Who owns the Internet’s Physical Backbone? Steve Song on Cables, Fibre and Connectivity

Steve Song on submarine cables, fibre networks, resilience, community connectivity and the infrastructure behind the internet.

About this interview: This conversation with Steve Song was recorded remotely on 31 July 2026 and later broadcast on 10 August 2026 at 7:00 p.m. as part of my radio show, Do You Speak Tech?, on Dublin community radio station Near FM.

Patrick Domanico: Steve Song is a Senior Director of Infrastructure Mapping and Development at the Internet Society. Thank you very much, Steve, for joining us today.

Steve Song: Ah, it’s a pleasure to be with [you].

Why submarine cables are critical to the modern internet

Patrick Domanico: Thank you very much. I am very excited about this interview, Steve, because I was looking for some interesting maps about submarine cables, and then I came across ManyPossibilities.net. That is a website that you published, and since 2009, you have been keeping track of both fibre-optic networks and submarine cables [1].

Before we focus specifically on the African Undersea Cable Map, the map you published on the website, could you explain why submarine cables are so important to the modern internet? What role do they play in our everyday digital lives, and what could happen if this infrastructure were seriously disrupted?

Steve Song: Sure. Fibre-optic technology is tremendously important because there is one thing that really makes it different from any other kind of access technology, and that is its capacity.

Fibre-optic networks have orders of magnitude more capacity than any other kind of access technology, which makes them ideal for carrying the backbone of the internet. Of course, they also operate at close to the speed of light, so they are extraordinarily fast.

Once you have access to the internet via a fibre-optic cable, you are only milliseconds away from almost anyone else on the internet. To some degree, that really defines how much of a participant you are on the internet.

While there are other kinds of access technologies that provide individual access, such as LEO constellations (Low Earth Orbit) like Starlink, fibre-optic technology is more of a wholesale access technology. It can enable many operators to provide many services. A single undersea cable can potentially enable dozens of terrestrial fibre-network operators.

So, not only do you bring all this capacity when you bring fibre, but you also bring competition, which brings prices down.

Patrick Domanico: I get it. So, submarine cables play an important role because [they] carry signals or traffic from one point to another. Okay, that’s super cool.

How AI could change demand for global fibre infrastructure

I have to ask this question because, as you see, the rise of artificial intelligence is changing different aspects of our lives and society. Is the growing use of AI significantly increasing international data traffic, or is it placing new demands on submarine cables and infrastructure?

Steve Song: I think the short answer is that nobody really knows exactly how AI is going to reshape demand for global submarine fibre-optic cable infrastructure.

I think there is a general increase in traffic, as we see AI generating more scraper traffic. Also, when you ask an LLM a question, it often goes out and searches multiple sources. So, there is a general increase in traffic associated with AI.

However, you can also see a lot of AI being devolved to the edges of the internet, where local LLMs are actually processing information and providing answers.

So, I think there is going to be an impact, in that there will be an overall upward trend in the use of fibre-optic cables. However, I don’t think it is going to be a dramatic change. I think it is going to be a very manageable increase.

How undersea cables transformed internet access across Africa

Patrick Domanico: Thank you very much. Let’s talk about the African undersea cables that you have been updating since, like, 2009, if I remember well. But what I can see from the website is that we have arrived at version 57, so a long list of updates, I would say. So, very well done with the job you’ve done.

So, you began mapping Africa’s undersea cables at a time when the continent’s international connectivity was much more limited. Looking at the development of the network since then, what has generally changed for African internet users, and which countries or regions are still being left behind?

Steve Song:Well, prior to the arrival of undersea cables, in the early 2000s, there was only one undersea cable, called SAT-3 [2], going down the west coast of Africa, and it was very expensive and very limited.

The rest of the continent depended on satellite connectivity, except for North African countries, which are on the Mediterranean and had closer access to European infrastructure. But for sub-Saharan Africa, it was almost all satellite connectivity, and not the kind of modern LEO constellations, but what are called geostationary satellites, which are much further out in space and take much longer for traffic to go up and back to them.

So, with connectivity that comes through those satellites, you get a lot of latency. And so, you know, if you can perhaps remember the bad old days of making international phone calls, you would get these long delays that made conversations very difficult.

