Problem is that the Bells are fighting municipal broadband tooth and nail. And winning.
Municipal Broadband is a tactic to get municipal governments to pay millions and give huge amounts of free spectrum to a technology that has no business model on its own. There are substantical commercial products already available. See bold at bottom of article if you don't want to read the whole thing.
Seybold dismembers Wi-Max business model once again:
The WiMAX Push Is On!
Thursday, August 4, Intel held an analyst's briefing entitled "WiMAX Spectrum Overview." The briefing was conducted by Margaret LaBrecque, director of WiMAX spectrum policy for Intel's Broadband Wireless Division, and chair of the WiMAX Forum Regulatory Working Group.
The briefing started with the Intel WiMAX vision, which was depicted in a slide that called out:
High throughput access for business, backhaul and some residential customers
Mobile/portable broadband
Consumer broadband access (WiMAX to the home, Wi-Fi within the home)
Hotspot backhaul
Once these objectives for WiMAX were presented, she moved on to spectrum goals for the next three years. These include sufficient "harmonized" (read world-wide) spectrum to provide true broadband, high-volume WiMAX products. At present, two licensed bands are being considered (2.3-2.4, 2.5-2.69 and 3.3-3.8 GHz with the initial focus on 3.4-3.6 GHz) along with "license exempt" spectrum in the upper portion of the 5-GHz band (5.725-5.85 GHz) because these frequencies permit a much higher output power (4 Watts EIRP) than is permitted in the rest of the band.
With these bands, there will be a single RF IC in the 2007 timeframe to cover these portions of spectrum that will "allow flexibility to address national situations." Intel also expressed longer-term goals of employing DTV (TV bands) spectrum. (I will comment on that later.)
One slide showed the regions of the world where Intel can roll out WiMAX in the 2005-2007 timeframe using these bands -- virtually all of the world. It is obvious that Intel is working diligently all around the world to make this "harmonized" spectrum available for WiMAX. Intel believes that any operator that deploys WiMAX should ideally have 30 MHz or more of licensed spectrum.
The final slide in this section of the presentation showed a view of the world and illustrated the fact that if it is successful, WiMAX will have more spectrum available to it than today's 2G and 3G voice-and-data networks. A lofty goal to be sure, and some of the spectrum was labeled as being both 4G and WiMAX. I find this to be interesting since Intel's contention appears to be that 4G will need a lot more spectrum than existing 3G systems. And here I thought new technologies were supposed to be more spectrally efficient. The explanation was that more spectrum will be necessary because of high demand for very-high-speed data services.
Intel also made the point that rule changes are needed in several areas. A prime example was adjusting the term "fixed wireless" to include nomadic operation in which the terminal location may change but it will be stationary when in use. Intel and the WiMAX folks are working the world's spectrum mangers hard and long. Several updates indicate that they are making inroads in India, China, Malaysia, Europe and the United States, especially in the 3.3-3.4 and 3.6-4.2 GHz band. When discussing the use of unlicensed spectrum in the 5-GHz band, excuse me, "license exempt band," they are targeting the upper portion of the band where the permissible RF output is 4 watts and the spectrum is specifically for outdoor use. One of the analysts questioned the use of unlicensed spectrum for WiMAX services and was told that there are many thousands of WISPs (wireless ISPs) using Wi-Fi for wide-area services that would switch to WiMAX because it provides faster data services.
When asked why Intel was not targeting the UMTS spectrum available for TDD services that are already licensed by many wireless operators in Europe and elsewhere, the response was that there isn't sufficient spectrum available in that band, but if three or more vendors decided to build equipment for any band it would support it.
Plans for the spectrum around 2 GHz are taking shape around the world and, according to Intel, it is just a matter of time before it is all stitched together and WiMAX becomes the technology of choice within these large chunks of licensed spectrum. Voice in the form of VoIP was not addressed, but cell site deployment, sectorization, smart antennas and other technologies that will provide better coverage, more throughput and better reuse of channels were.
Moving To the Lower Spectrum
Intel's strategy for the TV bands is twofold. First, it believes the fate of 698-806 MHz is pretty much decided and it will be licensed and auctioned, but when I asked if Intel would support the new Wireless Caucus that is pushing for 20 MHz of unlicensed spectrum (see last week's commentary), they were not sure of Intel's stand on this matter. This came from the director of Intel's WiMAX Spectrum Policy who went on to talk about unlicensed use of vacant TV channels in various parts of the nation. The game plan for the below-1-GHz spectrum is to work to re-claim spectrum from the digital TV transition and to work toward making this spectrum available for use on a more global basis.
