|
Written by Dave Burstein
|
|
Friday, 27 January 2012 14:01 |
40%, not 92%-120% “Data consumption right now is growing 40% a year,” John Stankey of AT&T told investors and his CEO Randall Stephenson confirmed on the investor call That’s far less than the 92% predicted by Cisco’s VNI model or the FCC’s 120% to 2012 and 90% to 2013 figure in the “spectrum crunch” analysis. AT&T is easily a third of the U.S. mobile Internet and growing market share; there’s no reason to think the result will be very different when we have data from others.
 With growth rates less than half of the predictions, a data-driven FCC and Congress has no reason to rush to bad policy. Wireless technology is rapidly moving to sharing spectrum, whether in-building small cells, WiFi, White Spaces, Shared RAN or tools of what the engineers are calling hetnets - heterogenous networks. The last thing policymakers should do is tie up more spectrum for exclusive use; shared spectrum often yields three to ten times as much capacity.
Bad compromises on the video spectrum are unnecessary because plenty of spectrum is unused. That includes the 20 MHz that M2Z would be building out today if Julius hadn’t blocked them; the 20 MHz the cable companies are sitting on and want to sell to Verizon; and the 30 MHz or so Stankey identifies as fallow at AT&T.
40% growth is still substantial, but wireless technology is improving at a breathtaking pace. LTE has about 10x the capacity of 2.5G and 4x the capacity of 3G. LTE Advanced, deploying beginning 2013 at Verizon, is designed for 10x the capacity of LTE. Putting more spectrum to use would be great, but let’s do it right.
Wireless speeds are actually going up dramatically, with AT&T delivering 2-5 megabits to most of the country and Verizon’s LTE delivering 5-12 megabits to 2/3rds of the population. Verizon is ahead of schedule to bring 5 megabits+ to 92% of the country in 2013 and 96-98% in 2015-2016. AT&T and Sprint have raised capex to catch up. 80%+ of the U.S. will have a 5 megabit offering in 2013-2014, 90%+ by 2015 or sooner. That’s without any additional spectrum.
Today’s wireless networks are designed to be shared: towers, WiFi, White Spaces, DAS and small cells all working together. The best engineers in the world are working on RAN sharing, SON, hetnets, 8x8 MIMO and techniques I’m writing about in my next book, Gigabit Wireless. AT&T in fact is one of the world leaders in DAS, WiFi and femtos and behind the scenes a key thought leader. There’s wonderfully exciting stuff I’ll be doing my best to translate for non-engineers.
Takeaway: The future is sharing the airwaves so let’s get the policy right.
|
|
Read more...
|
|
Written by Dave Burstein
|
|
Saturday, 24 September 2011 12:01 |
2012-2014 LTE
- Speed: Decent 5-15 megabits average, 30-100 megabits peak
- Capacity: Lousy, only a few hours of TV/month. 2 gig caps would disappear absent market power, but it’s hard to offer caps above about 15 gigabytes
- Design: Mostly towers, with about 20% WiFi and small cells
- 20 MHz maximum
- Users: 10’s of millions, not hundreds of millions, with 3G still dominant
- Policy goals: get it built, with the right price and capacity
2013-2020 LTE Advanced 4G can be 10 times better. (Shared gigabit)
-
The technology will come quickly, but only in a few regions will it be deployed
-
Speed: Better 10-50 megabits average is possible, 1 gigabit shared peak
-
Capacity: Better 25-75 gigabyte caps at an under $50 consumer price practical, although won’t be available without strong competition or regulation.
-
Design: Towers will go to 8 MIMO small antennae, doubling or tripling throughput100 MHz, where available.
-
Massive policy changes required most places to make contiguous large blocks available
-
30-70% WiFi, femto and small cells “bottoms-up”
-
“HetNets” Heterogeneous networks. For better capacity, not peak speeds
-
Interference smart, self-organizing networks (SON). Mesh features
-
Users: it will be 2015-2018 before LTE dominates in most markets
-
Policy goals: Total spectrum policy revamp for more sharing, use it or lose it, and obligatory network design for efficiency., including number of towers and small cells/offload
-
Where weak competition means deployment delays or high prices, more effective strong policy.
2015-2025 5G, probably with five times LTE Advanced capacity
Early discussions of how to use better coding, higher MIMO, cognitive radio and other interference reduction measures, meshes etc.
Think of a small cell on every street light or similar. |
|
Written by Dave Burstein
|
|
Monday, 19 September 2011 17:32 |
|
Takeoff possible then. CFO Braxton Carter of MetroPCS told investors the handsets are currently too expensive to attract many customers. After subsidy, MetroPCS is offering the handsets for $199 & $299, which implies a manufacturer price over $400. By the second half of 2012, he expects handset prices of $200 to drive material growth. At that price, he can sell them to retail customers for $99. Verizon CFO Fran Shammo, also speaking to the Merrill Lynch conference, pointed to battery life as the other major obstacle to LTE. He hopes to overcome that as well in 2012.
