On Fri, 09 Nov 2007 12:30:27 -0500, Randy Yates <
[email protected]>
wrote:
>This may be a stupid question, but when WiMax gets popular, so that a
>cell is within range in 98 percent of the populated areas, why would
>folks continue to use the cellular networks (GSM, etc.) for voice
>phonecalls?
Oh, man, don't get me started... ;)
*If* WiMax gets popular, and I think that's a pretty substantial "if",
it has limited capacity, just like any system. Consider the number of
cell towers that are currently deployed, and consider the effect that
it would have on the data carrying capacity of a WiMAX (or similar,
e.g., LTE) system to have to carry all that voice traffic capacity. A
device made for WiMAX to be able to carry data at a reasonably good
rate (say, a couple of megabits) won't necessarily be as efficient at
handling voice traffic as a device designed specifically to do that.
A nice thing about the cellular systems is that they're optimized to
do one thing really well. It makes scheduling, power management, and
all sorts of things, much easier when the assumptions about data rate,
etc., are within a very narrow range. Voice traffic places a nearly
fixed resource demand on the system, and that demand is reasonably
constant over the duration of the connection. The requirements to
carry voice traffic are much narrower than the requirements to carry
data at rates that vary over a wide range and for which the demand is
not predictable. That makes scheduling as well as baseband/radio
management (and engineering) a much harder problem for a WiMAX-like
sytstem and strains the compromises much harder than necessary when
carrying primarily voice traffic. Cellular signalling is reasonably
well optimized to reduce power consumption at the user terminal, and
while WiMAX takes that into consideration as well, the scales and
assumptions are quite different. Expecting a voice handset to have
the same battery life using WiMAX is a bit of a stretch, and it is
likely that the cost of the WiMAX handset would be higher as well.
A simple example point is the modulation waveform: Cellular systems
use low-PAPR waveforms in the uplink channel so that low-cost, high
efficiency PAs can be used in the handset. WiMAX uses OFDM/OFDMA.
Doh.
Similarly, if the idea of WiMAX is to provide a fairly healthy data
flow rate to users on demand, then expecting the system to also carry
a lot of voice traffic (i.e., enough to replace existing cellular
systems) may cut into that efficiency fairly substantially. It can be
done, and handling VoIP on a WiMAX system is well within it's
capabilities, I just don't think it's nearly as efficient.
The promises being made about WiMAX or similar systems are creating
expectations that are going to lead up to a lot of disappointed
people, IMHO. Meanwhile EVDO, HSPA, and LTE are either deployed or
waiting in the wings, and generally enjoy more support from the major
carriers than WiMAX.
WiMAX has enjoyed a lot of hype, much of it generated by Intel. Intel
isn't exactly a successful communications company, especially in the
last few years. Also consider that EVDO/HSPA/LTE were all developed
as organic extensions to successful networks and were designed by the
same people/companies that created those successful networks. WiMAX,
on the other hand, was a life-support effort tacked on to a
fixed-wireless IEEE WAN standard that has had little historical
success in comparison.
The real questions center around what people will really need and be
willing to pay for in the future as far as connectivity is concerned.
User models and user behaviors are hard to predict and anticipate, and
often don't work out anywhere near what was expected. A lot of
people do like the existing EVDO systems and they seem to work pretty
well, especially for people who don't have time to find or stop at a
WiFi hotspot. Personally, my mobile connectivity requirements are
fairly sporadic and I don't have a problem either doing without or
using a WiFi hotspot. Since a decent WiFi hotspot can easily be
expected to provide a much higher throughput than the expectations for
WiMAX, I find it a little difficult to see that a lot of people will
be willing to shell out $30-$50/month (or whatever it'll wind up
being) for something that isn't as good as what they might find for
free (or less money) if they just look around a little bit. Some
people will, naturally, sign up for WiMAX just like a fair number of
people are using EVDO. Whether or not there's enough market to
justify building out WiMAX and/or LTE remains to be seen, and I'm very
skeptical.
Eric Jacobsen
Minister of Algorithms
Abineau Communications
http://www.ericjacobsen.org