The Mobile Internet Transformation – Meeting Network Capacity Needs of Cities
Wireless connectivity is revolutionizing the way people live and interact with each other. Advances in wireless technologies such as 3G, 4G and WiFi have enabled the proliferation of connected devices at affordable price points. People now expect always-on Internet connectivity anywhere they go. The result is mobile data traffic continues to grow at an astounding rate every year. How can wireless providers continue to increase network capacity in urban cities? In this presentation we will look at innovative ways of deploying wireless capacity using a dense network of very small cells, reducing energy consumption and urban space requirements for cell sites and antennas.
- Matt Grob, Chief Technology Officer, Qualcomm
(Presentation slides not available.)
Introduction by Gordon Feller
We have 2 types of mobility revolutions unfold and they’re happening at the same time and they’re converging with each other. We have the mobility revolution on your devices that make it possible for you to stay in touch with home base while you’re here sitting with us. They know where you are and you know where they are. Cisco knows where everybody is at all times. Then we got the people who actually make those smartphone technologies work. That’s where Qualcomm comes in. We asked them to join us as a co-sponsor not only because they’re a multibillion-dollar company and they have a fantastic array of technologies and you’ll see some of them downstairs today, but because they have a chief technology officer who’s one of the few in the world that actually doesn’t only talk about doing experiments. He actually does them, believe it or not. You sit in meetings with CTOs and say “Wow, let’s do an experiment.” Afterwards, you talk about it and you talk about it. Actually, Qualcomm does these experiments and it’s always amazing when they roll this technology out in the marketplace. The enabling technology that Qualcomm provides us that’s embedded in almost every one of your phones is enabling the other revolution you just heard about − the energy revolution. I want to ask Matt Grob to join us and walk us through that and tell us a little bit about the mobile internet transformation − meeting network capacity needs of cities. Matt Grob.
Thank you. Thanks, everyone. It’s great to be here. I’m going to tell about what we’re going to do to dramatically improve the performance of wireless networks over the next 10 years. We all know how important those networks have become for future cities, for vehicles which we talked a lot about that this morning already, for machine-to-machine applications, smart grids, smart meters and we touched on that as well, and then of course the consumer and enterprise applications that we’re all using all the time. What you may not know is how dramatic the growth is and how important it is that we continue to evolve the technology so that the prices don’t start going up. We want to keep the service plans and those kinds of things at the current rates or lower despite the demand going dramatically up. To do that, I’m going to spend the next 20 minutes to give you an insight and hopefully, change the way you think about the way infrastructure and wireless networks are deployed and motivate some different paradigms and give you some interesting statistics on what’s really happening.
There are a lot of smartphones in the world. There’s going to be 1.2 billion by 2016. There’s going to be hundreds of millions of tablets. There are already hundreds of millions of cellular-connected devices. More people are using the devices. There’s a lot more data per device being used. It’s just exploding. At Qualcomm, the way we set the bar is one thousand fold improvement in performance over the next 10 years versus today. When we sit down and try to design the next networks and what the capabilities are going to be, we say “How good does it got to be?” Our answer is it’s got to be 1,000 times better today, roughly a factor of 2 every year for the next 10 years. That’s what we’re shooting for. I’m going to show how we’re going to get there with that. We’re going to do a little math. This is Shannon’s capacity law. This sets a bound on the upper limit of performance for a communications system. It’s got 3 major terms here that we’ll talk about: the number of antennas, the amount of spectrum, and the signal-to-noise ratio. If you think about improving a system by 1,000 times, you look at that and you go “OK, that’s the law. Which one of those things can we change?” If we improve the SNR, that means making a better modem, a better receiver, more sensitive decoder, a more sensitive receiver filters −those kinds of things. We’re going to work on that. Are we going to make that part 1,000 times better? Probably not because it’s already pretty good, but we’ll still improve that. Are we going to increase the amount of spectrum, the w? Yes, we are. The industry is constantly looking for new sources of spectrum. That is going to be a big part of it. Are we going to get 1,000 times more spectrum? Probably not but we’re probably going to get 10 times more over the next 10 years. I’m going to talk about the different kinds of spectrum and how we’re going to go about that. The n refers to the number of antennas. Can we have more complex radios that have multiple antennas, 2 antennas, 4 antennas, that sort of thing? Yes, we’re going to do that. Are we going to get 1,000 times more from that? No. Are we going to get a couple of times? Yes. So how are we going to get to 1,000 times? That capacity is the capacity between 2 antennas subject to these limits.
