, S.; Sugisaki, R.; Takahashi, H.; Tsuritani, T. Long haul MCF Transmission using Full C +
L-band 19-core Cladding pumped EDFA. In Proceedings of the 44th European Conference on Optical Communications, Rome,
Italy, 23–27 September 2018. paper Mo3G.2.
43.
Puttnam, B.J.; Sugizaki, R.; Rademacher, G.; Luis, R.S.; Eriksson, T.A.; Klaus, W.; Awaji, Y.; Wada, N.; Maeda, K.; Takasaka, S.
High Data-Rate and Long Distance MCF Transmission With 19-Core C + L band Cladding-Pumped EDFA. J. Lightwave Technol.
2019,38, 123–130. [CrossRef]
44.
Kobayashi, T.; Nakamura, M.; Hamaoka, F.; Shibahara, K.; Mizuno, T.; Sano, A.; Kawakami, H.; Isoda, A.; Nagatani, M.;
Yamazaki, H.; et al
. 1Pb/s (32 SDM/46 WDM/768 Gb/s) C-band Dense SDM Transmission over 205.6-km of Single-mode
Heterogeneous Multi-core Fiber using 96-Gbaud PDM-16QAM Channels. In Proceedings of the Optical Fiber Conference(OFC),
Los Angeles, CA, USA, 22–26 March 2017; pp. 1–3, Th5B.
45.
Jain, S.; Castro, C.; Jung, Y.; Hayes, J.; Sandoghchi, R.; Mizuno, T.; Sasaki, Y.; Amma, Y.; Miyamoto, Y.; Bohn, M.; et al. 32-core
erbium/ytterbium-doped multicore fiber amplifier for next generation space-division multiplexed transmission system. Opt.
Express 2017,25, 32887–32896. [CrossRef]
46.
Takasaka, S.; Maeda, K.; Kawasaki, K.; Yoshioka, K.; Sugisaki, R.; Tsukamoto, M. Cladding Pump Recycling in 7-core EDFA. In
Proceedings of the 44th European Conference on Optical Communications, Rome, Italy, 23–27 September 2018 . paper We1E.5.
47.
Wada, M.; Sakamoto, T.; Yamamoto, T.; Aozasa, S.; Nozoe, S.; Sagae, Y.; Tsujikawa, K.; Nakajima, K. Full C-band Low Mode
Dependent and Flat Gain Amplifier using Cladding Pumped Randomly Coupled 12-core EDF. In Proceedings of the European
Conference on Optical Communication 2017, Gothenburg, Sweden, 17–21 September 2017. Post deadline paper Th.PDP.A.5.
48.
Takeshita, H.; Matsumoto, K.; Yanagimachi, S.; de Gabory, E.L.T. Improvement of the Pump Recycling Ratio of Turbo Cladding Pumped
MC-EDFA with Paired Spatial Pump Combiner and Splitter; OFC: Washington, DC, USA, 2019.
49. Ono, H.; Miyamoto, Y.; Mizuno, T.; Yamada, M. Gain Control in Multi-Core Erbium-Doped Fiber Amplifier With Cladding and
Core Hybrid Pumping. J. Lightwave Technol. 2019,37, 3365–3372. [CrossRef]
50.
Yamada, M.; Ono, H.; Hosokawa, T.; Ichii, K. Gain control in multi-core erbium/ytterbium-doped fiber amplifier with hybrid
pumping. In Proceedings of the OptoElectron. Communications Conference, Niigata, Japan, 3–7 July 2016. Paper WC1-2.
51.
Abedin, K.S. Recent Developments of Multicore Multimode Fiber Amplifiers for SDM Systems. In Proceedings of the 2016 21st
OptoElectronics and Communications Conference (OECC), Niigata, Japan, 3–7 July 2016.
52.
De Gabory, E.l.; Matsumoto, K.; Fujita, S.; Nakamura, S.; Yanagimachi, S.; Abe, J. Transmission of 256Gb/s PM-16QAM Signal
through Hybrid Cladding and Core Pumping Scheme MC-EDFA Controlled for Reduced Power Consumption. In Optical Fiber
Communication Conference, OSA Technical Digest (Online); Optical Society of America: Washington, DC, USA, 2017; paper Th1C.1.
