On Monday, June 17, 2002, at 05:14 , Benedict Evans wrote:
> While it is certainly true that PDC is a crappy technology (and indeed
> its
> inefficiency is a further impetus to DoCoMo's move to W-CDMA),
Far more important than the PDC technology itself was the implementation
of it. There is no reason why you could not build a PDC network with
traditional uplink downlink assignment and traditional frequency
separation in one of the traditional mobile paired spectrum bands.
PDC in the 1.5GHz band is implemented fairly straight forward and
doesn't suffer as much from the Swiss Cheese syndrom of PDC as
implemented in the 800/900 MHz band.
Although I have not seen anything published on this, I am convinced that
PDC handsets for the 1.5GHz band are far less complex and easier to
produce than those for the 800/900 MHz band.
> it has had an
> upside. Because all of the intelligence in the system is in the network,
> rather than the handsets, the handsets can be substantially smaller and
> hence have substantially better battery life.
With the notable exception that a PDC handset for the 800/900 MHz band
has to navigate through a patchwork of cluttered spectrum where no two
pieces have the same frequency separation. So depending on which channel
the handset is assigned by the base, first it has to cover a range of
almost 200 MHz and then it needs to find the corresponding paired
channel which may be 130 MHz away or 55 MHz etc etc etc. This does not
only require additional circuitry which no other system needs (ANSI-41
and GSM use a fixed separation of 45 MHz) but also it is far more
complex in terms of RF design.
> The downside is PDC can't really do roaming, *even if there were any
> other
> PDC networks*. But then Europeans can have an exagerated view of the
> importance of roaming. It is certainly essential in Europe, given
> national
> operators and frequent travel within the continent. In the US there are
> c.3,500 licence regions and anything up to 30,000 licence units, so
> roaming
> is important there (which makes the failure to mandate a technology
> standard
> even more misguided, but that's another argument).
Far more important than roaming is the aspect of achieving economies of
scale and a global market for handsets and other gear. This is where the
money goes. This is where cost savings come from. Also, as previously
pointed out, the independence from vendors. A network that is free to
buy new gear from any supplier with every new lot, clearly has a
competitive advantage over another network that is locked in to a
supplier and would have to change all their gear or buy customised
products if they wanted to change vendors.
I have been working on GSM roll outs in Eastern Europe and in South
America on ANSI-41 based systems. I can tell from experience how
staggering the difference is between a) attitude, b) customer service
and c) pricing of vendors who know that you could just give the contract
to one of their competitors and those who know you can't.
> But for Japan? How much
> revenue is being foregone? From the point of view of Japanese operators,
> what really matters is not how many Japanese go abroad, but how much
> roaming
> revenue they are losing from not having foreigners roaming onto *their*
> networks.
I wish that was true, because it would mean that the current roaming
paradigm (pay your home network) would have shifted to the paradigm I
have been propagating for years (pay the local network).
In fact, most networks earn far more revenue from outbound roaming than
they earn from inbound roaming. The revenue potential from roaming for
Japanese networks under the current settlement regime is tremendous. And
that is due to the large number of outbound travellers.
> And I suspect the answer (this summer apart) is 'not much'.
Inbound tourism to Japan is about 4.5 million visitors per year, the
world cup is expected to increase this figure by about a 100K or 200K or
so.
Although inbound roaming potential isn't as interesting as outbound,
there are a few bucks to make, particularly because many travellers to
Japan spend more money than travellers to most other countries.
In any event, roaming is not the main issue.
> As a further observation on the current superiority of Japanese
> handsets,
> manufacturing capacity of colour LCDs is a big issue. There will only be
> 50m-odd handset-sized colour LCDs manufactured this year, up
> substantially
> from last year - of these 30m will stay in South Korea and Japan. This
> is
> not mature technology, and yield rates, while acceptable if you're
> manufacturing for the Japanese domestic market, have not hitherto been
> high
> enough to support the volumes that would be needed to put a 120x160
> colour
> TFT in every handset sold in Europe.
agreed.
> Looking at data services, it is undeniably true that mobile data
> happened in
> a smoother and more coherent manner because of DoCoMo's market power.
Which doesn't mean it couldn't have happened on top of CDMA or GSM or in
a mixed environment. After all it is a value added service.
> DoCoMo adopted a flawed network technology (albeit one with some
> incidental
> benefits, as mentioned above). It would probably have been better for
> Japanese consumers if DoCoMo (and the others) had adopted GSM.
I believe the Japanese way would have been to have CDMA, GSM and PHS,
possible even PDC. This is because Japan is a value added technology
exporting economy. The Japanese market is clearly large enough to
support this and the Japanese industry would have been able to ascertain
global market leadership in a variety of value added technologies. No
matter what the Japanese would have come up with, it would already have
been working on both the European and the American standard. WAP would
have never happened - European networks with cross Atlantic partnerships
and alliances would have said "We use the Japanese solution because it
works also across the Atlantic".
> If it had,
> Japan would have larger but cheaper handsets (still with colour
> screens),
How much larger/heavier are we talking here ? 10 grams ? 20 grams ?
Sure, it is not that 10 or 20 grams isn't noticeable. But then again, it
is not *that* noticeable.
Give me a break.
As for longer batter lifetime, again, how much are we talking ?
Last time in HK I bought a new NiMH battery for my Nokia GSM (8110)
phone. While in HK where there is a high density of base stations, the
battery lasted as long as any phone I have used in Tokyo. In Australia
it was notably less, but so was the experience with an earlier phone on
J-Phone when I spent the weekend in Hakone.
Then again, how come my KDDI au phone (CDMA) is just as tiny and light
and has just as much battery life than any of the DoCoMo handsets
(PDC) ? If it was because of PDC then clearly even Japanese CDMA phones
should be significantly heavier and use significantly more power, but
somehow they don't. What happened ?
It is not that the fundamental differences between the phones are that
big. In fact when comparing bananas to bananas and apples to apples it
is pretty close.
> and call charges would be substantially lower, both because the network
> equipnment would be cheaper, and because the greater network efficiency
> would mean that operators were not obliged to constrain usage through
> prices
> to quite the same degree.
agreed.
> However, this argument has nothing to do with
> DoCoMo's creation of a coherent, smoothly functioning and pretty-much
> open
> mobile data services model.
Which doesn't preclude that this or any similar functioning model could
not have been built on top of CDMA, GSM, PHS or any mixed environment.
kind regards
benjamin
Received on Mon Jun 17 14:19:14 2002