The Living Network - Our Future

In this final episode of the series, we join a round-table discussion among experts from the Smart Internet Lab at the University of Bristol. Join Dimitra Simeonidou, Simon Saunders, and Paul Wilson as they explore the future of telecommunications and the exciting advancements in network technology.

 Discover how AI and sensing capabilities are transforming networks into intelligent systems that not only communicate but also 'feel' the environment. This episode delves into the implications of these innovations for smart cities, crisis management, and the ethical considerations that come with them.

Chapters:
(00:00) Introduction
(00:30) The Evolution of Telecommunications
(10:15) AI and Sensing in Networks
(20:00) Future Networks Scenarios
(30:00) Ethical Considerations and Governance
(40:00) Closing Thoughts

This is an 18Sixty Production.

Transcript

>> Host: Welcome back to the Living Network from the Smart Internet Lab at the University of Bristol. In this final episode of the series, I'm going to hand over to the experts. So relax, sit back and join Dimitra Simeonidou, Director of the Smart Internet Lab, Simon Saunders, Honorary professor at the Smart Internet Lab and Paul Wilson, Chair of the Smart Cities World Advisory Board as they come together in conversation.

>> Prof. Dimitra Simeonidou: I'm going to talk about myself first. I have been in the telco sector all my life and I do identify myself as a telco enthusiast and that's where my heart is and my passion is. And telecoms, they have been through a very long time now perceived quite a boring industry to be in. Yes, really a utility that provides things that nobody cares about and everybody's building capabilities on top of what telecom infrastructure is. But at the moment there are a couple of things that they are happening that make me extremely excited of what the future is going to look like. And these two things, they're associated with what we call the next generation of mobile networks or 6G which not.

>> Prof. Dimitra Simeonidou: Many people want to talk about.

>> Paul Wilson: We got a bit bored with 5G.

>> Prof. Dimitra Simeonidou: Yeah.

>> Paul Wilson: So we're 6G and satellites.

>> Prof. Dimitra Simeonidou: Yes. So the thing is that let's take those as a general trends of networks going forward.

>> Simon Saunders: Can't we just talk about networks of the future? Because there's going to be a whole mix of 6Gs and 5Gs and Wi Fi's and satellites.

>> Prof. Dimitra Simeonidou: Let's talk about future networks. so those two things just to stop the discussion The one is of course the embedded intelligence into our networks. People are calling it AI native networks. But the thing is that we see actually adoption of AI into network infrastructure very very fast, whatever that means. And the second thing is actually what we call join communications and sensing that means that our networks do not only communicate, providing these services for people and things to talk to each other, but the networks actually feel. Yes. Feel the environment, feel themselves. Yes. So they are providing a sensing service.

>> Simon Saunders: It's like the very opposite of a dumb pipe.

>> Prof. Dimitra Simeonidou: Yes, it's exactly the opposite of it. And you know for somebody that has been work on fiber networks for a very long time, we knew that opportunity for a very long time because you take this fiber, it's a very long sensor and you lay this fiber in the ground and, and can detect for instance any kind of movements, vehicles running in the city and you put this fiber into the ocean floor and actually you can detect things like for instance any, and predict earthquakes and I.

>> Simon Saunders: Guess we should be clear. This is kind of a bug that turns out to be a feature. These fibers weren't designed to sense things. Right. It just turns out they're sensitive in interesting ways.

>> Prof. Dimitra Simeonidou: You have photons going inside the fibers and you know, the polarization of the photon, for instance, change through vibration. So if you are able to measure that change in polarization, you can.

>> Paul Wilson: Do you know how's that accuracy?

>> Prof. Dimitra Simeonidou: Very good accuracy. Because now they're developing special systems that they are actually detecting these changes, you know, in a meter accuracy. So within a meter of fiber, if you think that you may have 6,000 km for instance, of a system, a meter is quite good.

>> Paul Wilson: So under the sea, I can actually know if a boat is going by. You do know a large whale.

>> Prof. Dimitra Simeonidou: Yes.

>> Simon Saunders: Or a submarine.

>> Prof. Dimitra Simeonidou: Or a submarine.

>> Paul Wilson: Even down to an octopus.

