[ncolug] femto
- From: larry <larry@xxxxxxxxxxxx>
- To: ncolug <ncolug@xxxxxxxxxxxxx>
- Date: Thu, 01 Jul 2010 08:46:17 -0400
Small is beautiful: put a cell tower in your house
By Glenn Fleishman
Femtocells sound vaguely like a cross between a Feynman diagram and a
biology class, but they're the latest piece of gear that millions of
people will soon want in their homes without having missed them before.
A femtocell is a small cellular base station designed to provide
superior, short-range, indoor cellular coverage in a home or office. The
idea behind femtocells is simple: the hardware tries to capture the ease
of setup of a Wi-Fi network while allowing seamless connectivity for
existing cell phones.
Woojune Kim, the vice president of technology at Airvana, a mobile
broadband and femtocell equipment maker, explained the thinking behind
femtocells. "Can you take the economics of the last 10 to 20 years,
where we're able to make very small wireless transmitters like Wi-Fi
base stations—can you make cellular base stations small enough and at
that price point so that each of us can have our personal base station?"
The answer, after years of trying, is yes.
The compact base stations are a cheap way for mobile carriers to improve
coverage, while remaining relatively inexpensive for consumers to get
service outside a ground floor, in rural areas, or in places in which
their carriers have fallen down in meeting their needs. (Yes, we're
looking at you, AT&T, at least for now.)
Femtocells have been "coming next year" for at least four years, but
after successful introductions in 2008 and 2009 in tests and initial
rollouts, 2010 will be the first year for mass adoption. Analysts expect
hundreds of thousands of units to be in place this year, with tens of
millions sold each year by 2013 or 2014 worldwide.
Three US cellular operators have an in-home base station strategy in
place since an AT&T announcement in March. Sprint was first in 2008,
followed by Verizon in 2009, and then, nearly a year later, AT&T.
(T-Mobile chose a different tack than femtocells, which it recently
discontinued for new customers after four years in operation.)
The reason for the femtocell delay to market was twofold: 1) cost, and
2) the message it sent to the market. Until last year, femtocells could
cost $400 to $500 at retail; improvements in technologies combined with
large orders have pushed retail pricing as low as $100 to $150, with
carriers reportedly paying as little as $50 per unit with extremely
large commitments.
The messaging side was equally important: a femtocell told a carrier's
customers that the operator couldn't give them a good signal in the
home. The rise in popularity of 3G smartphones among average users,
starting with the iPhone but now far beyond it, has led to people being
dissatisfied with in-home coverage, whether or not they blame the
carrier for it.
Rob Riordan, an executive vice president at the Midwest cellular
operator and local exchange carrier Cellcom, said that the femtocell
message could be interpreted as "I offer you crappy service, and why
don't you buy this box from me, and pay me more money." Cellcom will
shortly start offering femtocells to businesses with a different
proposition behind it.
AT&T's notorious San Francisco and New York 3G undercoverage provoked
some unsurprisingly angry responses from residents who saw the company's
femtocell as a way to get customers to pay to improve AT&T's network.
See this New York Times article, with the provocative title, "Bringing
You a Signal You’re Already Paying For," for instance.
But not everyone feels that way. The industry believes that in the first
stage, there's a huge worldwide audience for people who are in the right
circumstances: living where coverage isn't expected (either by region,
topology, home building materials, or other factors), or isn't available
(rural or out of territory). In such cases, a carrier becomes a white
knight by having a cheap way to put their network in your home, even if
it's at your expense.
"Nobody actually expects their cell phone to work in the basement, or in
the elevator, or in any of those other unusual circumstances," said
Picochip's Gothard.
The next stage for carriers with broadband offerings or partnerships
will be far cheaper femtos built into equipment that they already
provide, such as set-top boxes and modems. When the femto comes
built-in, you won't feel line-item sticker shock.
The Case for Femtocells
Femtocells carry the same fundamental technology used in "macrocells,"
the large base stations deployed on towers and rooftops, but femtos are
designed to fit in a package appropriate for a home. There are also
microcells, used to build smaller cells in cities, and picocells,
typically installed to improve coverage in office buildings, campuses,
malls, and airports.
The larger siblings to femtocells pump out lots of power and cover
relatively large areas, from a building to square miles. That works well
for outdoor usage, but a macrocell is wasteful overkill to get a signal
into a house or office.
Simon Saunders, the chairman of the industry group The Femto Forum,
said, "The users that have the most marginal coverage actually occupy
the biggest part of those [macro] networks." He noted that with every
user moved to a femtocell, the network regains the bandwidth equivalent
of 10 outdoor users.