That latency defined all of internet access across African countries at the time. And so, the arrival of undersea cables, I mean, was a huge change, not just because, you know, you got rid of that latency, but also because it became much cheaper to access international traffic.

And, as I said earlier, the arrival of a single undersea cable actually spawned investment in terrestrial fibre-optic infrastructure, which created competition.

Starlink in Africa: access, affordability and infrastructure ownership

Patrick Domanico: Just a very quick, small question, because you mentioned satellite connectivity. What about Starlink at the moment in Africa? Is it well covered? 

I remember you also wrote an article [3] where you recapped a lot of experiments around this. So, for example, the idea behind Starlink is actually old, because someone had already invested in it in the past, like Bill Gates in the early 2000s [with Teledesic] [4]. But maybe the world wasn’t ready about that yet and then we also saw some different experiments like Google with balloon something [Project Loon] [5]. What’s your idea about this?

Steve Song: Um, there is absolutely no question that Starlink is a remarkable achievement, along with SpaceX, in terms of, you know, the reusable launch vehicles taking satellites into space. And it is providing access in places that were previously extraordinarily difficult to provide access to.

Um, and so, it’s hard to deny the kind of miraculous impact that Starlink has had. However, there are other issues to consider when thinking about access.

One is the economic impact. And so, for instance, you know, Starlink is very much a retail constellation. It wants to serve individual households. And where Starlink is being introduced, then typically, if you’re in an unserved community, it is the wealthiest people in those communities who can afford to purchase a Starlink connection.

And those people might otherwise be the loudest voices, you know, complaining that fibre should be coming to their community and serving the entire community. So, there’s a big distinction there between providing wholesale access and retail access.

Now, at the Internet Society, we fund community-centred connectivity initiatives around the world, where we encourage communities to band together, to organise, and solve their own access challenges.

And many of these community networks are using Starlink for backhaul, but they’re using the retail version of Starlink for backhaul, which is comparatively inexpensive, somewhere between $50 and $100 a month, but they’re violating the terms of service by reselling it.

But Starlink is turning a blind eye to this, which is fine for now. But I don’t think we can expect Starlink, or any satellite service provider, to turn a blind eye forever.

And eventually, you know, these community networks will be obliged to purchase a wholesale or business service, which will change the economics of their networks.

So, while I think Starlink is providing, you know, quite miraculous solutions in remote areas, which I think is great, my only concern is that it shouldn’t inhibit investment in terrestrial infrastructure to bring access to entire communities and entire regions.

And I think the other piece that I’m concerned about is the ownership of infrastructure. When communities and countries build their own network infrastructure, it’s theirs, and it’s an investment that can be a generational investment.

When we kind of take a shortcut and just connect Starlink everywhere, it’s great, but the long-term implication is [that] you are renting that infrastructure. You don’t own it, which means you don’t have any leverage over the cost of access, or even know whether it will continue to be provided in the future.

I mean, we saw that when Starlink turned off access in Ukraine for a short period. You are a little bit at the mercy of a non-national organisation for what is, you know, absolutely critical physical infrastructure for the digital economic life of the country.

So, I think, in thinking about Starlink, you know, and bringing Starlink to your country, you have to balance these two issues: addressing urgent access concerns, but also thinking in the long term about ownership of infrastructure, long-term costs and control, and agency at the community level and at the national level.

Patrick Domanico: That’s very interesting. Just one question: what’s the name of the initiative again? Community-centred central?

Steve Song: Community Centered Connectivity [6].

Patrick Domanico: That’s very interesting, and I can understand everything you said. And I think it’s important for a country and [a] community to own the main asset because, as you said, it’s for generations … for future generations, rather than something that can be like a patch for the moment. That’s a very interesting point of view. Thank you very much.

Why landing a submarine cable is only the beginning

Talking again about submarine cables, of course, they can bring enormous capacity to a coastal landing station, and I think this is what you were saying. Satellite connectivity is okay, but we don’t have to forget terrestrial connectivity either.

But when we talk about submarine cables and all the … how can I say … the benefits that they can bring, it doesn’t automatically mean that everybody can benefit from them.