LTE phones eat power and are expensive because currently they effectively contain both a 3G and 4G system. LTE requires significantly more processing power, hence a larger, power-hungry chip. The speed is much higher, also putting demand on the processor. Even the most advanced network, Verizon's, only covers about half of the country, so the phone also needs a robust 3G capacity. Especially at Verizon, that's a different technology and hence requires components and power.
Chipmakers including Qualcomm, Broadcom and Marvell believe they have solved the problem with the next generation of wireless chips at 28 nm, now starting to sample. The smaller design rule directly reduces power while allowing far more functions on-chip including the 3G fallback. Limited quantities of 28 nm chips are coming from TSMC, the largest foundry. 28 nm production at TSMC is virtually sold out for the next six months, EE Times reports, because Apple is shifting the iPhone/iPad tablet to TMSC from Samsung.
As the LTE chips mature and the foundry capacity becomes available, everyone expects dramatic change in the LTE handset market. |
|
|
Written by Dave Burstein
|
|
Friday, 27 January 2012 04:54 |
 Xavier Neil's 20 euro ($27) unlimited voice, data, and SMS deal is drawing a million customers a week to Free Mobile. He's changing the mobile world by selling at half the price of any major operator in the U.S., France, or Canada. It's not a gimmick. Xavi understands that wireless costs per minute go down dramatically with volume. Free's low cost, high volume model will net half a billion a year if he runs it well. High cost, low volume carriers like Stephane Richard's France Telecom can bluster but that won't be enough. Brian Williamson of Plum Consulting has a good look at wireless costs that suggests Free Mobile will do very well.
Every honest official is thinking how to copy France. Every operator in the world is scared they will succeed. Every investment analyst is trying to decide which carriers are vulnerable and need to be downgraded.
Xavi's trump card is that his "Bottoms-up" network design is years ahead of almost everyone else. He has 5M WiFi and femtocells to serve his mobile customers. Paris is now a Free WiFi cloud. Jennie, I and iPad had a wonderful two weeks in Paris in September. We rented an apartment in the Marais with Free DSL that came with WiFi. Everywhere we went, including Chartres, we could log on to Free WiFi,
WiFi and femtos aren't "tower offload" in the Free network design but are the primary connection. Less than half the traffic will ultimately go to the towers, saving both construction and spectrum. Everyone including AT&T and China Mobile has been talking for several years about building networks "Bottoms-up."
Xavi revolutionized the Internet once with the 20 euro triple play. He�s about to do it again to mobile.
|
|
Written by Dave Burstein
|
|
Saturday, 24 September 2011 11:57 |
World in 2015: Half a billion of these (hold up iPad/iPhone). 60+% used at home/office. Jennie often watches on iPad for convenience rather than 50 inch TV
- Bill Smith of AT&T explains “the sooner I get the bit to ground, the cheaper it is.” Except for Verizon, every large carrier I know intends to offload as much as possible. Physics is in favor: signals get weaker with distance. Economics are the driver, with 1/3rd to 2/3rd savings promised by optimists.
- Cisco’s estimate is that 22% of U.S. mobile data will be off-loaded in 2013 and I’d guess in 2015-2016 that will be much higher. In a world where everyone carries an iPad, iPhone or similar logging into WiFi/femto will be natural.
- Estimates of 30-70% offload a few years out are common and make sense to me. Different carriers and different countries are going at divergent paces.
- LTE is great but 3G will dominate most places until 2016-2018
- Don’t forget small cells like Alcatel lightRadio, Ericsson Het Nets, ?Ruckus. SON & interference management probably biggest factor in capacity. High backhaul costs currently killing. $3K is cheap for 100 meg microwave.
- Sidenote. We have the speed in LTE 5-15 megabits. We don’t have the capacity. 2-8 gigabytes/month is only a few hours of video/week. Advanced can do 25-75 gigabyte caps if competition/regulation pushes. Not preferred business model. Verizon demo’ing how to charge Hulu, NetFlix, Google for video.
“The World as One Big Hotspot” Ian Scales
- China Mobile announced 1M hotspots; Korea Telecom, KDDI, and many others will soon be in 6 figures. Softbank in Japan is virtually giving both a DSL connections and a femto/WiFi gateway shared to new mobile customers, Free in France has 4.7M WiFi gateways, many of which will be turned on for mobile roaming. The Revolution Freebox has 3x3 MIMO and a femtocell. DT’s new gateway has 3x3 MIMO and I’m sure they and Voda in Germany will put WiFi to major use.
