My recent article Life on the Road
with 3G sparked a little bit of controversy with some readers. I expected
that some people would take issue with some of my gripes about various phone
services and the qualities of my particular phone, but by far the most
controversial part of my article was contained in the title and the first
paragraph. In my article's title, I refer to SprintPCS's Vision service "3G," and in
the first paragraph I admit that it's really "2.5G." Respondents in the forums
on NewMobileComputing and at Slashdot, who linked to the article, and a few
who emailed me, seemed to come at me from both sides: most criticized me for
calling the Vision service 3G when it really isn't, and a few got on my case
for calling it 2.5G when it's really, clearly 3G. I decided to do a little
research on what exactly constitutes 3G, and who's really in charge of
deciding what's 3G and what isn't, and the more I dug up, the more confused I
became. I decided that I was going to get to the bottom of it once and for
all. First things first, the standards body that is the ultimate arbiter of
what's 3G and what isn't is the ITU [http://www.itu.int/home/index.html] (International
Telecommunications Union). In 1999, the ITU approved an industry standard for
third generation (3G) wireless networks. For the uninitiated, the first
generation was analog cellular voice networks, 2G was the replacement of those
with a digital signal (usually TDMA, GSM, or CDMA), and the third generation
is networks capable of high speed data transfer. The standard that the ITU
approved in 1999 was called International Mobile Telecommunications-2000
(IMT-2000), and it specified the following bandwidth requirements:
So that all seems clear enough; why the controversy? A lot of it swirls around the term "2.5G," which it seems was coined
because wireless operators started developing low-bandwidth data services and
calling them 3G, hoping to capitalize on the hype. This was most apparent with
GPRS, or Global Packet Radio Service, which is available to many GSM users
today. When people started hearing vendors refer to GPRS as 3G, they cried
foul, and eventually the term 2.5G was applied to these services. GPRS shares
some similarity to CDMA2000 in that it's packet-based and has always-on
capability. 2G networks are circuit-switched and require a dial-up to data
services. But alas, GPRS only allows for a real-world bandwidth of 30-40 kbps.
It's not really that simple, though. Depending on whom you ask, CDMA2000 1x
may or may not really be capable of the speeds required by the ITU to be
called 3G. Whether anyone has actually been able to get 384 kbps in an urban
environment or 2 Mbps with a fixed antenna is up for debate. The carriers are
so eager to achieve 3G and capitalize on the hype because there's potentially
a lot of money involved, but nobody has more to gain that Qualcomm, the
company that provides software and chips for 3G networks and owns most of the
intellectual property. That company, and the CDMA Development Group
[http://www.cdg.org/index.asp] (which is sponsored by Qualcomm, among others)
is the primary booster behind calling CDMA200 1x 3G. To sidestep the
controversy, some have started calling CDMA2000 1x "2.75G." Comparing bandwidth is actually beside the issue as far as real-world use
is concerned. Good bandwidth is nice, but the real killer for real-world use
is latency. Analog modems and satellite-dish internet are bad because they
have high latency, even if their bandwidth is acceptable or spectacular.
Latency, in a nutshell, is the lag between each bit of data that is
transferred on a network. Ethernet, DSL, ISDN, and Cable modems have good
(low) latency, so that's why they feel faster than a dial-up modem, even if
their bandwidth isn't that much higher, like a slow ISDN or DSL connection.
GPRS has terrible latency
[http://www.sourceo2.com/O2_Developers/O2_technologies/GPRS/Technical_overview/gprs_tech_reliability_latency_jitter.htm]
for most users, over 500 ms, even as bad as 4 seconds for some. DSL has a
latency of 20 ms or so. CDMA2000 1x will deliver a latency of 100-300ms.
That's roughly what you'll get from an analog modem connection. In my tests,
I've found CDMA2000 1x to be about 3-5 times faster (bandwidth) than dial-up,
with comparable latency. Here's the overview of the 3G landscape: The 2G phone landscape has come
down to two camps, GSM and CDMA. Because it's prohibitively expensive to
upgrade GSM to 3G (new equipment, software, royalties, and spectrum allocation
are needed), the GSM vendors have worked up two overlays to the GSM network to
provide packet-based data, GPRS and EDGE. GPRS has been available to most GSM
users for a couple of years, and EDGE is being rolled out now. EDGE is
theoretically faster than GPRS and CDMA2000 1x, though in real-world terms it
will probably be a bit better than GPRS, but perhaps not faster than CDMA2000
1x. The 3G successor to GSM is WCDMA, which is supposed to be great, though
very expensive to implement. From what I've seen, real-world performance of
WCDMA and EDGE may fall far below expectations. CDMA 2G networks were upgraded
to CDMA2000 1x networks. The faster CDMA2000 1xEV-DO 3G network is a data only
network currently in use by a handful of operators. The CDMA2000 1xEV-DV
standard is the successor to CDMA2000 1x, and it's clear into 3G territory. It
will also be rather expensive for CDMA operators to do the upgrade to 1xEV-DV,
though it won't mean a whole new infrastructure like WCDMA from GSM will. Like
CDMA2000 1x and every other 2.5G and 3G network, 1xEV-DV will probably be
slower than advertised. There is, of course, a raging ideological battle between the GSM camp (in
Europe and most of the rest of the world) and the CDMA camp (mostly US and
Korea). GSM has wide adoption and standardization on its side, while most
concede that CDMA has technological superiority, lower cost of implementation,
and an easier upgrade path to faster 3G capability. GSM proponents downplay
their disadvantages (that, frankly, don't matter at all to the average user)
and work on kludging together workable, "fast enough" data services on their
existing networks. They thumb their noses at Americans who can't use their
phones in Europe, to say, "where's all your superior technology now?" This is just my unscientific observation, but it seems that the people who
are the most up-in-arms about the CDMA Development Group and Qualcomm pushing
so hard to identify CDMA2000 1x as 3G are the promoters and users of GSM and
GPRS. I think there's a bit of sour grapes going on here. Heavy GPRS users,
who are probably a little dissatisfied with the performance of their system,
want to cast aspersions on a competing system that is not available to them.