>> Prof. Dimitra Simeonidou: Even down to an octopus.

>> Paul Wilson: Really?

>> Prof. Dimitra Simeonidou: Yes. But it's not only. So with fibers, we knew that this is happening, that this can happen. what is happening now is people start looking at the wireless side. For instance, your WI fi already can detect movements around, indoors environment like this. But furthermore, 5G systems, 5G advanced systems as we go forward, 6G systems. I said it again. But the thing is that there is now going through vendors and standards how we can actually use these RF signals that we use already for communications to provide us sensing information.

>> Simon Saunders: Arguably this is an update of actually a very old story. So radar was invented here in the UK based on an observation that radio communication signals bounce back off aircraft. And we're bringing that up to date now. It's almost like we'd forgotten that that property was there with radio signals in the spectrum, in the air, as it were. And now we can observe that we've got much more capable communication signals and that gives us better, resolution in terms of space. So we can get down to meters, perhaps even centimeters in that environment, and time resolution, because these are very wide bandwidth signals. So we can also detect distances and, motion and the nature, shape and materials of objects.

>> Paul Wilson: But people got really excited about Internet of Things, didn't they?

>> Prof. Dimitra Simeonidou: So, but this is not Internet of things.

>> Paul Wilson: There's all sorts of sensors out and about.

>> Prof. Dimitra Simeonidou: This is not about sensors. So what we say now, you don't need cameras to see if somebody comes into your door because your 5G or Wi Fi is going to detect it. You don't need IoT this extrapolation. We're extrapolating now, you don't need IoT platforms because the network itself, your 5G network that you receive, your Service detects what is happening.

>> Paul Wilson: I don't quite get this. I will say, like you, you've said the door can open. I don't need a camera to know the doors opened because I can tell it on my network, but the network doesn't know my. The network might know the door open, but it can't see me, can it?

>> Prof. Dimitra Simeonidou: But it could construct. You can construct imaging out of, you know, detection, signaling. But the thing is that you can get the same alert in your phone that somebody actually is on your door.

>> Simon Saunders: Might sound slightly abstract, but some of this is we're moving from electronics back to photons and photonics, by which I mean there's both. What Dimitri is describing about the sensitivity of fibers which are carrying photons to vibration and movement and so on, but also radio waves, we often forget, as radio engineers are also composed of photons. They're just photons that are in free space. Actually, when you think about it, a camera is just a radio receiver. So we have actually built cameras every time we've built a WI FI access.

>> Prof. Dimitra Simeonidou: Point or a don. It's a device that you add into your network here. Your network itself is. Detective.

>> Simon Saunders: Yeah, no, that's what I'm saying. The camera I'm talking about was already the access point that was doing the communication.

>> Prof. Dimitra Simeonidou: So what is happening? Quite a lot of the conversation that is happening up to now, people are looking at special technologies to enable that sensing.

>> Prof. Dimitra Simeonidou: Yes.

>> Prof. Dimitra Simeonidou: With meta surfaces and reconfigurable sensors and, devices and all these things that they're coming out of the labs. And actually vendors start also working out solutions. But the technology is not that important for me. What is important is if you start stitching this environment so you have your Wi Fi, you have your 5Gs, you have your fiber. You can see that. You scale this sensing to any kind of scales, potentially to a global scale, because you. You have all your ages, where people interact and receive services. You have your smart city, your smart home. Yes. And then everything is pipelined down into a fiber. And then that fiber go from regional to national to international. And actually the only thing that connects us globally is that fiber infrastructure. So we can have the data out.

>> Prof. Dimitra Simeonidou: Of sensing all the time, always on.

>> Prof. Dimitra Simeonidou: And Simon knows about always on environments which we can leverage. And with the AI on top, we can start actually developing very exciting solutions on the positive side or very worrying scenarios on the negative side. But I can see. What I can see is that these networks, let's avoid the word nervous systems, but these actually global networks, they could.

>> Prof. Dimitra Simeonidou: Be actually Global conscience. They can develop global conscience. Yes.

>> Paul Wilson: So Dimitri, you added in AI all of a sudden. I think where my mind went was the network is sensing all the time.

>> Prof. Dimitra Simeonidou: Yes.