53.
Zhu, J.; Yang, Y.; Zuo, M.; He, Q.; Ge, D.; Chen, Z.; He, Y.; Li, J. Few-Mode Gain-Flattening Filter Using LPFG in Weakly-Coupled
Double-Cladding FMF. J. Lightwave Technol. 2021,39, 4439–4446. [CrossRef]
54.
Sleiffer, V.V.; Jung, Y.; Veljanovski, V.; Van Uden, R.R.; Kuschnerov, M.; Chen, H.H.; Inan, B.B.; Grüner-Nielsen, L.; Sun, Y.;
Richardson, D.; et al. 737 Tb/s (96
×
3
×
256-Gb/s) mode-division-multiplexed DP-16QAM transmission with inline MM-EDFA.
Opt. Express 2012,20, B428–B438. [CrossRef]
55.
Ip, E.; Li, M.-J.; Bennet, K.; Korolev, A.; Koreshkov, K.; Wood, W.; Montero, C.; Linares, J. Experimental characterization of
a ring-profile few-mode erbium-doped fiber amplifier enabling gain equalization. In Proceedings of the 2013 Optical Fiber
Communication Conference and Exposition and the National Fiber Optic Engineers Conference (OFC/NFOEC), Anaheim, CA,
USA, 17–21 March 2013; pp. 1–3. [CrossRef]
56.
Ono, H.; Hosokawa, T.; Ichii, K.; Matsuo, S.; Nasu, H.; Yamada, M. 2-LP mode few-mode fiber amplifier employing ring-core
erbium-doped fiber. Opt. Express 2015,23, 27405–27418. [CrossRef]
57.
Bai, N.; Ip, E.; Luo, Y.; Peng, G.-D.; Wang, T.; Li, G. Experimental Study on Multimode Fiber Amplifier Using Modal Reconfigurable
Pump; OFC/NFOEC: Washington, DC, USA, 2012; pp. 1–3.
58.
Qiu, Q.; Gu, Z.M.; Shi, C.J.; Chen, Y.; Lou, Y.; He, L.; Peng, J.G.; Li, H.Q.; Xing, Y.B.; Chu, Y.G.; et al. Six-mode extended L-band
EDFA with a low differential modal gain. OSA Contin. 2021,4, 1676–1687. [CrossRef]
59.
Fang, Y.; Zeng, Y.; Qin, Y.; Xu, O.; Li, J.; Fu, S. Design of Ring-Core Few-Mode-EDFA With the Enhanced Saturation Input Signal
Power and Low Differential Modal Gain. IEEE Photon J. 2021,13, 1–6. [CrossRef]
60.
Bergano, N.S. Bit error rate measurements of 14,000 km 5 Gbit/s fiber-amplifier transmission system using circulating loop.
Electron. Lett. 1991,27, 1889–1890. [CrossRef]

Telecom 2022,3279
61.
Bergano, N.S. A 9000 km 5 Gb/s and 21,000 km 2.4 Gb/s Feasibility Demonstration of Transoceanic EDFA Systems Using a
Circulating Loop. In Optical Fiber Communication Conference; OSA: Washington, DC, USA, 1991.
62.
Bergano, N.S. Undersea Amplified Lightwave Systems Design. In Optical Fiber Telecommunications, 3rd ed.; OSA: Washington DC,
USA, 1997; Volume A.
63.
Bolshtyansky, M.A. Progress in Submarine Amplifiers. In Optical Fiber Communication Conference; OSA: Washington, DC, USA,
2019.
64.
Sinkin, O.V.; Turukhin, A.V.; Patterson, W.W.; Bolshtyansky, M.A.; Foursa, D.G.; Pilipetskii, A.N. Maximum Optical Power
Efficiency in SDM-Based Optical Communication Systems. IEEE Photon Technol. Lett. 2017,29, 1075–1077. [CrossRef]
65.