So, what needs to happen between the landing station and the end users, in terms of terrestrial fibre, competition, internet exchange points, regulation, and last-mile infrastructure, before new cable capacity produces lower prices and better services?

Steve Song: That’s a great question. I think that one of the interesting things about undersea cables is that, you know, the old expression from the movie Field of Dreams [7], “If you build it, he will come,” that certainly didn’t apply to people who built websites, but it definitely applies to undersea cables.

So, where undersea cables land, you see this wave of investment in terrestrial infrastructure, sometimes by governments, sometimes by the private sector, [and] sometimes in consortium together.

However, what we have seen across many African countries is that the cost of the terrestrial leg of the infrastructure is often higher than the international cost of connecting to global internet exchange points.

And one of the challenges we face is that there is a lot of terrestrial infrastructure … fibre infrastructure … across African countries, but it is often priced in a way that inhibits the growth of further infrastructure.

And so, there are things you can do to address that. One is to insist on open-access policies for terrestrial fibre-optic cables, which means that everyone should have access on the same terms and at fair prices on terrestrial fibre-optic cables.

And this is possible because the capacity of terrestrial fibre is so great that these are not rival goods at all. You know, they can accommodate an almost unlimited number of operators on the cable.

The second piece is that pricing is non-transparent. So, you find in European countries, for instance, that it’s part of the regulatory framework that they insist that operators publish a rate card, right? These are, you know, their kind of basic rates for wholesale access to their network.

These may not be the prices that purchasers ultimately pay, but it sets a kind of ceiling. So, having regulators insist that especially state-owned infrastructure providers publish a public rate card can begin to benchmark the cost of access across the region and begin to exert a kind of downward pressure on pricing.

And the last piece is the thing that I work on the most, which is also making more information available about the existence of terrestrial infrastructure.

You know, we see public maps of undersea cables … maps from TeleGeography and others, about where undersea cables are being built [and] where they go to. When it comes to terrestrial maps of infrastructure, there’s very little out there. Some operators publish everything, and some publish nothing.

So, what we’re doing at the Internet Society is trying to encourage some norms around information sharing, which we think will increase competition, but also open up opportunities for small operators to build out services into more remote areas.

Because once you’ve connected to a point where there is fibre-optic infrastructure, you know that the rest of your connectivity is going to be high-speed and high-capacity.

And so, there’s an incentive, if you know that there’s a point of presence in a particular region, [because] there may be a business case for building out an access network whether it’s wireless or fibre or what have you to offer services in unserved regions.

From AfTerFibre to the Open Fibre Data Standard

Patrick Domanico: Thank you very much, Steve. I think we can also mention another project you created to map terrestrial fibre-optic infrastructure across Africa. I’m talking about AfTerFibre [8]. What’s that project?

Steve Song: AfTerFibre was my first attempt to map terrestrial fibre infrastructure across the continent.

I had some early success in mapping undersea cables around the continent, and that map became quite popular. It gave me a false sense of confidence that I could easily extend that map to terrestrial networks.

And so, I attempted to take a kind of Wikipedia-style crowdsourcing approach to collecting maps of fibre infrastructure. And I got, you know? I probably succeeded in mapping over 50% of the fibre networks across the continent, but I realised that I was never going to get the other 50%, and that a different approach was needed.

And so, while I think that approach of crowdsourcing data was useful, something more profound had to happen, which is what led to the development of an initiative that we call the Open Fibre Data Standard [9].

It is essentially an open-data, open-standards approach to describing terrestrial fibre. It is basically a common language for talking about fibre.

And, thanks to the Internet Society, we’re supporting regulators and operators around the world to adopt and implement this standard, which means that, ultimately, everyone will be speaking the same language when sharing terrestrial fibre-network infrastructure data.

And that’s important not just for the reasons that I described earlier about increasing competition and improving the quality of investments, but also in terms of improving internet resilience.

So, these days, no operator just has one fibre-optic network. They need redundancy in their network. They need a backup connection so that, if one connection goes down, the internet can route over the redundant connection.

And in order to understand redundancy and how one network fails over to another, we need this transparency of data. We need to have this framework for describing fibre.

