Re: Daily Yomiuri editorial, 5/17/94

[David Farber <farber () central cis upenn edu>]

From: Max Morris <mgmorris () flab fujitsu co jp>


[Bruce quotes from the Yomiuri's editorial on research nets.]


Switching IMHO mode on full:


My impression is that WIDE is viewed in many circles as a research
network -- and only that.  I believe that it is still officially an
experiment, though clearly it is much more than that in practice.
Certainly, it doesn't compare well to the combined Nets in the States
(while it compares favorably to some Nets elsewhere).  But Japanese
companies and institutions haven't been standing still, as the article
implies.  Many companies, universities, and research organizations
have high-class machines connected to sophisticated private networks
(most running TCP/IP) -- and they have plenty of experience using and
managing it all.  An independent user community exists, and though
these folks mainly use phone lines and attendant protocols, I wouldn't
say they're backwards in their experience by any means.  WIDE serves
as the main (or most famous) way to connect those nets together,
despite its being a bottleneck.  In frustration, to deal with the
bottleneck, institutions limit traffic at the source (we're behind a
capped-access firewall here, for example), allow users to fight it out
and suffer heavy congestion, or simply hook up private lines.


I think institutions are interested in connecting their advanced
private networks together, because they see the growth and benefits
that are resulting in areas where the Internet really interconnects.
That means databases and other info coming online (for free or for
cost).  If the government decides to help out (as it seems it might),
the companies will be pleased and society could surely benefit.  But
meanwhile, there are other approaches to interconnect the nets on the
drawing boards -- look at AT&T/Jens, PSI, Fenix, IIJ, JCIX,
Tepco/Tokyu/Mitsu{i,bishi}, others we know and don't know about.  Some
of these efforts are fledgling; some may be a joke; some may take off.
And the list continues to expand, as the fact that our ISP FAQ is
outdated demonstrates.  The protocol convergence that's happened in
the States needs to happen here.


The fear, I think -- and here I also hesitate over the phrase "Japan
must approach this task with a grand design" -- is that the government
(and it's star, NTT) might seek to stifle the budding independent
Internet-building efforts through heavy regulation.  It would do so in
favor of solely the lauded (and laudable, if that's all it were)
research network that's being considered.  Somehow, I think such an
effort would be as successful as any attempt by the US government to
contain the CIX with NII might be: there's just no logic in it, and it
wouldn't work anyway.  (I think eventually some sort of coherent
regulatory structure will come to the Net in the States and elsewhere,
instead of the hodge-podge we have now, but that's another issue.)
Deregulate the market and promote an NII/NSFNet like alternative?
That sounds fairly sensible, affordable, reasonable, etc.


So, I don't view the state of the Net in Japan as all so backward as
it's hyped to be -- certainly not ten years backward.  Sure, there's a
lot of work to be done, especially political, as there are many
barriers to entry relative to the ISP market in the States.  What the
government will do with WIDE, how the Monbusho will approach this
change in culture regarding its experiment, how MPT and MITI will also
play their cards, and so on: these are the questsions that will decide
who will make how much money when.  But the technology is waiting in
the wings.


The appearance that networking in Japan is somehow way out of step
with the world, and that it might stay that way, is just that: an
appearance.  And as the maxim goes -- and it surely applies in Japan
if it applies anywhere -- appearances can be deceiving.


Never underestimate.


Back to lurking a bit...


---------------------------------------------------------------------------
Max Morris (MIT Japan Program)     ------ Kamikodanaka 1015, Nakahara
Multimedia Systems Laboratory     ------ Kawasaki 211, Japan
Fujitsu Laboratories, Ltd.       ------ (tel) +81 (44) 754-2627
Kawasaki, Japan                 ------ (fax) +81 (44) 754-2570
mgmorris () flab fujitsu co jp    --------------------------------------------
------------------------------------- These are my opinions, not Fujitsu's.