Is The Internet Reaching Its Limit?

Jul 15, 2013
Originally published on July 12, 2013 2:45 pm

There are more than 600 million websites on the Internet, according to the tech site Pingdom.

That includes about 18 billion Tumblr blogs, 36 billion photos uploaded to Facebook each year, an average of 175 million tweets a day and 144 billion emails sent a day.

And the internet keeps growing. But is there limit to the internet’s size? Markus Hoffman, head of Bell Labs Research in Holmdel, New Jersey, says yes — and we’re approaching it.


Copyright 2018 NPR. To see more, visit


It's HERE AND NOW - and Meghna, here are some big numbers. There are more than 600 million websites on the Internet.



HOBSON: That includes about 18 billion Tumblr blogs, 36 billion photos uploaded to Facebook each year, an average of 175 million tweets sent every day, and 144 billion emails a day.


HOBSON: I know. That's all according to the tech site Pingdom. Well, our next guest says all that stuff is a problem, and that we're actually reaching the limit of our Internet infrastructure. He is Markus Hofmann, who heads Bell Labs Research in Homedale, N.J. Markus, welcome.

MARKUS HOFMANN: Thanks, Jeremy. Thanks for having me.

HOBSON: Well, so first of all, what makes you think that the Internet is reaching its limit?

HOFMANN: Well, it is based on the personal experience we all have, right? We use all these applications on the Internet like browsing photos, browsing videos, doing indirect to video chats. And it works well most of the times. But every now and then, the video freezes; it takes forever to load those photos. For example, when I transmit a live video cast of my kids' hockey game to my parents in Europe, just in the most exciting moment, it might happen that the video screen freezes, right?

And this will increase over time, as the amount of data we transmit over the Internet today - this data will explode, and it will become more and more; and therefore, we need to continue increasing the capacity and the bandwidth of the Internet.

HOBSON: Well, but I mean, there are going to be hiccups along the way when you're using technology like this. What makes you think this is something that requires our attention now - and it's not just like, sometimes things are slower and sometimes, they go OK?

HOFMANN: So just like, for example, there are limits to the speed of light, there are limits to the amount of information we can transmit over a single communication channel, and we are getting closer and closer to that limit.

HOBSON: Well, what do we need to do?

HOFMANN: So we need to change, a little bit, the approach on how we try to push the envelope. So what we did in the past is, we tried to squeeze more and more information over one communication channel. We are reaching the limit to that. So now, we need to explore different dimension. In the simplest form, you could think about it - well, just using multiple cables in the ground. Rather than having one optical cable, we could think about putting two, three, four, five cables into the ground.

But it is very costly. We would have to dig up the ground; we would have to put multiple equipment in. So what we are looking at is an integrated approach where we can transmit parallel communication channels over one physical cable, basically. And this is allows us to then work with or bypass the limitations that we will reach very soon.

HOBSON: So you're saying that even though we're moving into a world where we individually are using more and more wireless devices - that the actual, physical wires that connect the Internet around the world are what we're lacking right now. We need more of those.

HOFMANN: That's an interesting question. So we clearly need more of those, but wireless communications will of course become even more important in the future and will also be a bottleneck. We need to apply the very similar techniques also in the wireless world.

So today, for example, we have these cell towers that we all know and love or hate, right? But we have those big cell towers, and they cover a relatively, let's say, large area, coverage area. What we will see in the future is that we move a little bit away from these large towers that cover these big areas and towers with technology where we have smaller access nodes, similar maybe to the Wi-Fi access points people have in their homes.

And they will cover a little bit smaller area, but we will deploy more of those. We will see maybe some of those access points deploying on lampposts, street lamps or in buildings. And this technology is very similar to what I just described, where we have now multiple, smaller access points covering the same coverage area than maybe for example some of these macro cell towers today.

HOBSON: But Markus, just like when there are too many cars on the road, and you have to build a new highway, eventually the newer, wider, bigger highway is going to get clogged up, as well. Don't we have a problem with what we're using and, you know, how much we are asking of the Internet, perhaps?

HOFMANN: You are absolutely correct. So what I described so far is one approach, one pillar that we need. So we need to continue driving up the bandwidth, but using these new techniques. That alone will not be sufficient, we believe, or I strongly believe. We need to be a little bit more clever in how we use those resources that we have from two fronts.

One is it is not really about the data that we transmit today, it is about information, right. So my mom, for example, is not interested to learn about what the temperature is in all these different locations. She's interested in knowing does she have to take an umbrella with her or not when she leaves the home. That's the kind of information she's interested in.

Yet today, what we do, we collect all this raw data, like temperatures or wind sensors, transmit that all around the world to then extract the relevant information at the last possible point, basically. So we believe that if we follow an approach we extract the relevant information from the raw data much earlier, much closer to the source, therefore can transmit much less data around the world, that this will also help to actually manage this (unintelligible), the explosion of data that we expect to see.

HOBSON: Well, how do you figure out what's relevant? I mean, I'm going to go to Google right now and type in weather, let's say, and I come up with 1.2 billion hits in .3 seconds.

HOFMANN: So at the moment, the way the Internet has been designed is that the network is really a black box for the user and for the applications and vice versa, meaning the network does not know anything about the data it transmits. We need to find ways where the network can become more aware of the data it transmits. So an application can, for example, indicate, OK, this particular data is high priority or has a specific meaning and therefore should be treated differently.

Then we can actually improve the efficiency and actually achieve that kind of what you just said, how does the network what kind of data needs to be treated differently and what networks have to be used for certain applications.

HOBSON: Now Markus, if we do what you say, and I read here that the number of smartphones, tablets and other network-connected gadgets is going to outnumber humans by the end of the year, and we make them able to figure out what's relevant to us, should we be worried that they're going to take control pretty soon?

HOFMANN: That's certainly a worry, and we need to be careful here. So my view is, and our view here is, that this should be completely automized. So I always get nervous about we talk about a system or an application or a service or a network trying to figure out what I want and then trying to make decisions on my behalf. So this is not really what I'm talking about.

What I'm talking about is that the systems, networks, services, applications support me in making those decisions and allowing me to express them, right? So for example, I don't know, if I have a conflict on my calendar, I don't want any applications to make a decision to remove that conflict but rather to alert me and to offer me smart choices how I could resolve the conflict.

But I at the end want to be in charge. I at the end want to decide what is the right thing for me to do. And in particular, I'm also worried about privacy rights because those systems need to maintain my users' privacy, as well. So this is a fine line. We need to be very careful about that. But there are absolutely ways to find the middle ground here that allows us to protect the privacy, as well as let user in control while still allowing the kind of flexibility that I talked about.

HOBSON: I was only half-kidding when I asked that, but now I'm really scared that you actually took me seriously and had a good answer for it. Markus Hofmann from Bell Labs Research in Homedale, New Jersey. Markus, thanks so much.

HOFMANN: Thanks for having me.

HOBSON: And still ahead today, how that train crash in Quebec is giving new arguments to both sides in the Keystone pipeline debate. News is next, HERE AND NOW.

(SOUNDBITE OF MUSIC) Transcript provided by NPR, Copyright NPR.