Subsea Cables & Internet Infrastructure

First Point: The Portugal and Spanish grids are tightly integrated with limited power connector to the rest of Europe. Solar and wind play a big role and both power sources suffer from what is known as frequency instability. Solar and wind generated power is much more volatile than traditional power sources. Traditional power generators have angular momentum inertia. It takes a while to up or lower the power due to the inertia in the spinning components. A natural gas turbine takes a few minutes to spin up. A nuclear reactor an hour to lower or increase output by 10% (French reactors do load following). Solar and wind create very volatile power fluctuations that can easily trigger a circuit breaker. In an isolated grid if a circuit breaker is triggered, the power in the remaining active part of the grid increases. This triggers more circuit breakers and usually brings down the entire grid. 2. The consequence of the first point is that avoiding a repeat of the Iberian Penisula outage re...

Top 10 ASN Ranking! Not Cogent. Not Hurricane.

`This article reflects discussions I have had with war planners, government officials, policy analysts, and subsea cable colleagues.  1. A striking fact is that there is no compelling evidence of subsea cable outages due to sabotage since the end of WWII. Subsea cables are poor terrorism targets. Terrorists create terror by killing and maiming people and damaging highly visible and important infrastructure like bridges, skyscrapers, prominent buildings or sites having symbolic importance. Intentional damage of a thin cable buried two meters deep in the English Channel does not have the shock value or cause sufficient disruption by itself to justify the great effort of clandestinely locating and severing it. Secondly, there are so many cables now that sabotage of one or two has little impact on voice or data traffic. RIPE analysis indicated that a country like Estonia experienced little layer 3 degradation despite losing subsea cables landing in the country or adjacent Finland, a r...

 4. The star-shaped organization model defines each department by function such as sales, provisioning, procurement, network operations and the like and has it directly report to the COO or CEO. Each department has a large degree of autonomy and independence which encourages kingdom building and departmental bickering.  We improve cooperation by recognizing that accountability requires some departments be subordinate to others. For example, procurement falls into two categories: backbone and the customer driven requests where a third party component like a long haul circuit, metro dark fibre pair or a local loop is required. Third party sourcing for customer driven requests should in the sales department. I know salesmen at one European carrier that source their own off-net requirements because procurement takes to 2 to 6 weeks to do it. Once a verbal yes is received, the salesman then points procurement to the third party offer. If sales determined the compensation and perfor...

Retelit is a competitive Italian carrier with a pan-European network. Retelit together with the Italian Ministry of Economy and Finance have purchased Sparkle with a 30/70 ownership split. Why the current Italian government keeps insisting on public ownership of telecommunications infrastructure is beyond me. There are no real benefits. Italian government interference in important sectors like banking and now telecom have brought only problems, no solutions.  Government ownership often leads to management sacrificing long term financial health to goals such as preserving head count. This purchase together with the recent EXA acquisition of Aquacomm for only $45 million despite a couple hundred million dollars in network investments raises alarm bells about the financial health of the subsea cable industry. Sparkle generated a billion Euros in fiscal 2023 or 1.14 billion dollars at the current exchange rate. Yet it was purchased for far less than current revenues, a sig...

I've been in telecom since 1992. This qualifies me as an 'old fart' or 'dinosaur fossil' as American teenagers would say. This means I've seen every mistake made by subsea cable capacity owners. 1. Buying lots of capacity between cable landing stations, but owning no fibre from the CLS to the customers' destinations, namely the popular carrier neutral data centers. You can't be competitive if you must buy 100G or 400G metro waves from a UK landing station to Slough Equinix. Lease a dark fibre pair ring and light it with DWDM. Don't be penny wise and pound foolish. Those network investments will dramatically improve operating margins. The amazing thing is that PPT members of cable consortiums make this mistake all the time. If it is worth spending $55 million for an undersea fibre pair, then it is worth adding a couple million to the pot for back haul IRUs. 2. Refusing to extend the network to new locations unless the order achieves an investment thre...