So that, you know, when your fibre connection fails, is your backup connection in the same area as that connection? Is it going to fail as well? Or do you have confidence that it’s going through a different route that is actually going to provide you with more confidence that your network will stay up when there’s a disruption?

And, of course, there’s going to be a disruption. Every operator suffers disruptions to their network. And so, having this common language is one way for operators to see that their networks are truly redundant.

Similarly, national governments need this language to be able to understand the overall fabric of their infrastructure and its resiliency.

And, indeed, even at the global level, we see a lot of discussion about undersea cables and the importance of undersea-cable resiliency. But that’s only describing half of the coin. It’s one side of the coin. The other side is this web of terrestrial infrastructure that carries, you know, the bulk of internet traffic.

Who owns the cables? How hyperscalers are reshaping subsea infrastructure

Patrick Domanico: That’s interesting. I was reading something from TeleGeography.com. They recently released some interesting research talking about the level of investment in submarine cables over the next 10 years [10]. So, this probably means that they [submarine cables] are still an interesting asset.

Do you have any thoughts about this? Do you agree or disagree with the projected investment in submarine cables over the next 10 years?

Steve Song: Well, I think the nature of investment is shifting. So, up until, say, five or six years ago, the vast majority of undersea cables were built by telecommunications consortia, right?

So, they were telecommunications operators in countries that shared a common need for international connectivity, and they joined consortia to build infrastructure.

Though what we’ve seen—maybe the last 10 years is a better description—is that now the hyperscale companies, like Google, Amazon, Microsoft, and Meta, see the value of having their own infrastructure and their own subsea cables.

And so, now the investment has tipped away from telcos towards these hyperscale companies that are investing in these undersea cables. And they can do that because, frankly, their reserves of cash are much, much greater than [those of] telecommunications companies.

But the way they go about it is something to pay attention to, and I think this is something that we don’t see enough attention paid to: the ownership structure of these cables.

So, for instance, if you compare the two largest hyperscale cables coming to the African continent, Google’s Equiano cable [11] and Meta 2Africa cable [12], the ownership structures of those cables are very, very different.

So, 2Africa is a true consortium, right? It is [a consortium] where all of the investors in the cable, of which Meta is just one, operate on roughly an equal basis. They own their shares in the cable, and it is, you know, a relatively egalitarian consortium that provides benefits to all actors.

Whereas the Equiano cable is overwhelmingly owned by Google. Although it has landing partners in different countries, it is essentially a Google cable, which is, you know, frankly disempowering to some degree for its partners in terms of control.

So, I think, you know, one of the things that we need to pay attention to, as we see the growth and dominance of these hyperscale companies, is the ownership structures of internet infrastructure at, you know, the undersea-cable level, the data-centre level, and the terrestrial-fibre level.

What real internet resilience requires

Patrick Domanico: That’s interesting, because one of the questions I put down was exactly: how can we determine whether a country has genuine resilience, rather than several systems depending on the same landing points, terrestrial routes, or international providers?

And I think this question is very connected to what you just said.

Steve Song: Yeah, I think we discovered in the last few years how risky it is to have undersea cables that follow a single route.

The disruption off the coast of Côte d’Ivoire (Ivory Coast) in March 2024, which took out five undersea cables [13], was devastating to the West African region. 

And so, you know, cables like Equiano, which took a different route, became absolutely critical in providing backup services and redundancy to the region.

And similarly, with landing stations, you know, previously, geographically separating landing stations wasn’t as big a priority as it is now.

And now, you know, we see the value of making sure that there is substantial geographic separation between landing stations.

Why fibre infrastructure data remains difficult to access

Patrick Domanico: Thank you very much. Steve, I have just a kind of small question before moving to another topic. How hard is it to get the data for the maps you have worked on so far?

Because, I mean, sometimes I consider myself a kind of data practitioner … although I think I offend the category … but I can understand how difficult it is to find the proper dataset to create something else. How hard is it?

Steve Song: It’s much harder than it should be, I will say that.

So, when it comes to undersea cables, there is a lot of good public data from companies like TeleGeography.com. However, when it comes to the nature of the consortia that make up those undersea cables, there is almost no public data.