Because carriers have a finite amount of expensive spectrum licenses,
the motivation is to have the greatest reuse of spectrum by having small
cells. Small cells aren't affordable, however, so the contravening
business logic is to have the largest cell size possible to cover people
inside and outside with the dropped calls, low data rates, and related
problems.
That's why the home problem can be intractable. In many parts of the
developed world and in some developing countries, homes are made of
thick building materials, like stone, or modern materials, like the
plaster-covered chicken wire that's prevalent in warm, dry US climates.
Stone blocks signals and chicken wire turns a house into a kind of
Faraday Cage that prevents signals from entering or leaving.
Andy Gothard, director of corporate marketing at wireless chipmaker
Picocell, described using a macro to push service into homes near his in
Bath, England, as "trying to fill a cup by firing a fire hose through
the window," the window being the only signal-permeable part of the home.
In researching residential metro-scale Wi-Fi network failures a couple
of years ago, I came across Rio Rancho, New Mexico, where a provider
from Michigan hadn't counted on the plaster-and-wire construction of
Southwest homes. (The firm has since switched to WiMax and a different
business model.)
Femtocells were, in part, an attempt to find an alternative to
picocells, which were the smallest previous option. Picocells have all
the requirements of a full macrocell, including special backhaul
provisioning, air-conditioning, power feeds, and so on. Femtocells, in
contrast, approach the wireless bandwidth problem from the bottom up,
trying to provide a better experience than installing a Wi-Fi router,
and only needing a few minutes to power up, establish a location lock
and network communications, and be available for use. Airvana's Kim said
that a femtocell can't be as perfect as a picocell, which is virtually
indistinguishable from a macrocell, but that current femtocells are
close enough to make no difference to most users.
Interference can be an issue between larger base stations and
femtocells, with some clever work required to make sure that, for
instance, someone placing a call inside a house that connects to a
macrocell-—if the number isn't whitelisted, as noted below—doesn't
interfere with or receive interference from another caller connected to
the home base station.
Phones attached to a femtocell burn far less power, too, possibly less
than a comparable Wi-Fi connection, because the signal is strong and
close. Cell phones tend to run down when they have to use higher power
levels to punch through interference or reach distant cell base stations.
Femtocells in practice: the GPS requirement
I could get way, way too technical here explaining all the signaling and
interaction between a femtocell and a carrier's core network. But it's
abstracted from the user experience, which is a good thing. So in order
to use a femtocell, you just plug it into a broadband connection via
Ethernet, follow a few simple steps in a manual, and unfurl up to 30
feet of antenna for a GPS receiver. Wait, what? A GPS receiver? That's
the only part that you really need to know about.
Femtocells in the US include a GPS receiver, which drives the cost up,
but has some ancillary benefits. Sprint and Verizon must include a GPS,
because the CDMA standard relies on timing signals that are precisely
synchronized. AT&T opted for GPS for simplicity's sake in fulfilling
other requirements (like Assisted GPS). The femtocells typically come
with a 30-foot GPS antenna, although longer antennas may be available, too.
Woojune Kim of Airvana explained that there are three separate reasons
to have a GPS receiver: first, to make sure the femtocell can operate
legally; second, to provide Enhanced 911 (E911) calling information; and
third, to provide location-based information, or gather data that's used
for marketing.
In the United States, carriers have licenses to the exclusive use of a
patchwork of frequencies across tightly defined geographic regions. In
some countries, a carrier may have an allotment that covers the entire
nation. The GPS receiver ensures that a femtocell isn't used outside the
United States, and that the only the correct frequencies are used
wherever it's in operation in the US.
Saunders of the Femto Forum said that regulators and carriers outside of
the US have opted in some case to eliminate GPS in favor of sniffing for
macrocell identifiers nearby. They also can pull information from the
broadband line that pinpoints the location. (Even a very weak cell
signal can provide the cell IDs.)
You can see how tempting it would be to take an existing cell phone and
use a femtocell while traveling internationally, plugging it into
broadband to avoid roaming charges. But the use of those frequencies in
this fashion is illegal in most countries, and could get you thrown in
the pokey.
The US wireless E911 calling requirement is supposed to require carriers
to provide a relatively close position to the 911 response center that
takes the call. The current standard is within 50 to 300 meters,
depending on the technology involved.
A much-delayed second-phase implementation may require substantially
greater accuracy, and a femtocell could provide that, both by having a
precise physical address derived from the GPS, plus additional
information from triangulation among the femtocell and macrocells. How
carriers will use this information is still unclear, however.