When you come to Europe, you go to Amsterdam. It's inevitable. Amsterdam is one of Europe's two main Internet hub, the other being Frankfurt. Also the city is a major logistics hub that is a good hunting ground for commercial telecom deals. 1. AM5 is one of the best data centers for Internet peering, but has no power for new clients or upgrades for existing ones. Novation is your best bet.  2. AM3, Nikhef, and AMS17 do have power.  3. Nikhef is the best deal for smaller players because there are no recurring cross connects fees, power is available, and fractional racks have no install charges. Peering opportunities are excellent with both AMS-IX & NL-IX present. Excellent site for both private ISPs and financial trading firms.  4. NorthC's two Amsterdam facilities have plentiful space to lease. Not a lot of connectivity providers yet, but could prove highly attractive to the server intensive crowd which needs a lot of racks and power.  5. Relined is highly reco...

The conflict between China and the West is exacerbating bandwidth shortages on key routes like Marseille to Singapore. AAE1 is maxed out just like SWM5. Both will be upgraded this year, but I believe all the incremental capacity will be snatched up even before upgrades are finished and the capacity delivered to customers. Furthermore, China Unicom is the lead AAE1 consortium member with China Mobile also selling capacity on the system. Avoiding carriers incorporated in China makes intercontinental capacity shopping is an excruciating exercise. I've been seeking Express AAE1 100G for almost a year for clients for whom China is a red line. Bandwidth sourcing has become a marathon. 😄 In light of this, I recommend buyers consider Peace despite the fact that it is a Chinese financed project. Encryption does work. It will not protect the IP overhead, which include the IP addresses, but the data payload itself will remain safe. Moreover, there are Peace providers such as PCCW or TELIN t...

6. Inability to fully exploit and integrate acquisitions. Lumen bought denial-of-service cybersecurity firm Black Lotus in 2015, but the company's full potential has not been realized. Lumen uses Black Lotus for its highly rated IP network, but it has not really fully developed commercial services for this acquisition. 7. Collapse of transit pricing. Cogent and Hurricane Electric has caused pricing to collapse. I just heard of a deal where a client got a 10G commit on a 100G port for $500 a month plus bursting charges at an Amsterdam data center. 8. Inability to control network expenses. The graph below shows gross margins declining sharply over the last 15 years. A lot of that reflects transit and wavelength price declines. But it also reflects the company's arthritic bureaucratic culture: revenue per employee is $509,000 versus over a million dollars for Verizon. Market pricing is often out of a company's control. In the competitive telecom industry there is only one way ...

Lumen continues its restructuring. It aims to sell its consumer fiber division to AT&T for a rumored $5.5 billion dollars. It sold its European and South American assets as well as its Atlantic cables over the last 24 months. In 2021 Lumen solds it local exchange assets, the old Centurylink copper line business, for $7.5 billion. Lumen is paying a high interest rate on its recently renegotiated debt of $17.5 billion. It is under severe pressure to continue paying down debt while also investing in network infrastructure. Lumen once had aspirations of being a global telecom player. It built South American and European networks from scratch during the New Millenium telecom frenzy. The carrier acquired a host of distressed competitors including Global Crossing, Progress Telecom, and US CLECs. In turn, Centurylink bought Lumen in a friendly merger in 2017. The idea was to infuse capital into the so-called sexy fibre optic business. The underlying idea was that Centurylink had a lot of c...

Half terabit of AAE1 Express available now. Both Marseille/Singapore and Frankfurt/Singapore via Bari are options.  A couple terabits of Topaz Tokyo/Westin Building in a few weeks. Topaz is the first cable to directly link Japan to Canada. It is also the second lowest latency cable for the Tokyo to Seattle route. Telstra's Topaz pricing is high. Here's a competitive alternative. Go! Go! Go!

What most people don't realize is that the following cables have not threaded the Red Sea yet due to the threat of Houthi missile attacks: 1. 2Africa. 2. Africa-1. 3. Blue Raman. 4. IEX. 5. SWM6. It is likely that none of these cables go live this year (2Africa Marseille/Mombasa segment among others ). When I ask industry insiders whether the Red Sea segments for these cables are finished, I get the roar of silence. The impending capacity drought reflects the failure of subsea cable designers and senior carrier management to take seriously their over-reliance on the Red Sea. Even Blue-Raman, which bypasses Egypt, traverses the Red Sea before landing near the Jordanian Aqaba data center. It is time for the telecom industry to work with Saudi Arabia to build routes that completely bypass the Red Sea and hit water on Israel's coast. Even it means neglecting Djibouti. Carrier culture is not innovative. Never has been. Technology comes from the network vendors like Ciena or Infinera...