As for sharing unused TV spectrum in various cities (each TV channel occupies 6 MHz of spectrum), there are a number of issues here, too. First, in many major metro areas police and fire systems have been using some of this "unused" TV spectrum for many years. Another issue is that if I were to make use of, say, channel 19, in an area where channel 18 or 20 was on the air, I would discover rather quickly that the spectrum was full of noise from either side. Because this spectrum was allocated for transmit-only TV services, noise generated to adjunct channels has not been a concern -- but it will be.
My Comments
I see several holes in the WiMAX business plan. First is the comment that in some cases indoor WiMAX to Wi-Fi devices might not provide sufficient signal from the nearest tower, in which case external equipment would be required. This makes the return on investment even tougher, especially if a truck has to be sent to install this device. And I am not sure how units will be installed and certified using licensed spectrum even the in home units. Today, if I order a DirecTV two-way Internet system, the RF portion must be installed by a licensed or at least trained technician, which greatly increases the cost of the service.
Perhaps my biggest issue with this presentation is the amount of spectrum Intel thinks it will obtain to support WiMAX. It is working on hundreds of Megahertz of spectrum in many different bands for, at this point, a data-only service. In underdeveloped countries and areas, this service might be well received. However, we still have the issue of a return on investment. If you recall in last week's commentary, once again I stated that I do not know of a single terrestrial, data-only service that is making money as a standalone endeavor and I challenged anyone with hard facts to come forward and show me they are. Guess what. I did not receive a single email or phone call from anyone.
So WiMAX will cover more area than Wi-Fi, it will be faster than Wi-Fi, it will include quality of service and many other features it will need, but it remains a data-only play. I still cannot pencil the return for services other than backhaul or the extension of existing services. Yet Intel is asking for hundreds of Megahertz of spectrum more than voice and data services that are serving vast populations. How does this compute?
At least Intel's spec sheet, attached to the end of the presentation, is the most realistic I have seen when it comes to range predictions. It is now claiming (at 2.5 GHz) 2 to 10 km in semi-rural areas and 2 to 5 in urban situations. At 700 MHz, the maximum range will be 35 km (what about the device in the field being able to transmit back to the cell site?). For 802.16e (the nomadic/mobile version of WiMAX), it is claiming 2 to 7 km for outdoor coverage and 1 to 5 km indoors. All of this with data speeds of up to 75 Mbps in 20 MHz of spectrum, 4-19 Mbps in 5 MHz, and for the mobile version up to 15 Mbps in 5 MHz of spectrum. These are impressive numbers, but not earth-shattering given the fast movement of other technologies today.
It would be interesting to take HSPDA or EV-DO Rev. A, spread it across 20 MHz of spectrum, and see what the data rates were. With 20 MHz of spectrum, even given guard bands, today I can put 15 EV-DO 1.25-MHz carriers in that spectrum and end up with a lot of broadband service for a lot of people. If I combined EV-DO channels to provide aggregate data speeds, I could get to some pretty high data rates with an existing technology, and HSPDA will provide high data rates as well. The same would be true of IP Wireless, Flarion or others. What I am saying is that WiMAX won't provide that much more in the way of broadband data speeds. WiMAX will be a standard and it will be built into a number of devices, but I expect both flavors of 3G to be built into notebooks and PDAs starting this year. There are already devices on the market that include EV-DO in handsets.
Of course, in order to deploy all of this new technology, we will need to employ existing cell sites and build more. In today's world, getting new cell sites approved by locals is problematic and takes a lot of time. It will be interesting to see how this all comes together. I do believe that there is a place for WiMAX, I just doubt that the potential market is as big as Intel seems to believe it is. If you combine all of the potential devices -- home, office, nomadic and perhaps mobile -- and then you weed out those where a customer is already using DSL/cable or existing voice and high-speed data services, I think the numbers are not there for standalone WiMAX systems.
WiMAX as backhaul and as an extension to existing services (BellSouth's roll-out in its existing markets) makes sense to me, but I don't believe WiMAX will be the huge win Intel and the WiMAX Forum believe it will be. Therefore, I don't believe they should get all of the spectrum they are seeking. How about if they start with a block in each country and be granted licenses for other blocks if needed? Make them prove, as the existing commercial network operators have had to, that they need it because the demand for WiMAX service is so overwhelming. Giving them access to all of this spectrum based on their belief that WiMAX will be a huge win for everyone simply does not make sense to me.
Andrew M. Seybold |