And so, you can understand where those undersea cables go, but if you want to understand the kind of economics of the companies involved and where the control lies, none of that information is public data.

It’s all kind of, you know, locked behind NDAs between the companies involved. And I think that’s actually problematic, in that there needs to be more public information about the ownership structure of these cables.

When it comes to terrestrial fibre, it’s so interesting. Some operators publish everything. They publish detailed maps, right down to the curbside of fibre. They publish downloadable KML files [14] so you can download the data and integrate it into, you know, a modern GIS system [15]. 

And some will publish absolutely nothing and refuse, even when commanded to [do so] by their regulatory authority. And it is really about norms.

So, you know, for the companies that publish, it’s a marketing strategy for them. They, you know, say: “Look at this amazing network we have. It’s everywhere. This is awesome. You should know more about it.”

But other operators don’t see it that way, and they cite the two biggest concerns. One is cybersecurity: that, you know, “We can’t publish this information because we would, you know, be making it available to threat actors who might disrupt our infrastructure.”

Well, this is contradicted, to some degree, by all those operators who do publish maps of their infrastructure and don’t seem to suffer disruption. So, we welcome evidence that fibre maps are being used to disrupt networks, but it seems very unlikely.

And indeed, you know, the largest fibre operator in Ukraine, a network called RETN, publishes a detailed map of its fibre network, right down to the curbside of its infrastructure.

And one would have thought that, if that were a cybersecurity issue, they would have taken their network map down, you know, years ago when the conflict began. But no, it’s still up to this day.

The other one is competition, and they cite competitive concerns about making their infrastructure [information] available. And I think it’s hard to kind of gainsay that argument.

But you can say that, if everyone shares the same amount of data, you know, using the Open Fibre Data Standard, then that data becomes a tide that raises all ships, and there are no information asymmetries. Everyone benefits from a common resource of information that’s going to lead to smarter investments and fewer network disruptions.

So, you know, the overwhelming [majority of] disruptions to terrestrial fibre-optic infrastructure come from accidents rather than malicious attacks.

You know, there’s a saying Hanlon’s razor “never ascribe to malice that which can be explained by incompetence”. And I think that applies in this context.

And if there were more transparency about terrestrial infrastructure, we would actually see fewer network disruptions rather than more.

Patrick Domanico: I think it’s interesting because, automatically, I can also think about the cost of maintenance. So, it’s okay putting a cable under the sea, but then you have to maintain it.

And I don’t know why, but it connects me to a book that I read. This book is called “The Victorian internet” [17]. I don’t know if you have ever read that book, but I remember it being very funny.

There is an anecdote about how [they tried] to connect one continent to another (Europe with America) and they found this kind of rich, wealthy guy, apparently a bit crazy, because he couldn’t understand the amount of money that it would really have cost to connect [the continents] with a cable.

But it’s very funny because, at the end of the day, we are still connected by a cable.

Village Telco, Wi-Fi mesh and the limits of alternative networks

You’re also known as the founder of Village Telco and as a longstanding advocate for affordable internet access, particularly in underserved communities, as you were already mentioning during this interview.

Would you like to tell us a bit about the experience of Village Telco and, especially, the Wi-Fi mesh technology, which can be another alternative for connecting small communities or areas that are not very well served by, I don’t know, satellites, cables, and so on?

Steve Song: Sure. The Village Telco came out of a certain time in the early 2000s in South Africa, where it was impossible to get access to any wireless spectrum except for Wi-Fi spectrum.

But the demand in South Africa at that time was more for voice services than for the internet, especially in remote areas. People wanted, you know—voice was, as they say, the killer app.

And that was before affordable mobile infrastructure had spread substantially into rural areas. So, Village Telco was really an attempt to address that issue by combining voice telephony with wireless mesh devices.

And so, we manufactured a low-cost wireless mesh technology that you could also plug an ordinary telephone into.

I think mesh networks have kind of, you know, caught the zeitgeist in terms of this kind of alternative infrastructure that could provide access anywhere. And I think it has been oversold, in that mesh networks are a tactical choice, not a strategic choice.