Ooredoo operates mobile operations in ten countries with 2024 annual revenues of $6.5 billion. It has mobile subsidiaries in Algeria, Indonesia, Iraq, Kuwait, Maldives, Oman, Palestine, and Qatar. The carrier reminds me of Vodafone, which entered the subsea cable market to cost effectively transport its international traffic. Led by Rick Perry, head of subsea network development, Vodafone owns capacity on many cables, and in fact, is a lead landing partner for 2Africa. Similarly, Ooredoo is entering the subsea market. It is the 2Africa landing partner for Oman, which must be extremely annoying for Omantel. Ooredoo is hosting a new CLS in Barka for 2Africa and also one in Salalah. Due to 2Africa's strict rules limiting predatory cross connect pricing, I expect Oman to become a leading Middle Eastern telecom hub. Right now Dubai is a telecom hell hole due to the mandatory cross connect fees that the two incumbents, Etisalat and Du, receive. It creates artificially high bandwidth pric...

AAE1 is the lowest latency path for the all-important Marseille to Singapore route. Also the lowest latency for Frankfurt via Bari, Italy to Singapore. Buy now to avoid your boss beating you with a big stick! 😃 

University of Eindhoven in the Netherlands has set up a 4.6 kilometer free space optical link that traverses the city center. Its throughput is 5.7 Tbps. Unfortunately, I don't see signal to noise measurements. Free space lasers suffer from atmospheric turbulence such as rain, fog, and clouds. They work well in outer space. The technology is the basis of the new NASA deep space communications network. Click here for the short paper: https://arxiv.org/pdf/2503.21058v1. 

In general, the American Digital Giants care a lot about protecting their customers from third party eavesdropping because it would undermine customer confidence in their services. Apple has refused on occasion to cooperate with the FBI when the latter has tried to get access to encrypted Iphone messaging or Icloud storage in criminal cases. Similarly, Google introduced Gmail encryption to protect its end users privacy from third party sniffing.  I have heard repeatedly through the telecom grapevine that the OTTs were not happy with the British Snooper Law. This uncontrolled mass surveillance provided another strong reason to build subsea cables that bypass the UK in order to directly deliver Continental Europe destined traffic. As a result, we have Marea, Dunant, Amitié, and now Anjana connecting Virginia Beach, Boston, and Myrtle Beach, South Carolina to France & Spain. Let me quote from a UK human rights group: "The UK has the most intrusive mass surveillance regime of any ...

At the beginning of the third millennium almost all Trans-Atlantic cables directly linked the US or Canada to either Ireland or the UK. Flag North landed at Northport, New York, and Bude, UK. Hibernia North (EXA North) was a rare exception. It came ashore at Halifax and also in Southport, UK. Yellow touched ground on Long Island and at Bude. AC2 came ashore also at Brookhaven and Bude. Hibernia South (EXA South) landed in Halifax and also at Dublin. Finally, Apollo North touched ground at Brookhaven and Bude. Clearly this is inadequate physical diversity in terms of landing points and cable landing stations. It is highly cost effective because subsea cables can share common terrestrial infrastructure, but cost and resiliency are sworn enemies. There is almost always a tradeoff. Almost all Trans-Atlantic traffic went to the 9th floor Telx facility at 60 Hudson and the 111 8th Avenue data centre in Manhattan. In London the destination was almost invariably the Telehouse North and East da...