The final advantage of GPS is in location-based awareness and targeting.
This gets the saliva running at carriers, which can figure out
additional services to offer to subscribers that rely on a knowledge of
where someone is located.
Cellcom's Riordan described several projects the company is testing, and
hopes to roll out, including deferred SMS messages. He described a
scenario in which his wife might think of chores and reminders for his
17-year-old son during the day, and she could send SMSes with a tag that
defers them. When his son walks into the house, he gets a pile of
messages on his phone, and his wife is notified that the messages were
delivered.
Riordan said, "They all get sent to Rob's [his son's] phone when he
walks in the door at the time he needs to see it." He added that the
system could be configured to alert when other people enter, too,
telling he and his wife when his son's girlfriend shows up at the house.
(Andy Gothard of Picochip said this technology is nicknamed "the
boyfriend sensor.")
Another project connects smart-home systems with femtocells. Riordan
explained that he has a system that uses motion detectors to fire up
lights and turn on the heat in his house. He's looking at extending
this. "You walk into the house, and the motion detector turns on the
lights," he said, but if the femtocell doesn't pick up his phone's
association, "the motion detector sends an alarm that there's an
intruder in the house."
Carriers may also pair home data plans with femtocell awareness. A
carrier that allowed unlimited bandwidth on a home plan might allow a
user to defer downloads over a 3G or 4G network until the user stepped
within range of a femtocell directly connected to broadband.
T-Mobile interlude
T-Mobile had the earliest home base station offering, launching
HotSpot@Home in 2006 in Washington State, and later expanding it
nationally after renaming it to Unlimited HotSpot Calling.
T-Mobile relied on unlicensed mobile access (UMA), an inexpensive option
that relied on Wi-Fi when femtocells costs were still hundreds of
dollars. With the assistance of handset makers and back-end voice
hardware providers, T-Mobile's service would let a handset swap off
seamlessly between Wi-Fi and cellular networks without dropping a call.
Calls could be received or originated on either type of network.
The company paired a Wi-Fi router that was ultimately priced at about
$50 (including a $50 rebate) with a service commitment that had two
Wi-Fi extensions to provide priority to voice packets over regular data
and minimize handset power use when a call wasn't active.
At one point, T-Mobile charged $10 per month for unlimited domestic
calling on top of a minimum cellular voice plan, including any calls
that started (but didn't have to complete) on a Wi-Fi network, whether
incoming or outgoing.
Earlier this year, T-Mobile stopped offering a related service, @Home,
which was most closely a competitor to Vonage, providing landline VoIP
with unlimited domestic calling using existing home phones. Sometime
after that, T-Mobile did away with its custom UMA plan and base station,
merging it into the Even More calling plans, which include fairly cheap
unlimited calling without a Wi-Fi restriction.
When prompted about the changeover, a company spokesperson provided this
approved statement: "Our new Even More plans, which feature options for
unlimited calling, text and data service, have taken the place of our
myFaves unlimited calling feature and our Unlimited HotSpot Calling
service."
T-Mobile continues to support previous UMA customers, and sells a
handful of BlackBerry models with UMA built in. A customer could
conceivably enable that feature, but it's not advertised.
The real issue for T-Mobile not providing this service is that as the
distant number-four carrier, it doesn't have network overuse. "Voice
offload is not a key concern at T-Mobile, so the only advantage of UMA
is in areas with insufficient coverage, but that makes it difficult to
market the technology," said Monica Paolini, the principal at wireless
analysis firm Senza Fili. That would be true of T-Mobile's four national
competitors from the marketing side, but each of its rivals could stand
offloading voice and data to keep their networks performing up to snuff.
T-Mobile may have delayed a transition to femtocells because of its
focus on rolling out a national 3G network, and trying to one-up AT&T
and competitors on network speed. Because T-Mobile has the fastest
national network—HSPA 7.2 Mbps up and running and HSPA+ at 21 Mbps
starting its rollout—it may be waiting for a next-generation femtocell
before moving into the market.
The Femtocell Deal
Now for the nitty gritty: does buying an in-home base station make sense
for you? It depends, because each US carrier is approaching these base
stations differently, and the capabilities are quite different.
None of the carriers currently offers any deal that discounts data, only
voice. Nor, to my knowledge and that of the several industry executives
and analysts I spoke with, does any carrier deploying femtocells. That
will change. AT&T has the technological advantage with its 3G MicroCell.
The AT&T femtocell can handle four simultaneous voice and data calls, as
well as push data out at a raw rate up to 3.6 Mbps using HSPA 3.6.