You know, Wi-Fi is a strategic choice, right? You might build Wi-Fi infrastructure, and in some cases you might use a mesh, but in other cases you might use a Star Topology [18] or Point-to-point (P2P) [19]. And so, it is really on a case-by-case basis that you would choose.

I think there was some wild optimism about mesh because each device is a repeater for the next, [and] somehow you could build out this massive web of mesh devices that would provide you with this kind of blanket of access.

But, you know, the devices don’t really work like that, in that they are not infinite repeaters. The service is diminished with every hop on the mesh, and so, you know, the performance of the mesh degrades quite quickly as you expand it.

Which is why, as I say, it is useful in some contexts, and it is important, certainly. I mean, Wi-Fi technology has been miraculous in terms of providing alternative infrastructure in general.

I mean, what we’ve seen in African countries is that, as an alternative to mobile network operators, which obviously use different access technologies and devices, the competition that has emerged through internet service providers using unlicensed spectrum and Wi-Fi devices has been hugely important.

Especially, you know, during the pandemic, when families were forced to self-isolate and we needed access. It was really important to be able to be connected to the internet without, you know, going out and about.

Wi-Fi providers offering fixed wireless access were transformative. They exploded during the pandemic across most African countries, providing competitive, unlimited access, in contrast to mobile network operators, which were mostly kind of, you know, pay-per-use or pay-per-gigabyte services.

Patrick Domanico: I think it’s interesting, but that “wild optimism about mesh,” do you think it has disappeared, or does it still make sense to invest a bit in it, or to use mesh as an alternative?

Because, for example, I know that here in Ireland there is a community of people who are trying to provide this kind of alternative internet for freedom, or something like that.

Do you see something interesting in this, or do you think, as you said, that there are technologies that work better?

Steve Song: I think, when it comes to access, you have to be agnostic about the technology choice, right? You have to approach it from a kind of toolbox perspective.

And in that toolbox, you’ve got, you know, wireless point-to-point networks, you’ve got mesh networks, and you have fibre-optic technologies. You want to look at each case and choose the most appropriate and affordable technology.

I think that what we see happening increasingly with community networks is that they’re embracing and building their own fibre networks, which, I mean, brings interesting opportunities and challenges.

On the one hand, fibre networks are kind of easier to deploy than wireless networks. With wireless networks, you have to know a lot about radio spectrum and propagation and kind of, you know, how your signal is being impeded by trees or by other buildings. It can be complex to build those networks.

Whereas, sometimes, you know, simply running fibre between two places can be simpler and more reliable and, ultimately, have more capacity.

That’s not always the case. And so, really, you just have to kind of look at the situation and choose the best technology.

And certainly, in some cases, mesh technologies are the best choice for access. But, I mean, there is no one true answer.

Community networks and the case for locally owned infrastructure

Patrick Domanico: Well, you’re right. Let’s talk about your current role at the Internet Society as Senior Director of Infrastructure Mapping and Development.

What are your main activities, and what are the main challenges that you face in your role?

Steve Song: Well, I work on two things at the Internet Society.

One is this Open Fibre Data Standard and trying to bring transparency to the bottom layer of the internet. You know, the internet is often characterised as having seven layers [20] [21], and this is the bottom layer, which I think hasn’t received enough attention.

We often think about, you know, the cloud, and people talk about putting their data in the cloud, but they don’t realise where it’s going and what’s happening.

And so, I think bringing more transparency to the physical layer of the internet is actually critically important in terms of understanding how the internet is going to evolve over time, and the choices that communities and countries are making—especially from the point of view, as I mentioned earlier, of this issue of infrastructure ownership.

So, that occupies, you know, quite a lot of my time at the Internet Society.

The other thing I work on is policy and regulation related to community-centred connectivity. You know, I don’t think it receives the attention it deserves, but some of the fastest networks in the world are being built through community efforts.

In northern England, in Lancaster, there’s an initiative called Broadband for the Rural North (B4RN) [22] and they are offering multi-gigabit speeds to farmhouses in an area that BT, British Telecom, once declared uneconomic to serve.