Takeshima, K.; Tsuritani, T.; Tsuchida, Y.; Maeda, K.; Saito, T.; Watanabe, K.; Sasa, T.; Imamura, K.; Sugizaki, R.; Igarashi, K.; et al.
51.1-Tbit/s MCF Transmission Over 2520 km Using Cladding-Pumped Seven-Core EDFAs. J. Lightwave Technol.
2015
,34, 761–767.
[CrossRef]
66.
Turukhin, A.; Paskov, M.; Mazurczyk, M.V.; Patterson, W.W.; Batshon, H.G.; Sinkin, O.V.; Bolshtyansky, M.A.; Nyman, B.; Foursa,
D.G.; Pilipetskii, A.N. Demonstration of Potential 130.8 Tb/s Capacity in Power-Efficient SDM Transmission over 12,700 km
Using Hybrid Micro-Assembly Based Amplifier Platform. In Optical Fiber Communication Conference; OSA: Washington, DC, USA,
2019.
67.
Kobayashi, T.; Hiraga, K.; Yamada, M.; Kawakami, H.; Matsuo, S.; Masuda, H.; Takara, H.; Sano, A.; Saitoh, K.;
Takenaga, K.; et al.
2
×
344 Tb/s propagation-direction interleaved transmission over 1500-km MCF enhanced by multicarrier full electric-field
digital back-propagation. In Proceedings of the 39th European Conference and Exhibition on Optical Communication (ECOC
2013), London, UK, 22–26 September 2013; pp. 1–3. [CrossRef]
68.
Ji, P.N.; Aida, R.; Wang, T. Submarine Reconfigurable Optical Add/Drop Multiplexer with Passive Branching Unit. US Patent
2015 0043920 A1, 12 February 2015.
69.
Nooruzzaman, M.; Alloune, N.; Tremblay, C.; Littlewood, P.; Belanger, M.P. Resource Savings in Submarine Networks Using
Agility of Filterless Architectures. IEEE Commun. Lett. 2016,21, 512–515. [CrossRef]
70.
Kovsh, D. Subsea Communications, IEEE, PTC Talk. 2020. Available online: https://comfutures2020.ieee-comfutures.org/wp-
content/uploads/sites/101/2020/02/ComFutures2020-Ses1-SubseaCom-Kovsh.pdf (accessed on 8 February 2022).
71.
Desbruslais, S. Maximizing the Capacity of Ultra-Long haul Submarine Systems. In Proceedings of the 20th European Conference
on Networks and Optical Communications—(NOC), London, UK, 30 June–2 July 2015. [CrossRef]
72.
Turukhin, A.V.; Sinkin, O.V.; Batshon, H.G.; Mazurczyk, M.; Bolshtyansky, M.A.; Foursa, D.G.; Pilipetskii, A.N. High-Capacity
SDM Transmission over Transoceanic Distances (invited). In Optical Fiber Communication Conference; OSA: Washington, DC,
USA, 2018.
73.
Liang, X.; Downie, J.D.; Hurley, J.E. Repeater Power Conversion Efficiency in Submarine Optical Communication Systems. IEEE
Photon J. 2021,13, 1–10. [CrossRef]
74.
Srinivas, H.; Downie, J.D.; Hurley, J.; Liang, X.; Himmelreich, J.; Perin, J.K.; Mello, D.A.A.; Kahn, J.M. Modeling and Experimental
Measurement of Power Efficiency for Power-Limited SDM Submarine Transmission Systems. J. Lightwave Technol.
2021
,39,
2376–2386. [CrossRef]
75.
HMNTECH Website. 18 kV Power Feeding Solution Leads SDM System Revolution. 2021. Available online: https://www.
hmntechnologies.com/enPressReleases/37856.jhtml (accessed on 10 February 2022).
76.
Dar, R.; Winzer, P.J.; Chraplyvy, A.R.; Zsigmond, S.; Huang, K.Y.; Fevrier, H.; Grubb, S. Submarine Cable Cost Reduction Through
Massive SDM. In Proceedings of the 2017 European Conference on Optical Communication (ECOC), Gothenburg, Sweden,
17–21 September 2017; pp. 1–3. [CrossRef]
77.