(AT&T sells several phones, including the iPhone 3GS, which support the
HSPA 7.2 standard for 7.2 Mbps of raw downstream data. However, the
MicroCell uses the 3.6 flavor, and AT&T says it will take until year's
end to get the backhaul to cell base stations to handle 7.2 on its macro
network.)
AT&T charges $149.99 for its femto with no add-ons, purely for coverage.
Separately, the firm offers a $20-per-month unlimited domestic calling
option, but that fee applies to each cellular line, even in a family
calling plan. If you sign up for the plan, you get a $100 rebate against
the femto. Sign up as a new AT&T DSL (3 Mbps or faster) or fiber
(U-verse) subscriber and get another $50 rebate.
The $20-per-month, per-line plan works only for a subset of AT&T
customers, because AT&T includes unlimited weekend and evening (9 pm to
6 am) in all its postpaid plans. For $30 per month per line, AT&T has
fully unlimited calling plans with no location restriction. However,
because AT&T relies on roaming agreements with T-Mobile and regional GSM
providers to handle coverage holes in parts of the US (notably in New
England), a 3G MicroCell might allow someone who wants an iPhone to not
have their service canceled by AT&T. AT&T will fire you as a customer if
you use too many roaming minutes, even if you're unaware you're off
AT&T's network.
Both Sprint Nextel, with its Sprint Airave, and Verizon Wireless's
Network Extender allow only three simultaneous voice or data calls.
Verizon is firmly in the "hey, this is for coverage expansion" school,
charging a whopping $249.99. This may be because Verizon has the best
home network voice coverage in the country. Sprint's Airave service is
seemingly cheaper at $99.99, but has a mandatory $4.99-per-month fee for
using the device. An optional unlimited domestic calling plan adds $10
per month for an individual plan or $20 per month for a multi-line plan.
The CDMA network standard both companies use doesn't allow simultaneous
voice and data over the same connection, and both firms' femtocells are
currently limited to 2G connection using 1xRTT for rates well below 200
Kbps. Both companies say that a 3G-capable phone will attempt to make an
EVDO connection to a macrocell, and only switch to the femto if such a
connection is impossible.
While Verizon has 3G coverage nearly everywhere it offers voice service,
Sprint's 3G footprint is about 15 percent shy of Verizon, with which it
partners for roaming. Sprint only allows 300 MB per month on its 5 GB
service plans on Verizon's network; exceeding that can get you canceled.
Forcing a Sprint phone to use slow 1xRTT at home could allow someone to
maintain Sprint service.
Now it might seem that data usage at home over 2G or 3G is a little
ridiculous, given that anyone with a femtocell would likely be operating
a Wi-Fi network and have a smartphone with Wi-Fi built in. But that's
not necessarily the case.
Airvana's Kim pointed out that even in his home, his wife leaves her 3G
phone set to 3G instead of switching to Wi-Fi most of the time. For
non-techies, he said, "Once you've got it set at one setting, why bother
with the other one." While many smartphones automatically swap over to
Wi-Fi for authenticated networks, or, in the case of AT&T, any for-fee
hotspots within range, standard 3G phones aren't that clever and may
lack Wi-Fi.
The Femto Forum's Saunders also emphasized that while Wi-Fi is in most
smartphones, the majority of the market for 3G handsets are so-called
"feature" phones, which run less-capable operating systems, but are
improving their browsers and Internet access tools. Owners of these
less-expensive, 3G-only phones will represent the majority of users "at
this point in time and for many years to come," Saunders said.
Using 3G over a femtocell will likely reach far closer to the maximum
limit of a given 3G service (3.1 or 3.6 Mbps downstream), because the
signal strength will be at or near its maximum. These higher speeds
won't drain battery life as fast because, as noted earlier, the phones
can use relatively small amounts of signal power to communicate.
Both Sprint and Verizon are rumored to have 3G femtocells coming,
according to a report in early April from IDG News Service. Samsung
apparently demonstrated a device at the CES trade show in January with a
Verizon logo on it, while Airave reportedly is providing Sprint's
next-generation product.
The new CDMA femtocells would handle EVDO Rev. A, the fastest CDMA
flavor currently deployed and likely to be rolled out in the United
States in the future. That standard has a raw data rate of 3.1 Mbps
downstream and 1.8 Mbps upstream. 3G CDMA femtocells have lagged because
of chip delays from some vendors. The Sprint device, at least, would
handle up to six simultaneous voice calls; the Verizon upgrade would
likely have a similar limit.