Here is what we’re going to do. If there is one slide to remember from this, this is the one here. We are going to get there by shrinking the cell radius so that you get to have that cell capacity c over and reuse that spectrum many, many more places. In the past, you may have a 3-kilometer, 10-kilometer cell radius and you get a certain capacity in there, say 10 megabits per second in 10 megahertz of spectrum. Now, we’re going to shrink that radius and be able to have a dozen sites like that in the same space, each one of those getting 10 megabits per second. That is the single biggest source of gain. There’s going to be a lot more infrastructures in terms of the sites, the number of locations, and the radius is going to be a lot smaller. We’re going to talk about what that means.
We’re also going to think differently about inside-out and outside-in. It turns out traditionally, the bay stations are outside up on towers. A lot of the usage, in fact, the vast majority of usage is indoors. So you have to overcome path loss and so forth for that to work. That’s the outside-in model. We’re going to improve that. We’re also going to talk about a new paradigm, the inside-out model which means small stations inside buildings being able to provide coverage outside. Let’s talk about those things a little bit.
Starting with the traditional outside-in. Here, you’ve got very common today macro coverage area: big bay station, a couple of kilometers of coverage. Over that area, you have this capacity c and you share that amongst the users. What can we do to improve the performance of that? An operator will split that. They’ll say, “OK, let’s put 2 cells.” That’s one step. Now we’re going to introduce small cells inside the coverage area of the large ones like this picture is depicting. If you do that, there are pros and cons to doing this. You get more performance but there are some challenges. One is interference management. These devices transmit, they interfere with the big ones, causes some interference, reduces the SNR in that equation. We got to figure out how to mitigate that. Two is backhaul. You’ve got to connect. You can’t just put a bay station without connecting it to the internet. You’ve got to supply backhaul to those sites and that’s a challenge. Then the whole thing has got to be cost-effective. If you’re putting 20 bay stations in and they each cost the same as a regular one, it’s not going to work. We’ve got to have a completely different cost structure for these kinds of products and maybe even a different business model in terms of who pays for them, who deploys them and why they do that.
This is depicting the interference. The little fuzzy stuff there is the interference. We’re doing a lot of research on efficient and integrated interference management so that you can have co-existence between small cells and large cells and solve these problems. We’ve made a lot of progress on that.
In this depiction, we’re showing what happens to the overall performance of the network if you place a certain number of small cells for every large one. The first bar there is if you put 4 small cells for every large cell, you get 5.7 times improvement. All the way up to where you have 32 small cells for every large cell, we can get a factor of 37 times improvement. That’s extremely significant although it’s still the outside-in model.
So we talked about radio spectrum. Where is it going to come from and what kind is it? People say spectrum is like real estate, no one’s making any more of it. There is a sweet spot of bands that if you go above or below, the propagation characteristics are not very good. In this picture, we’ve gone from a few megahertz up to 60 gigahertz and depicted the colors there what bands are the most suitable. Those are the ones that we try to go after. Traditionally, cellular is 700 to 800 megahertz, 1.8 or 1.9 gigahertz. That gives you good outdoor properties especially if the 700 or 800 megahertz band. But now for small cells, we can look at 3 or 4 gigahertz or higher because we don’t really need the 10-kilometer coverage. We actually don’t want that. We want to reuse the channel more.
It gives rise to 3 classes of spectrum. The first one is what we all know about today. It’s the kind of spectrum that Verizon, AT&T and operators like that have. It’s traditional licensed band. The regulator auctions it off and generates some revenue for the Treasury. They buy the spectrum. They have control over it. They deploy. That’s good stuff.
On the other end of the spectrum, so to speak, is the Wi-Fi and unlicensed band. No one owns it; anyone can use it. However, you have to comply with some emissions regulations and so forth. Other than that, everyone can use it. In fact, it’s being used in this room massively. You can turn on a little Wi-Fi analyzer app on your phone and look at how many access points there are. It’s a huge number. If you look at what it was a year ago, it’s grown enormously. Pretty soon, we’re going to have more congestion problems there. It’s pretty popular and it has its place.