And so, this is a very, very exciting story that is happening, you know, all around the world, where communities are choosing to build, manage, and own their own infrastructure.

However, many countries have barriers to doing that at the policy and regulatory level, right? So, many countries have regulatory frameworks that are oriented towards large network operators, like your [Vodafones?] and so on.

And so, part of what we do at the Internet Society is work with regulatory agencies to look at how they could expand their regulatory frameworks to embrace small operators and locally owned infrastructure.

And that affects things like licensing, access to spectrum, access to funding, and capacity building.

Patrick Domanico: I just have a question that connects with one of my previous questions, when I was asking about how hard it is to find data.

You talked about your experience as, let’s say, a “normal person” [I was referring to the manypossibilities.net private project] so, let’s use that word. You know, I have a very poor vocabulary, sorry about that, and I recognise it.

But even now that you’re covering this role, is it still easy or difficult to map terrestrial fibre infrastructure?

Now, I suppose you have access to some information that, as a citizen, you didn’t have. Or, even in this case, working for the Internet Society, is it still hard to find the data needed to map infrastructure?

Steve Song: Yes, it’s still challenging. It’s still much harder than it should be. It’s partly because of this lack of a common language, right? So, operators might submit maps in PDFs, or in, you know, JPEG images, or in Excel spreadsheets. And so, it becomes very challenging simply to, you know, use the same terms to describe these networks.

So, they might give an individual picture, but it’s very hard to build up a comprehensive picture.

This lack of a common language is absolutely fundamental. But it’s also [about] the practice of collecting this data.

You’d be surprised at how many places in the world [where] a rigorous practice of collecting this infrastructure data on the part of the regulator is just not happening yet. It hasn’t been organised.

So, there’s still a long way to go.

Patrick Domanico: That’s funny, because I thought it was more about, I don’t know, policies. But no, it’s simply a matter of standards.

Let’s talk about the Open Fibre Data Standards which aims to create a common language that operators, governments, and researchers can use to describe and exchange information about terrestrial fibre infrastructure.

So, what information should infrastructure owners be expected to publish, and how can greater transparency be balanced with legitimate security concerns and commercial confidentiality?

Steve Song: Uh well, you’ve touched on the hardest question, right?

Patrick Domanico: Okay, just at the end I invite you for the next episode 🙂

Steve Song: No, no, it’s fine. We try not to be prescriptive about what data is shared.

With the Open Fibre Data Standard, we try to be as comprehensive as possible in the descriptive language we have. But we believe that exactly what is shared between operators and regulators is a matter of negotiation and discussion.

Operators and regulators can find a level of comfort where they’re sharing information at a level that meets the requirements of the regulator but also, you know, doesn’t compromise the concerns of the operators. And there’s kind of a path to meet in between.

There’s also [a] conversation about what data should be public, and that is also a complicated conversation in terms of the things we discussed earlier about competition and risk.

It’s mostly about kind of working in an iterative way. So, you share a little bit, and that builds trust and confidence. And once you’ve shared a little bit, then maybe a little bit more can be shared.

So, it’s not [about] kind of putting everything at risk, but beginning a kind of conversation and iterative process where information is shared in a way that benefits everyone without putting anyone particularly at risk.

What better fibre data could mean for ordinary users

Patrick Domanico: Let’s play this kind of game. If reliable and standardised infrastructure data becomes available, how can it produce practical improvements for ordinary users like me, for example?

Steve Song: I think, well, you might want to know where, or how close, you are to fibre-optic infrastructure if you are buying a house, right?

You know, if there were no fibre-optic networks anywhere near you, that might affect your decision about whether to purchase a property, if access to the internet was important to you.

Access to this data can tell you something about, you know, how countries are serving the poorest areas. You know, is infrastructure only being built to serve the comparatively wealthy?

And if you don’t have access to this kind of data, you can’t have that conversation about whether the country is doing enough to build infrastructure to serve the most vulnerable and economically poor populations.

Similarly, you might want to understand the millions of dollars that are being spent on fibre-optic infrastructure. How is it changing people’s lives? Are there more jobs where there is fibre-optic infrastructure? Are wages higher when you’re closer to fibre-optic infrastructure?