- Femtos have only sold about 1M and probably 1BWiFi. AT&T postponed indefinitely their rollout of 5M for a cloud across the U.S. Ralph de la Vega says interference problems near the towers are not yet solved. I speak of WiFi/femto because many carriers preferred femtos, including Vodafone, to control the customer. My guess is that with ubiquitous WiFi phones/pads WiFi will win, but still unproven.
- Traffic growth gated by caps and pricing. Some expect it to dramatically slow outside of WiFi coverage. Unclear.
Drivers of even more demand for best in class WiFi
- DT and I believe Free also have CAT-IQ HD Voice. Tony Werner of Comcast and Tony Melone of Verizon Wireless also have promised HD Voice. The world will notice when they turn on, probably 2013.
- Swisscom is selling to customers Quantenna/Netgear 4x4 as a “Wireless HDTV connecting kit CHF 199” That’s the first carrier to publicly use WiFi for HD. Slews of others are in trials and RFPs. Going wireless and saving a truckroll is worth $200-400 home.
- 3x3 and 4x4 are getting cheap. Quantenna tells me less than $5 over the $3-4 basic WiFi chip, presumably narrowing.
|
|
Written by Dave Burstein
|
|
Monday, 19 September 2011 17:17 |
|
LightSquared's ambitious schedule.Frank Boulben, Chief Marketing Officer, "anticipates completing its nationwide build covering 260m POPs roughly 12-18 months before its YE2015 deadline." Working with Sprint and using Sprint towers and backhaul they expect will speed things considerably. That's faster than pioneer Verizon (5 years to 92%) or AT&T (4 years to 80%, unclear but implicitly 6-8 years for 97%.)
That would be a remarkable achievement from a company that has no operations. The technology is advancing rapidly so it's not impossible. Boulben also told the Merrill Lynch audience "It needs an additional $3.5b in funding to reach cash flow breakeven." That truly would be remarkable, The company has limited cash on hand and will have to cover operating losses for years before breakeven. He didn't break things out, but that figure implies the cost of a nationwide U.S. build of LTE of perhaps $3B.
We need more competition so best of luck to them.
|
|
|
Written by Dave Burstein
|
|
Wednesday, 12 October 2011 16:17 |
|
Massive spectrum waste in small blocks. Combining the spectrum of AT&T and T-Mobile will result in more efficient use of spectrum, AT&T tells Sascha Sagan of PCMAG. The result would be "improved network utilization, channel pooling, the elimination of redundant control channels, and other efficiencies that effectively provide the functional equivalent of new spectrum."
Using spectrum in contiguous larger blocks is significantly more efficient and emerging as a key factor in spectrum discussions. Currently, the most advanced spectrum policy in the world is coming from Kenya, where they are creating a 100 MHz network run by a private company but open to other carriers. That increases the effective spectrum by 30-70% and perhaps more. Merging spectrum into large blocks is the natural policy in countries counting on wireless for broadband where they need as much capacity as possible.
To the extent spectrum is short, the first step is using it efficiently. At Columbia October 14, I'll be speaking of obligatory network design for efficiency. That includes the number of towers, of course, but especially small cells for offload.
|
|
Read more...
|
|
Written by Dave Burstein
|
|
Friday, 23 September 2011 18:04 |
 Verizon is the great dissenter on the business advantage of offloading wireless data to WiFi, femtos and small cells. From China Mobile to AT&T to France Telecom, nearly every major carrier in the world is rushing to offload 50% of traffic and more from macro cell towers. Small cells closer to the user are thought to be much cheaper. "The faster we get the bit to ground, the less it costs." It's simple physics: signals get weaker with distance.
Tony Melone of Verizon is the one voice disagreeing on the move to bottoms-up networks. While Verizon will use WiFi in the home and crowded areas like sports stadiums, they are not going hard to a general deployment.“In my mind it’s much more effective to invest in your 3G and 4G environments than rely on Wi-Fi.” I respect Melone so take his opinion seriously. In a briefing on small cells the day I'm writing this, I presented the Bottoms-Up approach as the common wisdom but added his disagreement.
Melone may or may not be right that the company is more profitable not offloading so heavily. But it's clearly wrong from the public point of view. Spectrum isn't as scarce as the fearmongers have it but does have limits. It's a commons we need to share and protect. Melone's boss Lowell McAdam at Goldman Sachs urged the government to "get spectrum in the hands of people" who need it. McAdam sensibly spoke of transferring spectrum that's underused but also of auctioning more, which means enclosing more spectrum.
|
|
Read more...
|
|