Network resiliency defined as up time entails higher build and operating costs. Resiliency in the subsea cable world reflects two basic principles: good construction practice and physical diversity. Good construction includes an undersea route that minimizes damage and time to repair. In practice this means avoiding areas where there are geophysical threats. These threats include ships, debris slides, earthquakes, and strong undersea currents that erode the protecting shielding of deep sea cables. Good practices include deep burial, undersea repeater redundancy (the number of  spare amplifiers in an undersea repeater), cable armor thickness, etc. Physical diversity means putting a big distance end-to-end between the subsea network and other submarine cables. The farther apart, the less likely a common event disrupts two or more cables. In most cases this means longer undersea paths that increase the construction bill as well as planning costs. Good examples include  the Aprico...

A total of 9 cables land on the beaches near the small town of Bude, UK. There are four operational cable landing stations serving them in the Bude, UK area: two Vodafone CLS, a Colt (former Lumen) CLS, and a BT facility. Please click on https://lnkd.in/gGAP3QMA for a plethora of photos of the cable landing stations. The map illustrates the tendency for telecommunications networks to lack adequate physical diversity to ensure resiliency. Sometimes a laissez faire regulation is not the right approach. Most back haul fibre from the cable landing stations to London probably traverses the single road parallel to the beaches. See below.  When I worked at Hibernia Atlantic as an exclusive sales contractor, we cited the concentration of cables at Highbridge and Bude as good reasons to purchase capacity on the Hibernia North & South cables. North lands several hundred kilometers above Cornwall and at Halifax on the North American side. It was a compelling sales ptich. These cables toda...

AAE1; 100G; Frankfurt/Singapore; 139 ms RTD; $25K MRC; 2 Years; No NRC. Peace; 100G; Marseille/Singapore; 135 ms RTD; $20K MRC; 2 Years, $0 NRC. Peace; 100G; Mombasa/Singapore; $38K MRC; 3 Years ADC; 100G; Tokyo DC/SG1-SG3; 66 ms RTD; $13,500 MRC; 1 Year; $0 NRC.

Colt has bought capacity on Vodafone's Apollo South cable, one of the first Trans-Atlantic cables to directly connect France to the United States (along with Flag Atlantic 1). Apollo was RFS 2003. I was surprised Colt would buy capacity on a 8 terabit cable that has already 22 years of service under its belt. Apollo is a low capacity system facing similar operating expenses to higher capacity systems, hence its pricing should be generally much higher. In general, Colt appears to be mostly interested in serving the low latency end of the financial markets, mostly market markers (providing a bid and ask for a financial asset) and financial traders. Hence the most plausible explanation is that Apollo South provides a low latency route connecting Paris and Frankfurt to the NYSE and NASDAQ data centers in New Jersey as well as the other financial markets like BATS located at NY4, Secaucus Equinix. Another possible angle might be that the cable is not only low latency, but also highly di...

Turkey has bee a very tough market for global ISPs. Until EXA lit terrestrial fibre a few years ago across the Bulgaria's border with Turkey, 100G pricing was $15K per month and above. EXA now dominates the key Frankfurt to Istanbul route with lower pricing. But the situation is not ideal. Physical diversity is an illusion across the narrow border with one conduit system carrying much of the international traffic. Sparkle's old Mednautilus system does provide an alternative to the land routes, but it's primacy has been too high for foreign ISPs in Turkey to flourish. Sparkle has signed a memorandum of understanding with Turkcell to build a subsea cable that will land in the Instanbul suburd of Izmir and at Chania, Greece. The yet unnamed cable will interconnect at Chania, a Greek island, to the BlueMed cable which will carry the traffic to Marseille, Milano, and other key telecom hubs. We have few details on the cable itself other than a target design throughp...

A point: MRS1; Z point: SG1; Express Route; $31.5K MRC; 1 Year. Target RFS Date : Mid of June, 2025

 1. APG Cable; 100G; Hong Kong/Singapore; $8500 MRC; 1 Year. 2. ADC; 100G; Tokyo/Singapore; $13,850 MRC; 1 Year. 3. ASE; 100G; Tokyo/Singapore; $18,000 MRC; 1 Year. 4. ASE; 100G; HK/Tokyo; $8,500 MRC; 1 Year. 5. ASE; 100G; HK/Singapore; $5500 MRC; 1 Year Remarks: Customers responsible for cross connects. Pricing is the same whether from CLS to CLS or carrier neutral POP to carrier neutral POP.