Dar, R.; Winzer, P.J.; Chraplyvy, A.R.; Zsigmond, S.; Huang, K.-Y.; Fevrier, H.; Grubb, S. Cost-Optimized Submarine Cables Using
Massive Spatial Parallelism. J. Lightwave Technol. 2018,36, 3855–3865. [CrossRef]
78.
Thouras, J.; Pincemin, E.; Amar, D.; Gravey, P.; Morvan, M.; Moulinard, M.-L. Economic Impact of Multicore Erbium-Ytterbium
doped Fiber Amplifier in Long haul Optical Transport Networks. In Proceedings of the 2018 Photonics in Switching and
Computing (PSC), Limassol, Cyprus, 19–21 September 2018; pp. 1–3. [CrossRef]
79.
Asano, Y.; Jinno., M. Cost Comparison of Hierarchical Optical Cross-Connect Architectures for Spatial Channel Networks (SCNs).
In Proceedings of the Asia Communications and Photonics Conference(ACP), Hangzhou, China, 26–29 October 2018.
80.
Downie, J.D. Maximum Capacities in Submarine Cables with Fixed Power Constraint for C-Band, C + L-Band, and Multicore
Fiber Systems. J. Lightwave Technol. 2018,36, 4025–4032. [CrossRef]
81.
Downie, J.D.; Liang, X.; Makovejs, S. On the Potential Application Space of Multicore Fibers in Submarine Cables. In Proceedings
of the 45th European Conference on Optical Communication, Dublin, Ireland, 22–26 September 2019 . paper M.1.D.4.
82.
Bolshtyansky, M.A.; Sinkin, O.V.; Paskov, M.; Hu, Y.; Cantono, M.; Jovanovski, L.; Pilipetskii, A.N.; Mohs, G.; Kamalov, V.;
Vusirikala, V. Single-Mode Fiber SDM Submarine Systems. J. Lightwave Technol. 2019,38, 1296–1304. [CrossRef]
83.
Paximadis, K.; Papapavlou, C. Towards an all-New Submarine Optical Network for the Mediterranean Sea: Trends, Design
and Economics. In Proceedings of the 2021 12th International Conference on Network of the Future (NoF), Coimbra, Portugal,
6–8 October 2021; pp. 1–5. [CrossRef]
84.
Papapavlou, C.; Paximadis, K. ROADM Configuration, Management and Cost Analysis in the Future Era of SDM Networking. In
Proceedings of the 2020 11th International Conference on Network of the Future (NoF), Bordeaux, France, 12–14 October 2020;
pp. 168–175. [CrossRef]

Telecom 2022,3280
85.
Downie, J.D.; Liang, X.; Makovejs, S. Modeling the Techno-Economics of Multicore Optical Fibers in Subsea Transmission Systems.
J. Lightwave Technol. 2021,40, 1569–1578. [CrossRef]
86. Mauldin, A. Submarine Cable and Capacity Pricing Trends in Asia-Pacific; TeleGeography APRICOT: London, UK, 2018.
87.
Mariano, R. Jules Verne and the 20.000 Leagues of Subsea Cables: A True Tale about Submarine Cables. Available online:
https://www.youtube.com/watch?v=JW9bCeknNjo&ab_channel=LACNICRIR (accessed on 7 March 2022).
88.
Papapavlou, C.; Paximadis, K.; Tzimas, G.; Savelona, I. Optical Frequency Hopping Techniques for Secure Fiber-Optic networks.
In Proceedings of the 2021 12th International Conference on Information Intelligence, Systems & Applications (IISA), Chania
Crete, Greece, 12–14 July 2021; pp. 1–4. [CrossRef]
89.
Savva, G.; Manousakis, K.; Rak, J.; Tomkos, I.; Ellinas, G. High-Power Jamming Attack Mitigation Techniques in Spectrally-
Spatially Flexible Optical Networks. IEEE Access 2021,9, 28558–28572. [CrossRef]