>> Paul Wilson: What hasn't been done yet particularly well is harvesting all of that sensing into intelligence.

>> Prof. Dimitra Simeonidou: That is data.

>> Prof. Dimitra Simeonidou: Yes.

>> Paul Wilson: And that's data. And then AI allows us to now to accelerate this super fast.

>> Prof. Dimitra Simeonidou: I mean there is a very big question how we can use AI to extract knowledge out of this data. How this knowledge is going to look.

>> Prof. Dimitra Simeonidou: Like, is it going to be centralized or completely distributed. But you can see the opportunity now for me it's a very important time in telco infrastructures because we can see the industry potentially become something else and become something else that actually the connectivity service is completely hidden. But it's all that intelligence out of sensing that could come out of this and the opportunity that it creates.

>> Prof. Dimitra Simeonidou: And that is really quite,

>> Prof. Dimitra Simeonidou: I mean there is a lot to think about, there's a lot to argue about.

>> Simon Saunders: There is, I mean stepping back, I mean I find it an instantly compelling vision for a couple of reasons. One, you know, it's no longer the network as simply a carrier of words and pictures and texts and, and videos. It does all that. But now it becomes a very active part of the story that it's delivering sensing data that is its understanding of the world. And it's assimilating that data through the intelligence. That's an intrinsic part of it. And I think that creates an opportunity that says we're not just saying we can join the dots, but we can actually have this be a part of the world. Yeah, I mean I think we also though have to recognize that when we're sitting here talking about things that don't look like cameras but act as cameras and harvesting data on people and things all around the world. I can imagine a lot of people feeling a little freaked out by that. that, that sounds like something that could be used irresponsibly on a very large scale. So I guess, I guess we should explore where that take.

>> Paul Wilson: My mind has gone to the fact that you've got say you did this, which is coming, you know you need to do something next, don't you with this information. And your autonomous systems are ah, the perfect thing to do something because they are responding to the data in, let's call autonomous systems, robots, drones or various things in the real world. Actuators that make things happen in the real world. So now I can see from your network you start to the next step on is stuff starts happening in the Real world, physical things start happening in the real world with what we've called robots, drones, autonomous systems. So this is quite important, isn't it? We put some guardrails around how this happens.

>> Prof. Dimitra Simeonidou: Guardrails are absolutely important. But think about situations like crisis management. Yes. And usually when actually there is a conflict in an area or there's a disaster in an area, the first thing that people are looking is we lost connectivity. But now we have alternatives because we can have satellite that potentially could be unaffected. So immediately you bring your satellite, your satellite connects, gives Internet connectivity. But what you care more about is what the satellite sensing in the location that you are having a crisis could give you information where exactly you need to intervene. And so you can actually mobilize help with real time sensing data. not because people are going to be calling.

>> Paul Wilson: Do you think that's already today?

>> Prof. Dimitra Simeonidou: No, because now we are relying on somebody calls critical services. This is an autonomous thing. So you are sensing where things are happening and how they are happening. Because if you have casualties, huge amount of casualties in a location, it may take very long time to get this information in.

>> Simon Saunders: One of the things that enables that is very much that this isn't a central controller with a bunch of arms out to places. This is something that's very distributed and therefore very, very resilient. So bits of it can break, they can be removed and still you can have that functionality. I mean I think there's some great analogies for that. Right? In the natural world.

>> Paul Wilson: Yeah, I'm well in the natural world. Oh gosh, you're thinking of octopuses, aren't you? Where octopuses have nine brains, they have a brain in each tentacle. so they're very close to like edge computing in a sense, to sensing their environment around them. Their ninth brain, the one in the head, is coordinating a federating management system. And I think the useful thing for us to bring that into this topic is to say as humans we are we have one brain. So we're very it worked for us. We don't know anything different from that. All of us are organized with a central command and control system. So we tend to design then systems and solutions in our own image, as do governments and corporations. However, nature gives us some other models like the octopus, a distributed mind where it's closer to the edge, it's closer to what's happening in its environment. And what's more, some astonishing things that humans can't even believe almost that the octopus can change color and shape to fit its environment. Now for me personally, I find this very exciting because rather than this becoming some Orwellian command and control dark thing that's going to take over the world run by the tech bros, which I think a lot of people are pretty worried about, quite honestly. That for me is frightening, honestly that we see the biggest tech giants now running everything. What we need to do is to ground this in, I would say local democracy for me, and in our everyday lives, which are cities where we live and work and you know, do our thing. So for me in this discussion, the governance of what comes next is almost the critical issue. Let's assume the tech will come because it seems to be coming, next. We need to think long and hard about who's got the data. To what effect is this data. And I'm not sure we're really ready, with those governance of data and what to do with IT systems. So data trusts are often talked about but are few in number really. We're still grappling with basic legal agreements with sharing data. It seems really frustrating that we can't get to get grips with that quickly.