I think there is a dearth of research into the impact of investments in this infrastructure, and having this kind of information will actually deepen our understanding of how our lives are being changed by this—by, you know, internet infrastructure.

Patrick Domanico: And, at the same time, I think operators can see which areas are not served and maybe invest, or, you know, once you have the map…

Steve Song: That’s good.

Patrick Domanico: Yeah. Well, that’s cool.

Patrick Domanico: That’s cool, Steve. I have a lot of … many millions of … questions, but unfortunately the radio show is only 58 minutes long. And so, I have to invite you again, if you enjoyed this conversation, because I think it’s very interesting, as is your knowledge in general.

Where can people find you online?

Steve Song: Yyou can find me through the Internet Society, and you can just reach out to media{[@]}isoc.org and I’m very happy to chat with anyone.

Patrick Domanico: Okay. I can also invite people to go to your website, ManyPossibilities.net, because I think there is still a lot of interesting content, even if [it is] a bit outdated.

Thank you very much. You are more than welcome to come back anytime, Steve. And thank you once again.

Steve Song: Thanks for the time. I really enjoyed the conversation.

References

NB: Sources below were added during publication to provide context and supporting material for topics discussed in the interview. They were not part of the original conversation.

1. “African Undersea Cables,”Many Possibilities, 16 October 2008. Assessed 06 September 2026
2. “SAT-3/WASC,” Submarine Networks, n.d. Accessed 1 September 2026.
3. Steve Song, “Starlink and Inequality,” Many Possibilities, 6 November 2023. Accessed 15 August 2026.
4. “Teledesic,” Wikipedia, last edited 24 May 2026. Accessed 6 September 2026.
5. “Loon,” X, the Moonshot Factory, n.d. Accessed 4 September 2026.
6. João Paulo de Vasconcelos Aguiar, “What Is Community-Centered Connectivity and Why Should We Care?”, Internet Society, 17 July 2025. Accessed 6 September 2026.
7. “Field of Dreams,” Wikipedia, last edited 5 September 2026. Accessed 6 September 2026.
8. “AfTerFibre,” Many Possibilities, n.d. Accessed 6 September 2026.
9. “Open Fibre Data Standard,” Open Fibre Data Standard, n.d. Accessed 6 September 2026.
10. Alan Mauldin, “Forecasting the Next 10 Years of Submarine Cable Investment and Kilometers,” TeleGeography, 7 July 2026. Accessed 6 September 2026.
11. Michael D. Francois and Chris George, “Introducing Equiano, a Subsea Cable from Portugal to South Africa,” Google Cloud Blog, 28 June 2019. Accessed 2 September 2026.
12. “Announcing the Completion of the Core 2Africa System: Building the Future of Connectivity Together,” Engineering at Meta, 17 November 2025. Accessed 1 September 2026.
13. Staff Writer, “WA Regulators Call For Cooperation To Protect Submarine Cables,” CIO Africa, 15 July 2026. Accessed 6 September 2026.
14. “KML Tutorial,” Google for Developers, last updated 3 November 2023. Accessed 15 August 2026.
15. “Geographic information system,” Wikipedia, last edited 4 September 2026. Accessed 6 September 2026.
16. “Hanlon’s razor,” Wikipedia, last edited 21 August 2026. Accessed 6 September 2026.
17. Tom Standage, “The Victorian Internet: The Remarkable Story of the Telegraph and the Nineteenth Century’s On-Line Pioneers,” New York: Walker and Company, 1998. Accessed via Internet Archive, 20 August 2026.
18. “Star network,” Wikipedia, last edited 20 February 2026. Accessed 6 September 2026.
19. “Point-to-point (telecommunications),” Wikipedia, last edited 11 May 2026. Accessed 6 September 2026.
20. International Organization for Standardization, “ISO/IEC 7498-1:1994– Information technology – Open Systems Interconnection – Basic Reference Model: The Basic Model,” ISO, November 1994. Accessed 6 September 2026.
21. “OSI model,”Wikipedia, last edited 29 August 2026. Accessed 6 September 2026.
22. Broadband for the Rural North Limited, “B4RN,” B4RN, n.d. Accessed 6 September 2026.