🚀 go-pugleaf

On Tue, 23 Feb 1999 04:22:12 GMT, Usenet@NavasGrp.Dublin.CA.US (John Navas)
wrote:

><http://www.internetnews.com/isp-news/article/0,1087,8_71391,00.html>


Great article, and glad to see that you got mentioned.  I always enjoy
reading your posts.

mark

In article <36db5a0c.344542958@news>, markr@mbrnet.com says...
> On Tue, 23 Feb 1999 04:22:12 GMT, Usenet@NavasGrp.Dublin.CA.US (John Navas)
> wrote:
> 
> ><http://www.internetnews.com/isp-news/article/0,1087,8_71391,00.html>
> 
> 
> Great article, and glad to see that you got mentioned.  I always enjoy
> reading your posts.
> 
> mark
> 
> 

Ha!  You should see what our local Road Runner "Acceptable User Policy" 
is trying to pull off.  Probably not as bad as TCI, but it looks as if 
the cable modem "industry" is starting to become Big Brother.

--Kevin

In article <MPG.114ca9502d8f0ea9896c5@news3.newscene.com>,
hurnik@albany.net (Kevin Hurni) wrote:

> Ha!  You should see what our local Road Runner "Acceptable User Policy" 
> is trying to pull off.  Probably not as bad as TCI, but it looks as if 
> the cable modem "industry" is starting to become Big Brother.

I have been talking to some heavy-duty players in the provider world and
it is generally agreed that in the dedicated-connection arena, we are
headed for data-quantity pricing. With high-speed connections becoming the
norm, the "natural bandwidth limitations" are disappearing, and providers
are going to have to ultimately ration the amount of data that passes
through their limited facilities.

One would be to simply restrict flow, but that would result in much of the
chatter we see here about "slow connections" on high-speed lines. Another,
waiting in the wings, is the "per-megabyte" charge for in-bound,
out-bound, or both. Rather than have the provider become Big Brother and
dictate the types of equipment and services allowed on customers' sites,
they would simply have the users pay for data excesses. If someone began
streaming audio or video from a server, he would start paying for the
privilege.

Let the pooh-poohing begin, but mark my words, it is coming.

-- 
John Higdon  |    P.O. Box 7648   | +1 408 264 4115  |      FAX:
             | San Jose, CA 95150 | +1 500 FOR-A-MOO |+1 408 264 4407
         abuse@ati.com            |        http://www.ati.com 

abuse@bovine.ati.com (John Higdon) writes:

> I have been talking to some heavy-duty players in the provider world and
> it is generally agreed that in the dedicated-connection arena, we are
> headed for data-quantity pricing. With high-speed connections becoming the
> norm, the "natural bandwidth limitations" are disappearing, and providers
> are going to have to ultimately ration the amount of data that passes
> through their limited facilities.
> 
> One would be to simply restrict flow, but that would result in much of the
> chatter we see here about "slow connections" on high-speed lines. Another,
> waiting in the wings, is the "per-megabyte" charge for in-bound,
> out-bound, or both. Rather than have the provider become Big Brother and
> dictate the types of equipment and services allowed on customers' sites,
> they would simply have the users pay for data excesses. If someone began
> streaming audio or video from a server, he would start paying for the
> privilege.
> 
> Let the pooh-poohing begin, but mark my words, it is coming.

	Foo. I'm having X.25 flashbacks.

	Anyone here remember the "kilosegment" charges? Which were
a pain to keep track of sometimes, because the echo was provided by
the nearest switch in some cases, as opposed to remote hosts. Which
in turn made some full screen editors (like EDT) unbearable to use.

	Anyways, if this metering start, It Will Suck. PSI used
to tout the flat-fee, no per-packet charges as a Good Thing.

-jav

In article <36e36027.39055325@news.ncal.verio.com>,
Usenet@NavasGrp.Dublin.CA.US wrote:

> p.s.  The song being sung by the "heavy-duty players in the provider world"
> in really no different today from last year, the year before, etc., and in
> my opinion is mostly wishful thinking and/or trial balloons.

But up to now, it was not a real issue. Relatively slow connections
provided a natural data flow limitation. With more and more customers
demanding megabit speeds, the possibility for massive throughput is
exponentially higher.