>> Prof. Dimitra Simeonidou: It's a very, very good analogy with Octopus and the tech pros because what I worry about, and there is the last few weeks we have seen it just renew interest from some very big AI, vendors on networking. And I think that one of the reasons is that new emerging capabilities that we can see in networks. So if you want footprint beyond the big data centers, you have to have the distribution scale that networks provide. And that is a perfect platform for actually placing AI, resources and being able to extract intelligence from your local environments. What we don't want to see, and I agree with you Paul, what we don't want to see is actually having somebody controlling all these networks and actually create armies of robots that they are coordinated and actually, you know, obliged to commands from the big tech companies. But this can happen. It's not far fetched because I can see, I can see that this actually is a, very real possibility.

>> Simon Saunders: It is. I mean one of the things I think, I think helps with some safeguards in that comes from another natural analogy. So when I'm not playing with telecoms networks, I spend a lot of time in forests. And you know, we used to think that individual trees just suck in nutrients and water and grow as best they can, fighting for space and light and aren't too interested in what's happening to the other trees around them. But in fact in recent years we've realized that they often depend In a very kind of tight symbiosis on fungi that are growing right intimately around their roots and up their trunks. And they're interconnected as mycorrhizal fungi, with the networks of other trees. And that isn't just the fungi doing its thing. It actually acts as a signaling network between the trees. And sometimes the trees then collaborate between themselves. Other times they actually compete. So each tree, if you like, has a part of the puzzle. Just like a network would have the ability to sense certain parts of the environment. And they're sort of locked in this interesting dance where they're each choosing whether to share or to hoard. And that, that sort of builds in. Weird kind of coopetition, to use a lousy word, which in some ways I don't want to claim this is everything, but, in some ways I think that creates some kind of intrinsic checks and balances on the data that there isn't a single controlling mind that has access to everything. Rather, there's a whole ecosystem of elemental networks which are local in cities and in rural places and in countries, sometimes competing, sometimes cooperating. And that becomes a good, you know, that you have the opportunity to trade in the useful data and, you know, to all wake up when there's a disaster and deal with that in a very efficient fashion.

>> Paul Wilson: Simon, I'm fascinated by. You call it the wood wide web, I think, don't you, that the, is it the mycelium? Is that what they call it? The mycelium, the fungus? I read some of these can go for hundreds of miles. The single living organism of mycelium under the soil and the vast networks underground that are, ah, some of the oldest life forms on our planet.

>> Simon Saunders: That's right. There's about a mile in every cubic centimeter of soil.

>> Paul Wilson: That's astonishing.

>> Simon Saunders: It is astonishing, isn't it? Yeah.

>> Paul Wilson: I don't even know if I even can comprehend what that means, actually.

>> Simon Saunders: It's incredibly coiled and compact. It gives you a sense of the vast scale and I mean obviously that's the fungi just doing its own thing because it needs to get to those nutrients. It's often interacting with the tree and as I say, very symbiotic fashion. So it needs to find its way everywhere.

>> Paul Wilson: It's almost the ultimate definition of resilience, isn't it? We've got almost one of the longest living things that we know of in the universe because our planet's the only place we know of life.

>> Prof. Dimitra Simeonidou: Is it a peer to peer example?

