Subsea Cables & Internet Infrastructure

Despite what we think, telecom wholesale service is not really a tech industry, at least not in leadership. Engineers, scientists, technicians, mathematicians, and computer scientists lead the real tech companies. Google is an excellent example. Sergey Brin and Steve Paige, two Stanford educated computer scientists, founded the company. German companies almost always select mechanical engineers to lead them. Obviously there will be cases of generalists who do quite well. Bjarni Thorvardarson was not a STEM graduate, but he has a critical, yet open mind and led Hibernia out of bankruptcy to a $610 million sale sale to GTT. He accepted a sales force initially composed of four sales contractors who had no salaries, but received 10% of net revenue, (gross revenue minus third party network costs), used a contract manager to handle standard NSA and MSA negotiations as opposed to a lawyer, and took the daring step of building the new, ultra-low latency Express cable in an era o...

EXA has deployed G.652d fibre across much of its European backbone with the important routes receiving priority. Chart below shows the paths that have been upgraded. EXA's western European network consists largely of the Interoute assets plus a Bulgarian company that had fibre and conduit into Turkey. It is a highly dense network that gives EXA's Atlantic cables and Trans-Atlantic capacity (it owns a fibre pair on Dunant as well as some capacity on Amitié and other cables) the ability to deliver traffic from Chicago and Equinix all the way to Instanbul. This dense footprint allows to deliver traffic at lower costs then its main competitor, Telxius. The carrier has the same basic reasons to upgrade as EUNetworks, namely fibre depletion, hyperscaler demand for long haul dark fibre, and the improved performance of modern fibre with less bending and linear attenuation. Its marketing emphasizes that it is deploying G.652d fibre, but I believe this is just marketing to reassure pot...

With most European long haul fibre having been deployed in the 1998-2002 periods, it is not surprising that after 20 years that carrier upgrades were in the cards. There are several reasons. One is fibre exhaustion. Secondly, although fibre itself does not rapidly depreciate because glass is highly stable, technology has evolved. For one, the optimal ILA spacing has sharply increased due to the advent of ultra-low loss fibre, which has attenuation of .2 dB or even less per kilometer. Since most repeater huts are old, replacing closely spaced old ILAs with new, more widely spaced facilities lowers power and equipment costs long term. With fewer ILAs, less can go wrong. There were other factors as well driving EUNetworks' upgrade decision. Although old fibre is generally in remarkably good shape, there are often problems at joints. At these locations, the fibre pairs must be removed from the cable for splicing. Moisture often penetrates these joint enclosures. It attacks and erodes t...

During most of my telecom career terrestrial long haul fibre was synonymous with the ITU's G652 standard. This is single mode optical fibre that has zero chromatic dispersion at the 1310 nanometer wavelength and good wavelength performance in the 1260 to 1625 nanometer range. This include the O, E, S and workhorse C bands. G.652d fibre quickly became the de facto standard for long haul terrestrial fibre optic networks. The G.652 standard was introduced in 1984. Its motivation was a dual purpose optical fibre suitable for data centers and long haul. Most optical equipment for internal data center traffic transmitted in the O band, which ranges from 1260 to 1390 nanometers. But long haul traffic faces significant attenuation as distances grow. This required repeaters, another source of capex and opex expense. It was already known that attenuation or the diminishing of light's intensity was minimized in glass at 1550 nanometers. The versatility of the G.652 standard made it hugel...

I An optical fibre has three structural components. The glass core transmits the light. It is the optical highway that the light travels. If the light is single mode, the core is 8.3 microns. A single mode fibre can be thought of as a single lane road. If the light is multi-mode, the core is 50 or 62.5 microns.  The 125 micron cladding is also glass. It is made ofglass that keeps the core glass from leaking light as long as the angle of incidence is not too big. The buffer consists of a hard and soft layer of plastic, often urethane acrylate. The soft layer surrounds the cladding and cushions it. The hard layer prevents abrasion. For a more complete explanation, click on https://learn.aflglobal.com/home/the-basic-structure-of-optical-fiber. II Modes of Optical Strands A single mode strand has one path. Think of it as a straight path down which the light travels. To keep the path straight the diameter is relatively small at 8.3 microns. Multimode glass allows the glass to trave...

Juno is the Queen of the Olympian gods in Roman Mythology. This 20 fibre pair cable has design capacity of 320 Tbps, which makes it the highest capacity subsea network to connect the US to Asia. It lands at Grover Beach, California, and also has Japanese landings at Shima and Minami-Boso. NTT is the project's main backer and owner. The carrier clearly felt that two Japanese branches along with three cable landing stations (Softbank Maruyama, NTT Shima, and NTT Minami-Boso) would significantly improve resiliency given Japan's reputation for earthquakes and other natural disasters such as tsunamis. As is usual for a NTT project, it created a standalone company known as SerenJuno in which it has majority ownership to minimize risk. This way any misfortunes would not result in the parent company having financial liabilities. NTT is also extremely reserved about revealing capacity owners on the system. I have learned that PCCW is on the system. Details Ownership: NTT and Mitsui. Cap...

 1. Damage to cables is often the result of a cascade of events. For example, the 2006 Taiwan Earthquake pushed coastal sediments down a steep slope and triggered a massive landslide that traversed over a hundred kilometers. In many cases snapping deep sea cables like they were toothpicks. The sediment build up is due to Taiwan's heavy rains that carry eroded soil to the sea. 2. Lots of subsea faults arise from events that unfold over years such as deep sea abrasion of the outer sheaths of a cable. A cable may rub against a rough surface on the ocean floor for years until the polyethylene layer and the wound steel strands are removed and water hits the copper power conductor. 3. Another interesting example is the Congo Canyon where the Congo river deposits sediments offshore that high spring tides or flooding push down the undersea canyon and crush subsea cables. 4. Volcanoes during discharges have buried cables in molten lava. Also pose a danger to cable landing stations and terre...

In a recent interview META subsea engineers suggested their next project would be a 1 Petabit Trans-Atlantic cable. They pointed to three possible ways of accomplishing it: multicore fibre, use of the C and L spectrum bands, and 50 fibre pair SDM. Pros and Cons 1. Multicore is subject to cross talk. Right now the only multicore cable is the Taiwan-Phillipines-US cable scheduled to go live this year. A Google project with NEC as the vendor, it has two cores per fibre strand and 26 Tbps total per fibre strand or 13 Tbps per core. The challenge is cross talk. The light spills from one core to the other and vice versa. Obviously this distorts the signals and at high transmission rates the distortions become greater and greater and the signal-to-noise ratio disintegrates. The advantage of multicore fibre is that no redesigning of the optical amplifiers is required. Since everyone acknowledges that SDM alone is unlikely to surpass the half terabit mark, Google is obviously interested in mult...

As noted in my previous post, optical amplification allow light signals to be boosted without being first decoded into a digital representation. Optical-electrical-optical conversions go away. Hence the computer necessary for OEO conversions disappears. This in turn sharply improves the amplifier's reliability and life span. It also eliminates the conversion delay so end-to-end latency is improved. Finally, no computer means less cost. But these benefits are really secondary. More importantly, the advent of optical amplifiers led to a quantum leap in bandwidth that can be attributed to two related developments. The first is that optical amplifiers impose no transmission limits on computer technology. This means that we can lay a cable in the water and then upgrade it at regular intervals as Moore's law improves the ability of computers to process optical signals. Nothing on the wet side changes. Indeed, the introduction of digital processing allowed 10G wave subsea cables like ...

 A 1997 article celebrating how new optical amplifier technology will raise subsea cable output to an amazing five gigabits per second. It was a revolution at the time because the first optical fibre cable, TAT8, operated at 240 megabits. 😆

Layer 2 Switched Ethernet From Tokyo Data Centers To AWS Cloud 1. 10G; Layer 2 Pseudo Line Emulation Ethernet ; $775 MRC. 2. 100G; Layer 2 Pseudo Line Emulation Ethernet; $3800 MRC. Remarks: A. 802.1q frames. B. Jumbo framing available. C. See RFC 3985 to understand pseudo line emulation, which is basically a private line Ethernet service defined by a constant bit rate. D. Standard MTU size of 1522 bytes.

Picture below shows subsea amplfiers aboard a cable ship. They cost roughly $250K a piece, can support up to 40 Tbps, and are spaced every 70 kilometers or 42 US miles. These remarks apply to these specific amplifiers. As noted in Part 1 of this series, erbium doped fibre eliminated optical-electrical-optical (OEO) conversions for amplifiers. So the limiting factor of computer hardware involved in the conversions was eliminated. The result was a 'transparent' technology that placed no limits on submarine termination equipment upgrades. Subsea cables became highly scable because the amplifiers could easily handle higher throughput achieved via DWDM improvements. These amplifiers are tough buggers. They must withstand severe pressure at depths as great 8,000 meters for 25 years as well as nasty seawater corrosion. The amplifier hull is generally made of titanium or a special beryllium copper alloy, C17200, which conducts heat and electricity very well and is highly resistant to ...