To wit: for years it has been common for companies to connect to a
provider via T1, but only pay 56K rates based upon actual data volume
transmitted. The T1 provided burst speed, but the agreement was that the
sum total of data would not exceed that which could fit in a 56K pipe. If
the company went over that amount, it got surcharged for the extra
effective bandwidth.

Now that consumer connections approach (and in some cases exceed) the
speed of that T1, those same arrangements will of necessity find their way
down to the consumer level. Else what would stop anyone from putting a
nice Sparc10 at home and serving a mondo website? 

Are consumers going to want bandwidth charges, or would they prefer
restrictions on types of equipment to be served by the connection?
"Neither" is not a workable answer.

-- 
John Higdon  |    P.O. Box 7648   | +1 408 264 4115  |      FAX:
             | San Jose, CA 95150 | +1 500 FOR-A-MOO |+1 408 264 4407
         abuse@ati.com            |        http://www.ati.com 

[POSTED TO ba.internet]
abuse@bovine.ati.com (John Higdon) wrote:

>In article <MPG.114ca9502d8f0ea9896c5@news3.newscene.com>,
>hurnik@albany.net (Kevin Hurni) wrote:
>
>> Ha!  You should see what our local Road Runner "Acceptable User Policy" 
>> is trying to pull off.  Probably not as bad as TCI, but it looks as if 
>> the cable modem "industry" is starting to become Big Brother.
>
>I have been talking to some heavy-duty players in the provider world and
>it is generally agreed that in the dedicated-connection arena, we are
>headed for data-quantity pricing. With high-speed connections becoming the
>norm, the "natural bandwidth limitations" are disappearing, and providers
>are going to have to ultimately ration the amount of data that passes
>through their limited facilities.
>
>One would be to simply restrict flow, but that would result in much of the
>chatter we see here about "slow connections" on high-speed lines. Another,
>waiting in the wings, is the "per-megabyte" charge for in-bound,
>out-bound, or both. Rather than have the provider become Big Brother and
>dictate the types of equipment and services allowed on customers' sites,
>they would simply have the users pay for data excesses. If someone began
>streaming audio or video from a server, he would start paying for the
>privilege.
>
>Let the pooh-poohing begin, but mark my words, it is coming.

If downstream metering is tried, I predict an enormous backlash against
bloated websites.  People aren't going to be thrilled at the notion of
paying to download ads and fancy graphics.  As a result, I doubt that any
such metering will stick.  Regardless, I personally don't see any real
issue, since sophisticated ISP's should be able to solve most bandwidth
problems by caching on their own networks.  The obvious consequence is that
it will become even harder for smaller ISP's to stay in business.

I do agree that upstream metering will become more common -- it's already a
fact of life at most web hosting services.  But that's an issue on the
server, not the client, side.

p.s.  The song being sung by the "heavy-duty players in the provider world"
in really no different today from last year, the year before, etc., and in
my opinion is mostly wishful thinking and/or trial balloons.

-- 
Best regards,
John  mailto:Usenet@NavasGrp.Dublin.CA.US  http://www.aimnet.com/~jnavas/
FIGHT EXCESSIVE DOMAIN NAME FEES: <http://www.aimnet.com/~jnavas/namefee.html>

I would be quite happy with bandwith limits. Most web sites come up as fast on my
ISDN line as they do my cable modem. Yeah, downloads are considerably slower, but
still workable. They still charge for Frame Relay based on pipe size-there is no
reason that FR's pricing model can't be applied to DSL and/or cable.

John Higdon wrote:

> In article <36e36027.39055325@news.ncal.verio.com>,
> Usenet@NavasGrp.Dublin.CA.US wrote:
>
> > p.s.  The song being sung by the "heavy-duty players in the provider world"
> > in really no different today from last year, the year before, etc., and in
> > my opinion is mostly wishful thinking and/or trial balloons.
>
> But up to now, it was not a real issue. Relatively slow connections
> provided a natural data flow limitation. With more and more customers
> demanding megabit speeds, the possibility for massive throughput is
> exponentially higher.
>
> To wit: for years it has been common for companies to connect to a
> provider via T1, but only pay 56K rates based upon actual data volume
> transmitted. The T1 provided burst speed, but the agreement was that the
> sum total of data would not exceed that which could fit in a 56K pipe. If
> the company went over that amount, it got surcharged for the extra
> effective bandwidth.
>
> Now that consumer connections approach (and in some cases exceed) the
> speed of that T1, those same arrangements will of necessity find their way
> down to the consumer level. Else what would stop anyone from putting a
> nice Sparc10 at home and serving a mondo website?
>
> Are consumers going to want bandwidth charges, or would they prefer
> restrictions on types of equipment to be served by the connection?
> "Neither" is not a workable answer.
>
> --
> John Higdon  |    P.O. Box 7648   | +1 408 264 4115  |      FAX:
>              | San Jose, CA 95150 | +1 500 FOR-A-MOO |+1 408 264 4407
>          abuse@ati.com            |        http://www.ati.com

In article <36E38934.A8944E9E@ecis.com>, jthomas@ecis.com writes...
> I would be quite happy with bandwith limits. Most web sites come up as fast on my
> ISDN line as they do my cable modem. Yeah, downloads are considerably slower, but
> still workable. They still charge for Frame Relay based on pipe size-there is no
> reason that FR's pricing model can't be applied to DSL and/or cable.


It's already being done.  Most of the responsible ISPs that 
serve DSL customers have some kind of bandwidth limit or 
surcharge.

Another method is charging more for more IP addresses.. but
that is a very unreliable model for a variety of reasons.

Tis a pity having to deal with customers who want something 
for nothing, and pretending you're giving it to them. (and 
then deal with their complaints when their mega-oversubscribed 
ISP accounts don't download fast enough for them)

However, I like the current model, where such charges 
originate at the ISP, rather than the ILEC or CLEC.  If 
someone wants to deal w/ a fly-by-night/oversubscribed-up-
the-yin-yang ISP, which *might* give them decent performance
only at 4am when noone's on.. they should have that ability..
just as I have the ability to deal with a provider that is 
more oriented to assuring their customers a reliable connection 
with predictable bandwidth. (at a higher price)

Cable of course is a bit different, since you don't have
a choice of who your ISP is.  Personally I think the
cable IP providers would rather deal w/ the occasional
complaint from the more sophisticated users, than the
major bad press they'd get from the net-neophytes
if they announce bandwidth-charges.



-- 
Philip J. Koenig         The Electric Kahuna Organization          [see below]
-----------------Computers & Communications for the New Millenium-------------
References to my email address in this message have been modified to foil
address-collection robots.  If you wish to send email, use the following 
address by removing numbers and spaces:  pjkunet64 @  ekahuna27 . c o m

In article <abuse-0703992323400001@bovine.ati.com>,
	abuse@bovine.ati.com (John Higdon) writes:
 
> But up to now, it was not a real issue. Relatively slow connections
> provided a natural data flow limitation. With more and more customers

"Relatively slow connections" created a "natural" data flow?  So "relatively"
faster connections are unnatural?  What comes from these unnatural data flows?
Obviously the data gods never intended for us to go any faster than 30cps so 
I guess we've been in trouble for a while now ... huh!!

Of course as the backbones and such, speed up maybe the relative "natural"
data flow will return ... pleasing the data gods to no end I'm sure.

> demanding megabit speeds, the possibility for massive throughput is
> exponentially higher.
> 
> To wit: for years it has been common for companies to connect to a
> provider via T1, but only pay 56K rates based upon actual data volume
> transmitted. The T1 provided burst speed, but the agreement was that the
> sum total of data would not exceed that which could fit in a 56K pipe. If
> the company went over that amount, it got surcharged for the extra
> effective bandwidth.

To wit:  so what?  When these things were "relatively" rare they were sold
a bit differently.  Now they are commodities and sold as such.  It's one of the
side effects of "economics of scale", larger scale lower costs.

> Now that consumer connections approach (and in some cases exceed) the
> speed of that T1, those same arrangements will of necessity find their way
> down to the consumer level. Else what would stop anyone from putting a
> nice Sparc10 at home and serving a mondo website? 

Yeah, so "anyone" puts up a mondo website? So what?  The traffic is limited
to a T1, about 1.5Mb/s.  That's what they were sold.  If companies can't 
survive selling T1 access for what they are charging they either fail or they
raise their rates.  You are also assuming that people will be interested in
this "mondo website".  

> Are consumers going to want bandwidth charges, or would they prefer

You don't get it, we are all paying bandwidth charges now. If your complaint
is that the increments aren't what you'd like to see then start your own
service and charge in increments you do like.

> restrictions on types of equipment to be served by the connection?

Types?  Glad to see you support the "free" market. Each provider is going to
allow only certain "types" of equipment to be connected?  So each time I move
I might as well leave that modem behind since the next provider will most
likely have a different "type" they allow?  I can only have a P200 connected
because the provider has that one CPU on their list of "types" they allow?

> "Neither" is not a workable answer.

Being charged for bandwidth/access is a given, so the possibility of a choice 
between one or the other isn't even on the table.

Paul
-- 
Use that connection!! Idle connections are the devil's playground!!
I, the Diabolical Ruler of the Universe, command it, OC-192 to every residence.
I'm a low resistance pure metal, so don't get me started.

[POSTED TO ba.internet]
abuse@bovine.ati.com (John Higdon) wrote:

>In article <36e36027.39055325@news.ncal.verio.com>,
>Usenet@NavasGrp.Dublin.CA.US wrote:
>
>> p.s.  The song being sung by the "heavy-duty players in the provider world"
>> in really no different today from last year, the year before, etc., and in
>> my opinion is mostly wishful thinking and/or trial balloons.
>
>But up to now, it was not a real issue.

Tell that to the "heavy-duty players in the provider world" -- they have
been saying it was a "real issue" for years.  ;-)

>Relatively slow connections
>provided a natural data flow limitation. With more and more customers
>demanding megabit speeds, the possibility for massive throughput is
>exponentially higher.

I respectfully disagree.  While modem connections were of course limited to
modem speeds, aggregation (some would say overselling) of those connections
to megabit speeds, combined with megabit traffic from ISDN, frame relay, and
(fractional) T-1, has been common for many years.  Nothing has really
changed except the price for and availability of higher speeds, something
that is long overdue thanks to artificially high pricing of ISDN, frame
relay, and T-1.  All that has happened is that the free market is finally
asserting itself.  That may be painful to the dinosaurs, but the world
really isn't coming to an end.

>To wit: for years it has been common for companies to connect to a
>provider via T1, but only pay 56K rates based upon actual data volume
>transmitted. The T1 provided burst speed, but the agreement was that the
>sum total of data would not exceed that which could fit in a 56K pipe. If
>the company went over that amount, it got surcharged for the extra
>effective bandwidth.

I presume you are speaking of fractional T-1.  Again, that was just an
artificial artifact of the lack of free market competition.

>Now that consumer connections approach (and in some cases exceed) the
>speed of that T1, those same arrangements will of necessity find their way
>down to the consumer level.

Not necessarily.  The real cost of shipping bits around has fallen
dramatically over the years.  Price did not keep pace thanks to market
controls.  Now it's catching up in a hurry.

>Else what would stop anyone from putting a
>nice Sparc10 at home and serving a mondo website? 

Nothing.  It's just a matter of buying whatever upstream bandwidth you want,
a system that is already in place.  For that matter, even a decent old home
computer can saturate a T-1 -- a SPARC 10 is unnecessary.

>Are consumers going to want bandwidth charges, or would they prefer
>restrictions on types of equipment to be served by the connection?
>"Neither" is not a workable answer.

Consumers already have bandwidth charges.  You're in effect just arguing
that the charges are too low.  My counter is that the free market will set
the price, now that we finally have a measure of free market competition.

This kind of debate happens over and over around transitions from controlled
markets to free markets -- there are always lots of naysayers claiming that
greatly lower prices simply aren't possible.  Yet over and over the free
market winds up proving the naysayers wrong.  The free market wildcard is
efficiency.

-- 
Best regards,
John  mailto:Usenet@NavasGrp.Dublin.CA.US  http://www.aimnet.com/~jnavas/
FIGHT EXCESSIVE DOMAIN NAME FEES: <http://www.aimnet.com/~jnavas/namefee.html>

newhouse@nospam.crap (Paul Newhouse) writes:
>In article <abuse-0703992323400001@bovine.ati.com>,
>	abuse@bovine.ati.com (John Higdon) writes:
> 
>> But up to now, it was not a real issue. Relatively slow connections
>> provided a natural data flow limitation. With more and more customers
>
>"Relatively slow connections" created a "natural" data flow?  So "relatively"
>faster connections are unnatural?  What comes from these unnatural data flows?
>Obviously the data gods never intended for us to go any faster than 30cps so 
>I guess we've been in trouble for a while now ... huh!!
>

Reading comprehension, Paul.  He said "...a natural data flow *limitation*."

  -Greg

In article <36e3ee16.75374101@news.ncal.verio.com>,
Usenet@NavasGrp.Dublin.CA.US wrote:

> This kind of debate happens over and over around transitions from controlled
> markets to free markets -- there are always lots of naysayers claiming that
> greatly lower prices simply aren't possible.  Yet over and over the free
> market winds up proving the naysayers wrong.  The free market wildcard is
> efficiency.

Maybe so. If bandwidth is becoming the "by the ton" commodity in the
Internet market, then I would expect address space (at least until V6
becomes widespread) to be the next item in short supply. You do now pay a
considerable premium for service if you need any kind of address space.

-- 
John Higdon  |    P.O. Box 7648   | +1 408 264 4115  |      FAX:
             | San Jose, CA 95150 | +1 500 FOR-A-MOO |+1 408 264 4407
         abuse@ati.com            |        http://www.ati.com 

Usenet@NavasGrp.Dublin.CA.US (John Navas) writes:

> [POSTED TO ba.internet]
> abuse@bovine.ati.com (John Higdon) wrote:
> >I have been talking to some heavy-duty players in the provider world and
> >it is generally agreed that in the dedicated-connection arena, we are
> >headed for data-quantity pricing. With high-speed connections becoming the
> >norm, the "natural bandwidth limitations" are disappearing, and providers
> >are going to have to ultimately ration the amount of data that passes
> >through their limited facilities.

Frankly, this is the way it "ought to be done" IMO.

I don't see anything wrong with it fundamentally, and it has a lot of
nicely self-regulating aspects to it.

> Regardless, I personally don't see any real
> issue, since sophisticated ISP's should be able to solve most bandwidth
> problems by caching on their own networks.

You can't cache a lot of the really bandwidth intensive
applications. (Sure, you can cache the porno newsgroups and websites,
and large commonly visited web sites, but you can't (usefully) cache
faxes, email, telephone, point-to-point video, ssh sessions, encrypted
VPN setups, etc.)

---Jim

In article <7c11uj$5bj$1@shell1.ncal.verio.com>,
	gerg@shell1.ncal.verio.com (Greg Andrews) writes:
> newhouse@nospam.crap (Paul Newhouse) writes:
>>In article <abuse-0703992323400001@bovine.ati.com>,
>>	abuse@bovine.ati.com (John Higdon) writes:
>> 
>>> But up to now, it was not a real issue. Relatively slow connections
>>> provided a natural data flow limitation. With more and more customers
>>
>>"Relatively slow connections" created a "natural" data flow?  So "relatively"
>>faster connections are unnatural?  What comes from these unnatural data flows?
>>Obviously the data gods never intended for us to go any faster than 30cps so 
>>I guess we've been in trouble for a while now ... huh!!
>>
> 
> Reading comprehension, Paul.  He said "...a natural data flow *limitation*."

Are you suggesting that relatively fast connections will provide a natural
data flow UNlimitation?  If slow limits then fast must unlimit.

Paul
-- 
Use that connection!! Idle connections are the devil's playground!!
I, the Diabolical Ruler of the Universe, command it, OC-192 to every residence.
I'm a low resistance pure metal, so don't get me started.

In article <36e3ee16.75374101@news.ncal.verio.com>, Usenet@NavasGrp.Dublin.CA.US 
writes...
> abuse@bovine.ati.com (John Higdon) wrote:
> 
> I respectfully disagree.  While modem connections were of course limited to
> modem speeds, aggregation (some would say overselling) of those connections
> to megabit speeds, combined with megabit traffic from ISDN, frame relay, and
> (fractional) T-1, has been common for many years.  Nothing has really
> changed except the price for and availability of higher speeds, something
> that is long overdue thanks to artificially high pricing of ISDN, frame
> relay, and T-1.  All that has happened is that the free market is finally
> asserting itself.  That may be painful to the dinosaurs, but the world
> really isn't coming to an end.
> 
> >To wit: for years it has been common for companies to connect to a
> >provider via T1, but only pay 56K rates based upon actual data volume
> >transmitted. The T1 provided burst speed, but the agreement was that the
> >sum total of data would not exceed that which could fit in a 56K pipe. If
> >the company went over that amount, it got surcharged for the extra
> >effective bandwidth.
> 
> I presume you are speaking of fractional T-1.  Again, that was just an
> artificial artifact of the lack of free market competition.
> 
> >Now that consumer connections approach (and in some cases exceed) the
> >speed of that T1, those same arrangements will of necessity find their way
> >down to the consumer level.
> 
> Not necessarily.  The real cost of shipping bits around has fallen
> dramatically over the years.  Price did not keep pace thanks to market
> controls.  Now it's catching up in a hurry.

[snip]
> >Are consumers going to want bandwidth charges, or would they prefer
> >restrictions on types of equipment to be served by the connection?
> >"Neither" is not a workable answer.
> 
> Consumers already have bandwidth charges.  You're in effect just arguing
> that the charges are too low.  My counter is that the free market will set
> the price, now that we finally have a measure of free market competition.


This reasoning only works if the upstream providers (ie ISPs) 
are also somehow benefitting from this "free market".

Last I heard, T-1 etc. connections to major backbone providers 
were going UP in price, not down.

If UUnet, Sprint, GTE, etc. and all the rest of the backbone 
cabal continue to charge the same high prices for connecting
to their backbone/customers, why does it make any difference
what the cost of the local loop is?

T-1 bandwidth at <$100/mo. means nothing unless the ISP
"middleman" can get bandwidth out to the outside world at
nearly those rates.  Caching is not a fix-all.  All these
1.5Mbit connections at cheap prices will be really cool for
a few months.. until their providers go bankrupt from 
simple economics.



-- 
Philip J. Koenig         The Electric Kahuna Organization          [see below]
-----------------Computers & Communications for the New Millenium-------------
References to my email address in this message have been modified to foil
address-collection robots.  If you wish to send email, use the following 
address by removing numbers and spaces:  pjkunet64 @  ekahuna27 . c o m

newhouse at rockhead dot com writes:
>In article <7c11uj$5bj$1@shell1.ncal.verio.com>,
>	gerg@shell1.ncal.verio.com (Greg Andrews) writes:
>> newhouse@nospam.crap (Paul Newhouse) writes:
>>>In article <abuse-0703992323400001@bovine.ati.com>,
>>>	abuse@bovine.ati.com (John Higdon) writes:
>>> 
>>>> But up to now, it was not a real issue. Relatively slow connections
>>>> provided a natural data flow limitation. With more and more customers
>>>
>>>"Relatively slow connections" created a "natural" data flow?  So "relatively"
>>>faster connections are unnatural?  What comes from these unnatural data flows?
>>>Obviously the data gods never intended for us to go any faster than 30cps so 
>>>I guess we've been in trouble for a while now ... huh!!
>>>
>> 
>> Reading comprehension, Paul.  He said "...a natural data flow *limitation*."
>
>Are you suggesting that relatively fast connections will provide a natural
>data flow UNlimitation?
>

Why, 42, of course.

Now, if you'd care to ask the question using reasonably decent english
grammar, I'll be happy to give my answer.

