Wireless Phones: Untethered and Unreliable
Network Computing Online
April 5, 1999
By David Willis
People have been hanging up on
me a lot lately. I'll be in midsentence and
suddenly...dead air. It's not that I'm being more
obnoxious than usual, it's that I'm trying to do the
impossible. I am vainly attempting to conduct
business on a wireless phone that is simply not
ready for business.
In the past three years, I've
subscribed to five wireless phone
services. I've used analog cellular,
digital cellular and PCS networks.
I've used systems built on TDMA,
CDMA and GSM technology. My
advice is to disregard all the claims various
vendors make; the underlying technology doesn't
set their products apart. The major difference is in
the operators themselves. It's no small task to find
a company offering reliability that even
approaches wireline quality.
Now I don't want to single out any particular
service, nor do I need to--it's an industrywide
problem. On many online forums there appears to
be a long-running contest to identify the company
delivering the worst cellular or PCS service.
Name an operator and you'll find unhappy
customers, be it AirTouch, AT&T Wireless, Bell
Atlantic Mobile, Sprint PCS, Cellular One,
Omnipoint, Nextel or any other network. Each
has a predictable list of complaints: dropped calls,
busy network signals, phones that don't ring,
incomplete coverage and/or poor in-building
penetration--and those are just the most common
gripes.
One reason for the increased problems is more
users. According to the CTIA (Cellular
Telecommunications Industry Association),
wireless subscribership more than doubled
between 1995 and 1998. Meanwhile, the average
subscriber's monthly bill dropped 24 percent, and
per-minute costs have dropped 40 percent.
You don't have to be Warren Buffett to recognize
the economics. Wireless operators have modified
their principal target, shifting from high-paying
corporate users to consumers looking for
affordable service. Because profits per subscriber
are lower, operators must attract large numbers of
subscribers, and quickly, so they can finance
further expansion.
Wireless phone operators are pitching services to
anybody who can speak, and consumers are
responding. Cell phones are as mainstream as
cable TV and America Online. Some parents are giving their kids cell
phones when they start to drive, sometimes even earlier. At the park, I saw
a 10-year-old soccer player on the sidelines, cell phone in hand, describing
his big goal to dad.
You've Got Dead Air This increased load is changing the traffic patterns
for which these networks were designed. CTIA's numbers show that the
average call-holding time has held steady since 1988 at just under
two-and-a-half minutes. But flat-rate calling plans with bundled
minutes--not to mention all those teenagers--will drive up traffic and clog
the system. It's what America Online discovered when it introduced
all-you-can-eat pricing and the busy signal became its corporate anthem.
This phenomenon has been demonstrated in other wireless markets; for
instance, the story goes that parents in northern Europe began using their
cell phones as baby monitors once they lost the economic incentive to hang
up.
The increased number of users has made wireless telephone networks
volatile. Service is inconsistent and subject to wide variations based on your
location. To be fair to the operators, creating consistent service in a
general-purpose wireless telephone network is one of the most difficult
engineering problems. It requires an understanding of an ever-widening
target population's usage patterns, coupled with random clustering effects. It
may be obvious that you should build out additional capacity in Times
Square for Dec. 31, 1999, but there is no way to plan for hundreds of
reporters gathering around Monica Lewinsky's hotel on a day's notice. With
cell sites often costing about $500,000 each, optimizing these networks is
far from trivial.
Not only do wireless engineers have a substantial traffic-engineering
challenge, they need to support a wide variation of aging equipment. Most
networks are an amalgam of technologies emanating from a range of
suppliers, each with its own performance characteristics and technical
quirks. Sprint's network, touted as a unique ground-up deployment, was
built with Lucent's equipment in some areas, Nortel's in others and
Motorola's in others still.
One Size Fits All? Hardly For the foreseeable future, wireless telephones
will be a convenience, not an essential network component. Costs will
continue to fall in the effort to capture more users. The low prices may be
tempting, but corporate strategists should not adopt wireless phones for
critical functions, such as office lines, dispatch operations or even to replace
pagers. I certainly don't want to be a patient in a hospital where doctors are
summoned from the golf course via cell phone instead of by pager.
Recently in Washington, D.C., I was picked up by a taxi whose service was
being run via cellular. The company had replaced its SMR (Specialized
Mobile Radio) system with a common local digital offering. The cabbie told
me that some days he would give up on the dispatch system and search for
fares the old-fashioned way, at curbs and hotels.
Later that day at the ComNet trade show, getting anything but a fast busy
signal using AT&T's PCS service became next to impossible. (Did I
promise not to single out any vendor?) When a call actually rang through, I
felt like I'd won the D.C. Lottery--or at least the Scratch and Win. A friend
gloated about the superiority of his Cellular One digital service. But before
the day was out, he was lined up at the pay phones with the rest of us.
As the cab company learned, it pays to stick with what works. A cell phone
cannot yet replace a private radio system. It doesn't even offer the
messaging reliability of most pagers. These systems can't match the
remote-monitoring capabilities of proprietary telemetry networks from
companies such as Cellnet Data Systems, Itron or Williams Telemetry.
Partially due to AT&T's highly successful Digital One Rate flat-pricing plan,
the Wireless Services division has been a phenomenal success for AT&T,
accounting for almost half of its overall revenue growth in 1998. In my
experience, call quality is hardly close to that of a wireline service. In a
matter of seconds, the other party's voice breaks up, and the call gets
dropped. At other times, the network returns fast busy signals when I try to
dial out--this, too, is not an occasional annoyance; it happens all the time.
It's the Stupid Network, Stupid For 90 bucks a month, I expect more. So I
call AT&T's customer service, and the representative tells me that maybe
my dropped call was just one side or the other hanging up, and that's not
the network, that's the people using the phones. Most of us grasped the
on-hook concept some time around preschool, but this guy wants to make
sure I know that I'm hanging up when I press the End key. He also explains
that I won't be charged for those calls that are met with a fast busy, entirely
missing the point.
A business-grade service must first be reliable, and the hidden message in
these canned answers is that the service is not yet business-grade. AT&T's
own terms and conditions recognize that the service is subject to
transmission limitation and interruption, and state that the company isn't
liable for any outage of less than 24 hours. While other enterprise-class
carrier offerings move toward detailed service-level agreements, these
contracts specify almost no service at all.
If there's something good about the AT&T Wireless service, it's the Nokia
phones that are bundled with it. Incredibly, my Nokia 6160 has run for
three days without any external power, setting a record for incomplete calls
on a single charge. Fortunately, the phone has a display clock, so it still
serves a purpose no matter where I go.
But enough about AT&T. As I mentioned, it's not the only operator with
these problems. When I asked Pacific Bell about the dropped calls I
experienced while driving down Highway 101, the salesperson countered
that it really is best to pull over to the side of the road when making a call
anyway. Sprint PCS service didn't fare much better. My phone dropped
calls daily while I sat in my office.
Overall, public wireless phone systems continue to be the derelict sibling of
the telephony industry, lowering the bar for quality and serving as a
deflector for new entrants. Ask any voice-over-packet vendor about
service quality and you'll likely be told, "Well, at least it's better than a cell
phone." We should have much higher standards than this.
Send your comments on this column to David Willis at
dwillis@nwc.com.
networkcomputing.com:80/1007/1007colwillis.html |
