Subsea Cables & Internet Infrastructure

META's chief AI scientist, Yann LeCun, explains why neither Google nor META has deployed something like ChatGPT despite developing many of the underlying techniques. "I mean you have to realize that most of the underlying techniques used in ChatGPT were invented at Google and META. If you are META or Google, you can think about putting out a system like this that you know is going to spew nonsense and you know because you are a large company you have lot to lose by people making fun of you for that." Click for the complete interview : https://www.youtube.com/shorts/bdYylvhZ9T0

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...

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...

This past autumn I did a post based on two insider conversations about an under-the-radar Facebook cable that would span the world with a W shape. I was told the cable would directly dconnect the US to South Africa and then head straight to India and onward to Australia before landing on the West Coast of the US. My theory at the time was that this was an AI driven project since the routing really didn't match Internet traffic flows or did not connect to major Internet exchange points (like Singapore) even though the route passes by them. The purported route latencies would be quite high which discourages carrier interest in purchasing capacity on the system.  So I figured its purpose was to move 'Big Data'. I am surprised to say I was right.  The one deviation from my initial understanding is that the 24 fibre pair cable will land in Brazil before veering for South Africa. A Brazil landing makes perfect sense in retrospect because Facebook's current capacity down to So...

Facebook is building a six fibre pair cable directly connecting Taiwan to the the United States with landings at Toucheng, Taiwan, Hermosa Beach, California, and Manchester, California. RFS is 2Q2027. The Hermosa Beach CLS is the well known facility built by RTI Holdings before its bankruptcy. In the submarine cable landing license application, Facebook noted that its motivation was the fact that US-Taiwanese traffic is growing rapidly each year. Due to the 12,000 kilometer length of the cable and the absence of any island landings for power, the design throughput per pair is a relatively low 12.8 Tbps or 76.8 Tbps aggregate initial capacity.  ORCA is an open cable so each fibre pair owner will operate and control its own submarine line terminating equipment with only power being under collective control. Since Facebook is the cable's sole initial owner, the open architecture suggests it will sell capacity on the system to third parties to recoup capital expenditures and share comm...

Several sources have whispered in my ear that META is planning a new 16 fibre pair cable that will encompass the world going from the US East Coast to the US West Coast via the Atlantic, Indian Ocean, and the Pacific. The most ambitious subsea project ever undertaken. I do not know the exact routing. I know that the cable will launch from the American East Coast and will go down the West African Coast to South Africa and then head straight to Mumbai. It is not clear if Europe will be online or not. From Mumbai it will head straight to Australia and then up to the US West Coast. I speculate that there may be branching units to Singapore, Malaysia, and Indonesia. But the basic routing is US/South Africa/India/Australia/US. What is not clear is if there will be branching units to add more countries to the cable.  This semi-secret cable reflects META's desire for network resiliency given the four month Red Sea down time that AAE-1 and other subsea cables suffered during the first half ...

Meta's Anjana cable sets an Atlantic bandwidth record with 24 fibre pairs each transmitting 20 terabits per second for a total of nearly a half petabit. This huge capacity reflects its spatial division multiplexing (SDM) design. Traditional subsea cable designs maximize capacity per pair, but this leads to nonlinear signal distortions once a threshold is exceeded. Hence additional power to boost the rate above the threshold yields only small gains. This inefficient use of power limits fiber pair counts to 8 or less per cable with most cables rarely exceeding 150 Tbps. In contast, SDM increases the total bandwidth punch by operating each pair at lower transmission rates to avoid these nonlinear signal distortions. In turn, the lower transmission rates free up power to support enough more pairs to sharply raise total cable throughput, which ranges from 200 Tbps to a half petabit per second.  Another notable feature of Anjana is that it lands at Myrtle Beach, South Carolina, and Santa...