There's another parameter to consider about simultaneous calls. It would
be tempting for a carrier to let anyone within range, neighbors or
passersby, to make calls over your femtocell. Verizon calculates that
about 40 Kbps per call is needed, which is paltry if you've got 3 Mbps
DSL or almost any cable broadband connection.
Airvana's Kim said, "If you're selling [a femtocell] to a consumer, and
asking consumers to pay to put it in, how do you convince them, oh, by
the way, you're putting it in, but I want my other customers to be able
to use it." He noted that people's attitude has been educated by Wi-Fi
security warnings. He said, subscribers will think, "I bought this, I'm
putting this into my house, why should I allow other people to use it?
My Wi-Fi, I only allow my family to use it."
Despite this, both Sprint and Verizon allow anyone within range to have
calls handed off to a subscriber's femtocell unless that option is
disabled. Verizon requires a call, while Sprint lets you make the change
via a Web site. Both Verizon and Sprint let you whitelist up to 50 phone
numbers. Verizon says it only allows a single caller to use a femtocell
who isn't whitelisted.
AT&T has taken the opposite tack, only allowing authorized numbers to
interact with the femtocell, and requiring that you set a whitelist of
up to 10 numbers.
Right now, the three carriers require customers to pay for femtocells,
even though one might think it would make sense for the firm to figure
out its highest-paying, worst-reception subscribers and send them gratis
units. That's not happening yet on any wide scale, but while preparing
this article, a colleague on the east coast said Sprint sent him an
Airave with no charge and no monthly fees, possibly because of
complaints he'd made about coverage.
Future
This year's femtocell rollout is really just the tip of the iceberg,
industry experts say, and analysts tend to agree, with 4G networks the
focus.
The two primary 4G network technologies, LTE and WiMax, will both
benefit carriers from spectrum re-use, and that could be reflected in
deals to the consumer. 4G standards require wider channels—10 MHz to 20
MHz versus 1.25 MHz or 5 MHz for 3G—and that, in turn, means there's a
far greater need for carriers to conserve macrocell use.
The Femto Forum's Saunders said that the risk with LTE, particularly
with AT&T and Verizon pushing 700 MHz deployment in the US, is that with
an initial deployment, "You've used up so much of the bandwidth that's
available, you can no longer give the same levels of good coverage."
The 700 MHz spectrum penetrates into buildings and homes quite well, so
the first experience of users will be quite good. But the macrocells can
be saturated fast. He says the math works out to complement macrocells
and femtocells instead of focusing first on a macrocell deployment. A
carrier could build "a fairly sparse macrocell network, and they can use
femtocells to ensure that the network stays that way," said Saunders.
WiMax has a slightly different use case, because Clearwire and other
WiMax providers have marketed the technology both as a wireline
broadband replacement and as a complement to wireline services. Comcast,
for instance, resells Clearwire's Clear service under its own name in
some early markets as the mobile leg of an integrated bundle: fixed
voice, data, and television, and mobile broadband. And, soon, voice. A
femtocell for WiMax could plug into a cable partner's broadband in the
home, and allow mobile voice and data to reuse Clearwire's wide channels
of WiMax in the home. WiMax and LTE both have specific components to
prioritize packets and provision frequency use, as opposed to Wi-Fi,
which has optional add-on standards for packet priority alone.
At some point, femtocells will get both bigger and smaller, Alice in
Wonderland-like. Airvana already has a larger femtocell, designed for
enterprise and larger-scale use, as a cheap alternative to picocells.
Cellcom's Riordan said that the firm was already looking into using
femtocells to cover some of Wisconsin's 20,000 lakes—places where a
regular base station would be impossible, but an outdoor femto could be
solar-powered and backhauled wirelessly.
On the smaller side, the femtocell radio and technology could be
embedded into broadband modems and set-top cable boxes (which will
eventually be broadband modems with video outputs, anyway). Saunders of
the Femto Forum said, "For an integrated operator that's providing a
fully integrated bundle and wants customers to buy into that bundle,
then a fully integrated femto capability makes perfect sense."
Standalone femtos won't go away, because there will be plenty of
customers in every market who want to choose voice from one firm, and
broadband or video from another. But femtocells will disappear from our
field of vision as we get used to the idea of bringing the cell into our
house as yet another digital appliance we didn't know we couldn't live
without.
--
"We don't see things as they are
We see things as we are."
Anais Nin
To unsubscribe send to ncolug-request@xxxxxxxxxxxxx with 'unsubscribe' in the
Subject field.
Other related posts:
