Curt, thanks for your clarification.....
I simply could not see what you were getting at in the first post.

see comments below........

On Sun, 30 Jun 1996, Curt A. Monash wrote:

> At 05:31 AM 6/30/96 -0400, xerxes wrote:
> >In the US, the RBOCs and LD telcos (distinctions will blur in the future)
> >have complex regulatory relationships with the FCC and State Public Utility
> >Commissions that must be kept in good repair.  You seem to suggest special
> >deals for higher qualty access limited to web sites owners with special,
> >technology-driven, relationships with one or more of the telcos.......this
> >treatment could generate tremendous political heat, ...and is the reason
> >for the MANY stringent common carrier provisions....... the communications
> >bar doubts that legislatures will permit the oligopolistic, differential
> >treatment you suggest----but who knows.
> >
> 
> --I don't see why.  

What, precisely, do you not see?  My point was about equallity of access
to the market.....not about tiers of service open to all.  


Paying for a T-3 connection is really just the same
> thing. 

My point is that all customers for a T-3 have to pay the same thing, i.e.
no side deals.  That is what I thought you seemed to be talking
about....but were you?  There cannot be backdoor deals froma single
provider in a single market which will allow one customer to get a better
deal for a T-3, or anything else, than any other customer.  Of course
different providers can offer different prices... 8-)

 But thanks for pointing out this constraint on industry evolution.
> 
> >At 12:54 AM 6/29/96, Curt A. Monash wrote:
> >
> >>Our story goes like this:
> >>
> >>Bandwidth is subject to "The Tragedy of the Unregulated Commons".  The
> >>first/best fix will be for large ISPs to emerge and take responsibility for
> >>end-to-end quality and reliability of connections.
> >
> >Yeah, it appears that Internic has fallen down on the job......bring in the
> pros
> 
> --You got me here.  I'm not sure what Internic does except pass our domain
> names, and that's not what I'm referring to.

Well, if the DNS servers are unable to cope with the load of recent
demand, as was alleged on this list several days ago, AND, the bandwidth
of the backbone  is not the problem, (also here alledged) then, the
organization responsible for the DNS system would be responsible for the
bottleneck....the perceived lack of bandwidth.  The pipes are open, it is
the routers, alledgedly, that are insufficent to demand.

> >> (We presume these ISPs
> >>will in many cases be today's telecom behemoths, such as AT&T, MCI, British
> >>Telecom, et al. -- I think AOL will be one too, but probably after
> >>acquisition by a telephone company.  Dunno about Netcom.)
> >
> >Everytime AOL's stock price drops, it becomes a more likely acquisition....
> >
> >>
> >>This suggests that the Web sites to which the high-quality access is
> >>provided must themselves have deals of some sorts with the leading ISPs.
> >>(Not unthinkable, if only a few ISPs wind up mattering.)
> >
> >Stream of logic breaks down a bit here, Curt.
> 
> --It's based on the technical argument you seem to have dismissed.

I have not dismissed your case.  As of yet, you have not made a
sufficiently clear case that I can unravel.  How bout some more detail,
ok? Unpack the shorthand.

> >
> >>One technical
> >>reason is that caching will stop working.  Why?  The old dynamic/static
> >>distinction.  If every Web page is dynamically generated "on-the-fly",
> >>there's nothing to cache.  What replaces caching?  Maybe database
> >>replication; certainly DBMS vendors and hangers-on (me included) are
> >>salivating about that possibility.

I am sure you have a fine point here,.....I'd just enjoy a bit more
specificity.  For example, which parties are replicating whose databases
in the paragraph above?

> >>Note that this leads to a self-fulfilling circular arrangement, in which the
> >>high-traffic sites get superior access quality, leading to even higher
> >>"market share" (whatever that means), vs. the gazillions of
> >>contenders/pretenders.

This is real cryptic.  Why assume that a high traffic site "gets" better
quality access?  Server owners will get the quality of access they pay
for, independent of traffic.  No?  Are you instead making an economic, not
a technical case, that servers with high traffic, can garner more
advertising revenue, become wealthier, and thereby afford better
bandwidth?  Well, if you cannot afford bandwidth, which continues dropping 
in price in most venues, then you are not much of a contender/pretender.

> >Not necessarily.  Attend a PUC hearing sometime.  The NY ones get pretty
> >darn acrimonius.  If the current common carrier provisions do not
> >adequately pertain to the backbone systems yet,  .... ask yourself how long
> >this will take once the perception of special deals reaches the press.....
> >
> >>  It's not the truly personal or otherwise quirky
> >>sites that go away; people will not much care about bandwidth in deciding
> >>whether they care about those.  It's the "secondary" commerical sites which
> >>lose; i.e., the me-toos.  In other words, the rich get richer, and the
> >>runners-up evaporate (no surprise there).
> >
> >The runners-up have trade association that hire Reed Hundt's future law
> >firm, and they litigate
> >
> >
> Hmm.  Maybe I should try to spell out the technical case again.

Yes, please.

> Right now, today, ISPs provide differentiated access to popular sites.  

Some ISPs do this....not all, right?  Clark.net I doubt does this.

> How?
> By caching the pages most often retrieved.  This requires no discrimination,
> no choice, no volition -- just count accesses, and cache the pages that get
> hit a lot.

Sure, this assists in providing an acceptable speed for users of large
ISPs.......

> However, this level playing field MUST change, once the most popular Web
> pages are dynamically generated on-the-fly.  In other words, technology
> mandates that the connectivity suppliers actively choose which Web sites to
> give the "best" access to.  

Ok, I see what you are saying.  Some "connectivity suppliers" may in fact
do this. But not all, eh?  Because if these suppliers are in the regulated
industry, the telcos, they are open to serious problems from the courts
and the regulators, .....unless they can clarify the issue beforehand. 
A differential quality of access related purely to demand may be
defensible, if it is not arranged for by compensation to the
provider,.....like Open Text has been reported as doing in a similar
situation, no?  

> On what basis will the ISPs make these choices?  Well,  there are actually
> TWO ways they can give differentiated access to specific Web sites.  One, as
> outlined above, is by son-of-caching; the other, of course, is by provision
> of prominent links (and perhaps even more explicit recommendations) to
> favored Web sites.

The latter has a longer history of acceptability. 

> The second point is, I presume, pretty well understood; just think of AOL.
> So let me loop back to the first point, which isn't.  In two words, the key
> is DATABASE REPLICATION.  The more copies you have, in more convenient
> places, the shorter the access time.  

Yep, I see now.  

> Now, the issue is a bit confused by a stupid debate as to what form this
> replication will take.  Much as IBM/Lotus tries to muddy the waters, the
> fact of the matter is that relational databases have all the "bidirectional"
> replication functionality anybody would seriously need, or ask for.  What's
> more, I think most interesting Web sites will be generated by RDBMS or their
> UDBMS successors, rather than "flat files".  Shoehorning the STO archives
> into a Sybase server was just an early (and surprisingly extreme, if
> anything) form of that.  

what is the U in UDBMS again?   STO archive?  

> Anyhow, even if you believe the Lotus/Landry/Ozzie Notes-still-wins tripe,

an industry debate I do not follow...........but not necessary to your
thesis, right?

> one way or another major Web site owners are going to replicate their sites,
> physically, across multiple computers.  As an educated guess (but admittedly
> just a guess), I suspect by far the best way to do this is to have the
> replication targets strategically located at prime locations in various ISPs
> networks.  And that's the technological tie I was talking about.

ISP networks or backbone service providers, like Sprint or Alternet?
Little ISPs will not go this far....nor will web owners pay them to do so,
eh?  You assume the smaller ISPs will disappear......but they may not, at
least not all of them.

> 
> As for the financial arrangements associated with those technological links
> -- I don't see why money WOULDN'T change hands.

I see no reason why differential service for differential revenue will not
happen, so long as the better service is available, across the board, to
anyone who pays for it.  The telcos cannot refuse to sell better access to
a customer who can pay, or cut an under table deal for a differential
price plus other consideration.......  this I believe would cause major
controversy.  Certainly this is what Malone and the Cable industry is more
used to doing..... 

The issue becomes far more important when you go to video or 3D, or other
advanced services web sites.  Tell me, do you see any reason why the owner
of the REPLICATED DATABASE (not the original), might accumulate or
otherwise reproduce that replicant, rather than being paid by the dbase
owner?  If accumulated, then the better service might be sold to
end-users, rather than sold to dbase owners, no?  Does anything prevent
this?  Are the web owners protected from competition by the BIG
connectivity suppliers, or do they take their chances?  The phone
companies do not generally compete with the firms that buy their
YellowPages ads....but in an info context....you could see that in your
example above, no?


+---------------------------------------------------------------------------+
This message was posted to ONLINE-NEWS. http://www.planetarynews.com/o-n.html

------------------------------

End of online-news-digest V1 #712
*********************************


From owner-online-news-digest@marketplace.com Mon Jul  1 00:12:59 1996
Received: from marketplace.com (majordom@marketplace.com [206.168.5.232]) by cnj.digex.net (8.6.12/8.6.12) with ESMTP id AAA21468 ; for ; Mon, 1 Jul 1996 00:12:58 -0400