  -Greg

For_Email@ddress.see.below_plea.se (Phil Koenig) writes:
>In article <36e3ee16.75374101@news.ncal.verio.com>, Usenet@NavasGrp.Dublin.CA.US 
>[snip]
>> >Are consumers going to want bandwidth charges, or would they prefer
>> >restrictions on types of equipment to be served by the connection?
>> >"Neither" is not a workable answer.
>> 
>> Consumers already have bandwidth charges.  You're in effect just arguing
>> that the charges are too low.  My counter is that the free market will set
>> the price, now that we finally have a measure of free market competition.
>
>
>This reasoning only works if the upstream providers (ie ISPs) 
>are also somehow benefitting from this "free market".
>
>Last I heard, T-1 etc. connections to major backbone providers 
>were going UP in price, not down.
>

Really?  Quest (or is it Qwest?) isn't offering high-speed
connections for less than Alternet/Sprint/GTE, etc.?


  -Greg

In article <7c4uuc$5tl$1@shell1.ncal.verio.com>, gerg@shell1.ncal.verio.com 
writes...
> For_Email@ddress.see.below_plea.se (Phil Koenig) writes:
> >In article <36e3ee16.75374101@news.ncal.verio.com>, Usenet@NavasGrp.Dublin.CA.US 
> >[snip]
> >> >Are consumers going to want bandwidth charges, or would they prefer
> >> >restrictions on types of equipment to be served by the connection?
> >> >"Neither" is not a workable answer.
> >> 
> >> Consumers already have bandwidth charges.  You're in effect just arguing
> >> that the charges are too low.  My counter is that the free market will set
> >> the price, now that we finally have a measure of free market competition.
> >
> >
> >This reasoning only works if the upstream providers (ie ISPs) 
> >are also somehow benefitting from this "free market".
> >
> >Last I heard, T-1 etc. connections to major backbone providers 
> >were going UP in price, not down.
> >
> 
> Really?  Quest (or is it Qwest?) isn't offering high-speed
> connections for less than Alternet/Sprint/GTE, etc.?
> 
> 
>   -Greg
> 


Even if true, is Qwest the only company that any ISP ever buys
transit/connectivity from?  I daresay they are not yet a 
majority player in the market.

Things may be different in a couple of years, but we're
talking about prices for DSL/cable customers NOW.



-- 
Philip J. Koenig         The Electric Kahuna Organization          [see below]
-----------------Computers & Communications for the New Millenium-------------
References to my email address in this message have been modified to foil
address-collection robots.  If you wish to send email, use the following 
address by removing numbers and spaces:  pjkunet64 @  ekahuna27 . c o m

For_Email@ddress.see.below_plea.se (Phil Koenig) writes:
>In article <7c4uuc$5tl$1@shell1.ncal.verio.com>, gerg@shell1.ncal.verio.com 
>writes...
>> For_Email@ddress.see.below_plea.se (Phil Koenig) writes:
>> >
>> >Last I heard, T-1 etc. connections to major backbone providers 
>> >were going UP in price, not down.
>> >
>> 
>> Really?  Quest (or is it Qwest?) isn't offering high-speed
>> connections for less than Alternet/Sprint/GTE, etc.?
>> 
>
>Even if true, is Qwest the only company that any ISP ever buys
>transit/connectivity from?  I daresay they are not yet a 
>majority player in the market.
>
>Things may be different in a couple of years, but we're
>talking about prices for DSL/cable customers NOW.
>

Forgive me, but I understood the phrase "...were going up in price..."
as indicating a trend that has been occurring *over time*, with the
implication that the trend will continue through the future.

Qwest certainly isn't a majority player yet, but that doesn't mean
they can't have an effect on the market, does it?

  -Greg

John Higdon wrote:
> 
> In article <920886078.545019@rockhead.com>, newhouse@nospam.crap (Paul
> Newhouse) wrote:
> > You don't get it, we are all paying bandwidth charges now. If your complaint
> > is that the increments aren't what you'd like to see then start your own
> > service and charge in increments you do like.
> 
> Are you serious? If a DSL customer has a circuit capable of 1.54 MBPS, for
> which he pays < $100/month, why would ANYONE continue to pay for a T1 at
> twenty times that amount from a first-tier provider if they are buying the
> same thing?