>> Simon Saunders: Yeah, I think it is a peer to peer Example, there isn't a single kind of controlling oversight in this. I mean, that said, you know, they're all affected by common things. So if there's a lot of shade in the tree canopy, then they all are affected by that consistently. There's a kind of coherence, across that, you know, the sun rises and falls for all of them. So it's not that they're operating in a separate bubble with completely different conditions. There's another piece of that. Oak trees, for example, when they drop their acorns, they do it very non uniformly year by year. And there are certain years that are so called mast years when there's a ridiculously large number of acorns and weirdly that happens in the same year all across Europe, even when the climate is quite different in those places and we still don't know why and how that happens, that all the oak trees across the whole of Europe have a mast year together.

>> Paul Wilson: Wow. Gosh. Never knew that. I think this biomimicry is fascinating in this topic because we need to get into this because technically we're getting to the point where we can see networks start to become sensing environments. This is so important for many things. Right. It's a, it's a revolution.

>> Prof. Dimitra Simeonidou: Biological living things.

>> Simon Saunders: Yes.

>> Paul Wilson: We need to next think about life itself and biomimicry and how these things work in tandem and well, rather before one starts destroying the other, our modern technical environment starts further destroying our natural environment.

>> Simon Saunders: That's right. I think there's a lot for us to learn from the natural environment and to help the natural environment. And I think we should bring to the table that we're not just talking about this theoretically, we're actually putting it into practice. so for example, at the Smart Internet Lab, we're the centre of a UK wide initiative called the Joiner Network where we've realised as researchers we've spent a lot of time playing with little bits of networks, building little toy simulations and claiming great things for it. And it doesn't have a lot of credibility to it because networks are big things, they're very multidimensional, they have scale in space, across countries and continents, they have scale in terms of the number of nodes and components that they have and they have scale in terms of the compute resources, they have scale in terms of the numbers M of calls and texts and all the rest of it that they handle. So for us as researchers and as technologists, to be credible in anything we do with this, we have to work at that scale. That's Kind of expensive from most research grants. So what we decided to do, and we're in the process of building, is build a network that's at commercial scale, has an intensity of compute and storage resources that is actually beyond today's networks, is physically distributed all across the country and potentially into other countries. We've got it all across the nations of the uk, but also now into Ireland. And we're using that to breadboard all of the techniques that we're talking about here, right? From applying AI to make the network work at its best, potential performance, through to adding sensors to that. So one of the parts of that that we're working on is, is sensing the spectrum around all of those locations, which generates absolutely vast data sets, many terabytes to spot patterns and understand how that interaction happens. And we can now start to dig into the data we've collected and look for what it tells us about what's moving around those things. Not only what it tells us about how networks are using the spectrum, but also how the spectrum is revealing the natural world around us.

>> Prof. Dimitra Simeonidou: I think that when we started putting together Joiner as a consortium and being very passionate on how we can support network innovation, telecom innovation in the uk, we didn't think about these aspects and the whole opportunity about linking AI into the network with environmental sensing, it has been evolving the last two years as we are building Joiner, but we are finding ourselves now in situation that I think we have a responsibility to use this platform, to actually start understanding what it means for the future that sensing over the network at that scale and what it means to have a conscious network, what it means to have a living network and actually reflect on that. Not from a tech point of view, but we are actually landing in 15 cities in the UK. 15 cities. And we can link this back to applications, you know, and citizen experience and we can start making some sense. What are the issues you mentioned data trusts, Paul? Yes. What are the issues around privacy? What are the issues about responsibility, especially across responsibility and AI models that we are using and we are going to be using. And from my point of view, I think that we should accelerate this activity because the learning is very important and informing the community beyond our telco community, what are the opportunities and the risks of, are also very important. So I would very much like to work with your cities colleagues and get, you know, just, just reflect some of these opportunities.

>> Paul Wilson: It seems very important to be doing that right now because the technical leap is coming super fast and also the multiple crises facing humanity are coming super fast and with climate coming super fast. So this seems very prescient to be going deep into this together. Very exciting.

>> Host: Thank you to Dimitra Simeonidou, Simon Saunders and Paul Wilson. If you would like to find out more about the research happening at the Smart Internet Lab, then head to bristol.ac.uk research research groups smart. You've been listening to the Living Network from the University of Bristol's Smart Internet Lab. This was an 18Sixty production. The producer was Kate White.