Canada and US account for slightly over 45% of the world's IVP4 IP addresses. Europe is far less. Yet the 2025 budget for ARIN, the regional registry that serves the US, Canada, and Caribbean, is $28.8 million versus 41 million Euros for RIPE. If I use purchasing price parity to convert Euros to USD, and this is the correct approach, then the RIPE budget is 1.3*41 million Euros or $53 million USD. So RIPE's budget is almost twice the ARIN budget, yet their core functions are the same. The budgets are recouped via fees on IP addresses. It is plausible in my opinion that RIPE is feather bedding, which means it has a lot of unnecessary positions. There have been rumors for years that RIPE maintain a huge, but generally useless staff of community representatives. Let me add, that I have used ARIN to get /24s for an ISP client. The organization is highly efficient and professional. So is Arin underfunded and RIPE just the right size? Highly unlikely. Given the price pressures that I...

***It is worth emphasizing that the DRC 2Africa CLS is a carrier neutral facility where all clients are treated equally with non-discriminatory and cost-based pricing and consortium-imposed performance standards. Moreover, there are several diverse fibre routes from the CLS to Kinshasa, which has several new carrier neutral data centers. The DRC market could be particularly interesting as it traditionally been a hell hole for telecom carriers. This means wholesale ISPs can sell at high transit prices. ***The other Congo (Brazzaville) is also on-net for 2Africa. And I can get you fibre transport across the river at wholesale prices to the DRC. 🙂 ***Special Note: ACR2, a Digital Realty data center in Accra, Ghana, is where most 2Africa capacity carriers keep their SLTEs.

TAT-8 was the first Atlantic optical cable. It was RFS in 1988 and signified a bandwidth revolution with its capacity 10x that of its coaxial predecessors. By modern day standards it was a pygmy with 280 megabits total throughput. But it heralded the beginning of a long period of rapid throughput growth. Optical repeaters were spaced every 50 kilometers or 30 US miles. Photo diodes received the weak incoming signals, converted them into a digital representation of zeros and ones, and then laser diodes generated fresh light signals. A copper conductor provided power. Lots of redundancy in terms of components were built into this amplifiers to avoid repairs. However, this approach had a severe drawback, namely the throughput could not be increased in a time where Moore's law was rapidly increasing transmission rates. The pace of the optical-electrical-optical conversion was set in stone because an upgrade would require replacing all the hardware on all the amplifiers. TAT-8's ca...

***Senegal is the one of the most challenging on-net 2Africa countries. 1. Sonatel appears to have a monopoly on Dakar metro fibre. This spells trouble. 2. Sonatel's landline infrastructure monopoly may be one reason many 2Africa consortium members declined capacity into Senegal, despite its strategic location as the natural Northwest Africa telecom hub. Another reason might be their lack of mobile operations in that country as Senegal is a small market. 3. ONIX data center is the place to be in Senegal. Carrier neutral and on-net for 2Africa. Most African ISPs coming to me for advice plan to make ONIX their home. 4. I expect 100G pricing to be high due to the dearth of capacity. Figure low 30s for 100Gs to Europe and upper 20s for 100Gs to West African countries. ***Ghana is much friendlier. Many 2Africa carriers are on-net at the CLS and also the very popular PAIX data center. In fact, I know one that has built a ring from the CLS to PAIX reflecting strong demand for the latter. ...

Facebook engineers during an interview on the Waterworth project provided details about their next big project. 1. Facebook is planning to build the first one petabit per second Atlantic cable. No details yet available on end points. Given traffic flows it is most likely to directly connect Continental Europe and the US. 2. META engineering is looking at three options to achieve one petabit per second throughput. A. Using both the C and L spectrum. This would effectively double the bandwidth. B. Cable will definitely be SDM (spatial division multiplexing). Strong likelihood that will be 48 pairs. C. Another possibility is two core optical fibre in order to double the bandwidth per pair. 3. I believe the most likely option is using C and L bandwidth. Arelion has incorporated L band spectrum into its DWDM Layer 1 service between Atlanta data centers and Ashburn Equinix using Infinera gear. Colt and Sparkle have used the L band on terrestrial routes. Most DWDM equipment today offers both ...

1. The Waterworth cable will link the US to South Africa with a branching unit to Brazil. It extends from South Africa to India & onto to the Pacific and the US. The total length is greater than the Earth's circumference at approximately 50,000 kilometers. 2. Subsea cable projects are taking 3 to 5 years from initial idea to commercial service. 3. US traffic goes to Europe. It's aggregated with European originated traffic and then traverses the Mediterranean Sea to reach Egypt and takes terrestrial routes (Telecom Egypt) to the Red Sea. Then the traffic flows down the Red Sea. From there it either heads to India or bypasses it with Southeast Asia being the destination. 4. In the eyes of Facebook's subsea engineering team, the standard cable routing described above creates a host of problems. First of all , the Red Sea is a single point of failure. Same holds for Egypt. Secondly, the Mediterranean Sea requires many government permits as cables inevitably goes through ter...