Nevertheless, both sides are heading toward roughly the same place, 3G
Nirvana, and Qualcomm controls much of the technology that's likely to be used
there, by carriers in both camps (CDMA2000 and WCDMA). Qualcomm wants to be
perceived by the general public and the wireless carriers alike as the
standard bearer for 3G, but it also stands to gain more from the adoption of
CDMA2000. Let us contemplate the possibility that the CDMA2000 1x technology
may have, in some parallel universe, failed to live up to its full potential
in real-world performance. Would it be possible, in that case, that Qualcomm
and the CDMA Development Group would exert its influence over the carriers,
the trade press, the general public, and even the ITU to sweep that little
complication under the carpet and go ahead and call it 3G, as planned? Just an
idle conspiracy theory. Besides SprintPCS, about 35 other companies have also set up CDMA2000 1x
networks around the world, including telecom giant Verizon in the USA. NTT
DoCoMo has a WCDMA network working in Japan that's been up for a while, and
AT&T and NTT DoCoMo are in the process of setting up one in several large
metropolitan areas in the USA. WCDMA has also been deployed by Hunchinson in
the UK, Italy, and Australia. There are a few CDMA2000 1xEV-DO networks. To
see the companies that have 3G networks set up, here is a list
[http://www.3gtoday.com/operators_flash.html]. The fact that there has been so much exaggeration and outright
misinformation about the cost, ease of implementation, and performance of
wireless data systems made by the equipment and software vendors to the
network operators may have severely damaged the wireless telecommunications
industry. Customer disappointment in current services makes them wary of
spending money or becoming enthusiastic about the "next big thing" and the
carriers' enthusiasm for buying new spectrum for their 3G networks from
governments caused financial ruin a few years ago. Some CDMA boosters
speculate that the GSM equipment vendors' claims of GPRS and WCDMA performance
prompted them to make unwise decisions to stick with GSM. On top of this,
Wi-Fi, which promised little and has delivered much, will end up sapping much
of the market for 3G data services in urban areas. In conclusion, SprintPCS' service may not be quite as fast in real-world
terms as it needs to be to really be 3G in the strictest sense, but
"officially," it is 3G, as CDMA2000 1x has been anointed such by the ITU.
Every wireless data solution has fallen short of its advertised performance
claims, and CDMA200 1x seems to have hit its mark closer than the various GSM
solutions, which deflates the naysayers a bit, since they mostly seem to have
some stake in GSM/GPRS/EDGE/WCDMA. Especially if you take latency into
consideration, SprintPCS Vision is quite a bit better than GPRS, the poster
child for 2.5G. I haven't covered cost at all in this article, but I should
also remind everyone that I'm paying $10/month for unlimited data, which is
pretty darn cheap. Faster 3G service has been dealt a series of blows by dishonest vendors,
gullible operators, cheerleading trade press, apathetic customers, and new
wireless networking technologies that will erode 3G's market. Faster service
will come, probably at a cost that exceeds its value, but my conclusion is
that what I have right now is pretty good. My main gripes now are with the
devices, not the network. I'm still waiting for a small phone/small PDA combo,
linked via short-range wireless network, that allows me to make a call on a
wireless earpiece and browse the 'Net with my PDA (on a fast, full-featured
browser), all withouth taking the phone out of my pocket.
( Original Story URL at
http://www.newmobilecomputing.com/story.php?news_id=3091 )
There are 5 operating
modes approved in the IMT-2000, and the most widely used ones are CDMA2000 and
WCDMA. CDMA2000 1x is what the SprintPCS Vision service that I've been using
is based on, and its theoretical limits just barely make the cut to be called
3G. In real life, I've achieved speed close enough to 144 kbps with my Samsung
a500 and laptop that I'm willing to give CDMA200 1x the benefit of the doubt.
You can read more details about IMT-2000 and get details about where 3G
networks have already been rolled out at 3G Today
[http://www.3gtoday.com/standard.html] (Qualcomm-sponsored site).
| Copyright
NewMobileComputing 2003. All Rights Reserved. NewMobileComputing and the
NewMobileComputing logo are trademarks. The readers' comments are owned by whoever posted them. We are not responsible for them in any way. All trademarks, icons, and logos, shown or mentioned in this web site, are the property of their respective owners. Reproduction of NewMobileComputing stories is premitted only by explicitly receiving authorization from NewMobileComputing. Privacy statement - Notice to Bulk Emailers  |