In the middle of that is something that we’re calling authorized shared access. It turns out there’s a lot of spectrum in the world and in the US that is owned by an incumbent such as the government or the military. If you actually look at the logs, it’s not used very much, spatially or temporally in many cases. They got it and they need it. They’re not going to give it up. You introduce technology that can make use of it but you can yield it if need be. If you do that, we’re hoping and working to get access to a lot more spectrum that way. That work may be spatially or temporally in many areas. If the primary user needs to use it for some national situation, in one second, the commercial user yields and then the primary user can use it. Through a technique like that along with some technology that goes with it, we think we can enable some more spectrums.
Let’s talk about densification and the backhaul solutions. If you start putting these small cells all over the place, then you’ve got to connect them. How are you going to do that? There are 2 basic ways: wired and wireless. We’re working on both of those. We’re working on more efficient Ethernet, powerline modems, and all kinds of wireless backhauls. There are different models. For example, in one of the world, there might be an operator with 4G spectrum but not yet a lot of 4G phones. You can use the 4G spectrum as a backhaul and then a small base station that relays into 3G mode with a lot of phones. That way, you’re actually making use of that spectrum. You’re not having to wire up the backhaul in those kinds of models. That’s part of the picture as well.
The economics here, the ultimate goal is to increase the supply of bandwidth and bits as fast as demand goes up so that the price doesn’t go up. We really don’t want the price to go up. We’re already seeing strain. Operators have data caps. A few of them are resisting that. They’re offering unlimited and that’s now a selling point in some cases. Generally, if we don’t do something, you can’t allow everyone to have unlimited use at the same price. That’s why we’ve got to do something and that’s why we’re introducing these small base stations.
Just to give you a sense of these things, I just happen to have one in my pocket here. This is a cellular base station. I’m going to show you the card inside it. This one is not even a prototype. This is a commercial chip and this is a commercial reference design. The biggest part on this board as you can see on the corner is the Ethernet connector. You might say, “How much does this cost?” We had a little competition of what ordinary object costs the same as a cellular base station and the winner was a tattoo. If you see someone with a tattoo, they could have had a cellular base station instead. Think of a cellular base station like this. You’re used to thinking of a big giant tower but going forward, it’s going to be a little module. This actually has a lot less on it in some sense than a phone. A phone has got a display, a camera, a battery, this does not. The cost structure of this can actually go very low. We’re working on integrating it with Wi-Fis. In the future, a cellular base station is much less a big box and is actually more akin to an app or mode on a little access point, just one more mode of many modes. When you think about deployment of Wi-Fi and things like that in enterprise and cities, you can now have both licensed and unlicensed. That is what’s really, really exciting. That can improve the performance. We can get the 1,000x gains. From an operator’s standpoint, you can leverage the other kinds of deployments that are already taking place. You can incent users. We talked about car-sharing this morning. That peer-to-peer distributed model applies here, too. You can incent a user to go to the store and buy one of these and put it in their house. It’s not going to cost a lot. Like I said, the cost structure of this in many ways is less than a phone. Then you have a deal where if you put this in your house and turn it on, then you may get a higher cap on your plan and you get to use the ones everyone else bought in their homes when they pass by. That’s a sharing kind of model. That’s a new paradigm. The product I showed you is actually 3G-capable but we’re developing one that has 3G, 4G, and Wi-Fi altogether.
