Verizon CEO Ivan Seidenberg this week tried to put it to the cable community that they are in trouble from LTE. He to stood up this week at an investor conference and said that there would be some LTE-cable substitution ‘on the margin,’ as a modest substitute for in home cable broadband. Now how would that work exactly and is Seidenberg being a bit far-fetched?
Well yes, and then again no, because it’s more complicated than you think, and additionally there are two other issues here, one that Seidenberg is expected to retire in two years, so he won’t be held to his predictions. Also that over the past ten years cable has done so much damage to Verizon in terms of forcing it to invest in fiber and as part of the reason for Verizon losing so many fixed voice lines, that you can’t really blame him for trying to get back at it.
But the bald facts of it are pretty simple. Currently the LTE specification allows for 100 Mbps downlink rates and about half that as an uplink in a single 20 MHz spectrum slice. What that means today the way cellular networks are built, is that a single Macro base station, designed in much the same way as it is today, will be able to serve perhaps 50 people concurrently (a sensible average for those using data), which would offer around 2Mbps to each of them, which takes us back to where fixed line broadband was 5 years ago. How is that going to man up against cable with its potential for DOCSIS 3.0 offering 150 Mbps and beyond? It won’t.
Even when 4 x 4 MIMO antennas are commonplace in devices, this will probably only increase about 3 fold, and two by two antennas will be about a 2 fold increase, so 200 Mbps, let’s say 3Mbps or 4 Mbps each at best and the more people using data concurrently in any given base station, the less each gets.
But right behind LTE comes LTE advanced, and to get an idea of implementation schemes, LTE reached a standard in early 2008 and LTE advanced is close to a standard and may reach one in the next few months, so it sits around 3 years behind it. Does that mean that it will be implemented 3 years down the track, after LTE? Well again it’s complicated. Probably, because all equipment being built now can likely continue to be used on LTE Advanced.
The fundamental step is really simple. If everyone gets a base station in their home, like a low power femtocell, then everyone can have 100 Mbps bandwidth to themselves or with 4 x 4 MIMO maybe 350 Mbps. Because spectral efficiency today is close to its theoretical limits LTE Advanced is expected to simply sound out a new topology which is not so different in spectral terms from how WiFi works. It IS different in that services can be more easily guaranteed, because it is licensed spectrum and it has all the concerns about roaming. If you have more and more radio cells, you have to backhaul each of them and you have more and more cell edges.
At a cell edge the next cell is either using the same spectrum or neighboring spectrum and you have interference issues. This is where the complexity comes in. Each cell will be lower and lower power, and there has to be some dynamic management in each cell sensing the next and deliberately turning down its power so that it doesn’t interfere with the next.
So LTE gradually becomes LTE Advanced potentially with software upgrades and a new topology of where base stations need to be. Today’s Macro cells will stay in place, but then there will be pico cells and femtocells working within the reach of any given Macro cell once these dynamic management techniques are proven and nailed down.
Something similar is already going on in DSL technologies as we reported over the past two issues, with interference management in twisted pair bundles at exchanges and the technique is not so different. This way we get more and more spectrum re-use, with the same spectrum being used multiple thousands of times over the whole network. The whole thing has really come out of seeing what is possible in WiFi.
Which is why these numbers are familiar, because home WiFi systems for video distribution have virtually the same speeds, although they operate rather differently, and are able to move across a wide expanse of unlicensed spectrum. WiFi does however get carried on OFDM carriers, over short distances and that is at the heart of LTE and LTE Advanced.
But there is one important step where Seidenberg is perhaps washing over the issues, in that however future Verizon cellular networks are built, they will need backhauling and a strategy which relies on the backhaul from a home broadband fixed line will hardly work in unseating a cable operator from a home. Take out the cable modem and you have to backhaul it on something like a FiOS fiber. This doesn’t have to be one contracted for by the user for broadband, although it could be, but it could also be one which simply passes his home. So a FiOS fiber in the neighborhood could be used to backhaul multiple nearby femtocells. But in most cases plans to install femtocells will rely on existing backhaul to the home and if that’s a cable operator and if the US continues to not support Network Neutrality laws, then why wouldn’t the cable operator simply filter out Verizon packets from femtocells.
No, in our opinion Seidenberg was talking very casually, about LTE, not LTE Advanced and he was talking about Macro cell broadband wireless lines replacing broadband in cable homes where the requirement is 1 Mbps to 3 or 4 Mbps, and there is a mismatch with price and cable capability.
As Verizon offers more and more MiFi style devices, more and more cheaply, it could potentially do the jobs of all the broadband in a cable home. In order to make a case for replacement it will have to be offering this at a noticeable discount, say $20 a month beneath the price floor for cable broadband, perhaps $45 a month. Which implies a 3 or 4 Mbps service in line with what Clearwire is offering in WiMAX, for something like $25 a month.
So LTE attacking cable is a little like Tablets attacking home PCs ‘ it only works where the original solution was a sledge-hammer to crack and nut and no-one really needs it. But that might imply a significant number of homes right the way across America.
But as the LTE roadmap plays out there is no reason why it does not interfere substantially with cables. There are many scenarios where some form of backhaul, be it a FiOS cable at Verizon or a line of sight wireless connector operating in 80 or 90 GHz, couldn’t backhaul femtcocells which behave like WiFi (on steroids), with roaming.
Time and time again that future is laughed at by wired telecoms specialists, because wired will always beat wireless, but the truth is that if there is an economic basis for installing wireless in tight femto topologies, backed up with roaming Macros cells and WiFi hotspot offloading, and the price of LTE to WiFi conversion comes down (MiFi devices) then wireless can pack the same punch as cable, it’s just a matter of build out.
Back in 2007 NTT DoCoMo demonstrated what it called a real 4G trial at 5 Gbps using 100 MHz of spectrum, which translates to 1 Gbps in 20 MHz or 100 times what LTE is currently capable of.
Too many people think about operators filtered through the way they feel about them. They think , ‘Cellcos offer poor connectivity now, so they will never be able to do this.’ But the truth is that it is a deployment issue similar to cable. There used to be big cable loops and as there was more requirement and more income, the loops got split in two by buying more QAM. This is the future of cellular operators chasing an increase in the bits per second they deliver, for each megahertz of spectrum they own, for each area of the country they cover. And this is achieved by splitting cells into smaller and smaller cells, using more and more antennas (MIMO), more and more carrier aggregation and what they are calling heterogeneous networks ‘ multi-layer radio networks.
When cellular was all about voice then the architecture that it dictated sent radio across kilometers of distance and there was no point in chasing tiny cells, in the new economics of portable data it’s a case of bringing the cell closer to the people
Ten years ago everyone in cable said IPTV wouldn’t work. Eight years ago the IPTV brigade told me that video would never go over the internet. Today Seidenberg is telling us that cable could be ousted by wireless and we’re not going to be the ones to tell him he’s wrong, but it won’t be in his working lifetime because he retires in two years.