Let’s talk about inside-out. Again, I mentioned that most of the data traffic these days −and this is true actually for years even for voice− is indoors. Most of our infrastructure is outdoors. With devices like this, we can put the infrastructure indoors. What happens if we actually do that? Here’s a depiction of a typical scenario. Maybe there’s a user that has an apartment and the macro coverage there is poor. So they call up the operator and say, “My coverage is poor. I’m going to switch or do something for me.” The operator says, “Go buy a femtocell and go put it in your place.” A couple of people start doing that and now, they’ve got good coverage in their little spot there. What happens if you start doing that more? This is licensed band. In licensed band, you can do many things. You can put performance requirements. You can do interference management. You can force the protocols. You can require the devices to have sensitive receivers. If you do all those things, something magical happens here. You actually get pretty good coverage outdoors. If it starts to get more and more popular which we are going to try and promote, what can happen is you’ll get pretty good coverage outside. That’s inside-out. I’m going to show you a real exciting result. In fact, in the introduction it was mentioned about experimenting. We certainly do a lot of that. Of the 2 slides from this whole presentation that I wanted you to keep in mind, the first one was the capacity with all the C’s. This is the second one. What you’re seeing here is not a simulation. This is a picture of a neighborhood from Google Maps or Google Earth. You can see the streets. We actually went and found out that there were femtocells at certain addresses. That’s what those red squares are. We did not place them there. They were already there. These were customers that have called up their operator for whatever reason and decided to buy a femtocell and put it in their residence. That femtocell today is closed. It only works for you. A passerby can’t make use of it but nonetheless, it exists. So we went out there with our test van and drove up and down the streets and logged the pilot signal strengths from all these femtocells just to see how well it worked outdoors. The results are overlaid there with those little colored dots. You’ll see it range from red to green. Red is the weakest signal. The bottom there is -105 dBm to -115 dBm. Even at -105 dBm, you can get 700 kbps. You can see there that geez, we have pretty good outdoor coverage. There was no big tower, no nothing. Guess who’s paying for the backhaul? Guess who’s paying for the electric power for this? It’s a different entity and in this case, the end user. We were very encouraged by that. I don’t have the time to show you all the test results but we built a network like this. I can tell you, it’s kind of intriguing. There’s a drive loop around Qualcomm that we’ve been using for 20 years to test base stations in cellular. We usually have 1 or 2 towers up on the roof and we drive around in our van and we test stuff. Now, what we did is put devices like this, a whole bunch of them, in offices, in random places, not necessarily by the window, just wherever. We drove around that same loop that we’ve been driving around for 20 years and did voice and data knowing the outdoor coverage is being provided by an indoor site. It works quite well. We’re really excited about that possibility.
To wrap up here, if we start with 1x, meaning the capacity we have today with a macro base station, if we then introduce small cells and let’s say 9% of all the residents in a residential area were to have a small cell along with 10x spectrum because this is a 10-year plan, you can see the whole picture here. 10x more spectrum that we’re going to get from shared access, auctions, and unlicensed altogether, and introduction of small cells if we can get them, the combination of those 2 things will give us 500x more capacity than we have today. If we go to 20% with a 10x spectrum, that’s how we can get the mythical 1000x. In fact, you can even go farther. You can go on up as far as you want. 1000x is roughly a factor of 2 for 10 years. Just one other data point of interest, if only 2% of residents −this applies similarly to enterprises− have small cells, you get double. So if only 1 in 50 does this, you can double the performance. We’re not going to break Shannon’s rule. We’re going to make the best radio we can with the most antennas and all that, but the biggest gains are going to be coming from this new kind of deployment. That’s how we’re going to do it − 1000x. Thank you very much.
We’ll take one question and then you’ll be available during the coffee break.
Yup, I’ll be available.
Sir, stand up and introduce yourself.
My question is aren’t we already in this small cell era? We just call it Wi-Fi, right? I suspect almost everyone in the room has a device that will reach some macro cell and yet most of us are connected to Wi-Fi. If we switch our voice communications to VoIP, doesn’t that work without resorting to this new femtocell?
That’s a great question. Thank you. Yes, we are. Wi-Fi is prevalent. This is not intended to displace Wi-Fi. In fact, what we feel is going to happen is these small pieces of infrastructure will have both Wi-Fi and licensed. You really want both because you want connection management. You want the ability to put certain flows on Wi-Fi, if you’re downloading an email attachment or something like that. Compare that to emergency calls or the kind of service the operator really wants you to rely on with a fixed or known QoS and a large coverage. What you really want, we believe, is both. One does not replace the other. They complement nicely. You want the efficient ability to hand back and forth between those 2 modes. We think both of those are going to exist. Qualcomm has got a large investment in Wi-Fi as well and they’re both very important.
Matt, thank you so much.