Well I for one wouldn't but a business might since they can get:

1) Better service and likely on a provider closer to a backbone
2) *Guaranteed* T1 throughput at all times.  Show me an ADSL account
under 100 dollars in the US with *guaranteed minimum bi-directional
rates of 1.544Mbps

--

Paul Doherty
Systems Analyst/Programmer
http://www.dfw.net/~pdoherty
Home of PC DiskMaster

Greg Andrews wrote:


> 
> Really?  Quest (or is it Qwest?) isn't offering high-speed
> connections for less than Alternet/Sprint/GTE, etc.?

HAHAHAHAHHAHAHAHA!

Real "tier 1" backbone bandwidth providers have multiple private
interconnects (at geographically diverse locations) so that data moves
freely from one backbone to another when needed to get from/to its
destination.  Qwest doesn't, they rely on their "peering" at the
overcrowded MAEs.

Also, their pricing is allegedly all over the map.

jc

In article <36E8CB1C.7071173@rahul.net>,
	Me Again <magical@rahul.net> writes:
 
> HAHAHAHAHHAHAHAHA!
> 
> Real "tier 1" backbone bandwidth providers have multiple private
> interconnects (at geographically diverse locations) so that data moves
> freely from one backbone to another when needed to get from/to its
> destination.  Qwest doesn't, they rely on their "peering" at the
> overcrowded MAEs.

Your org is set to "Concentric" (apologies if you are talking about a different
Real "tier 1" backbone) so let's go to BoardWatch 

     http://boardwatch.internet.com/isp/spring99/bb/concentricpg3.html 

and see how they report Concentric is connected to the rest of the internet:

MAE West - San Jose, CA    (Most crowded peering point in the world)
MAE East - Washington D.C.
MAE LA - Los Angeles, CA 
Pac Bell San Francisco NAP - San Francisco, CA 
PAIX - Palo Alto, CA 
Ameritech Chicago NAP - Chicago, IL 

Of course you might mean something different by "overcrowded" than I am 
understanding.  I was guessing you meant number of packets peered.  MAE West
peers more packets than any other peering point on the planet.  The PacBell
NAP isn't exactly underpopulated (it used to be the second most crowded
peering point, followed by PAIX & MAE East).  

So are you suggestting that; peering where lots of peering is going on is bad 
and peering where very little peering is going on is good? 

Paul
-- 
Use that connection!! Idle connections are the devil's playground!!
I, the Diabolical Ruler of the Universe, command it, OC-192 to every residence.
I'm a low resistance pure metal, so don't get me started.

On 12 Mar 1999 13:48:41 GMT, 
newhouse at rockhead dot com <newhouse@nospam.crap> wrote:
>In article <36E8CB1C.7071173@rahul.net>,
>	Me Again <magical@rahul.net> writes:
> 
>> HAHAHAHAHHAHAHAHA!
>> 
>> Real "tier 1" backbone bandwidth providers have multiple private
>> interconnects (at geographically diverse locations) so that data moves
>> freely from one backbone to another when needed to get from/to its
>> destination.  Qwest doesn't, they rely on their "peering" at the
>> overcrowded MAEs.
>
>Your org is set to "Concentric" (apologies if you are talking about a different
>Real "tier 1" backbone) so let's go to BoardWatch 

[snip]

>So are you suggestting that; peering where lots of peering is going on is bad 
>and peering where very little peering is going on is good? 


IMHO, it's the difference between:

1) ISP1 <-> MAE <-> ISP2

and

2) ISP1 <-> BACKBONE1 <-> MAE <-> BACKBONE2 <-> ISP2

and

3) ISP1 <-> BACKBONE1 <-> Private Peer <-> BACKBONE2 <-> ISP2

#1 is often faster than #3.  #2 is always a loser.

Concentric is a borderline case.  It *has* a backbone but it's not
really a backbone provider.  It's just a big ISP.  This puts it
somewhere between 1 and 2 depending on quality of its internal
interconnects.

-- 
Real courtesy requires human effort and understanding.  
Never let your machine or your habit send courtesy copies.

Eric Edwards <ese002@news9.exile.org> wrote:
>IMHO, it's the difference between:
>
>1) ISP1 <-> MAE <-> ISP2
>
>and
>
>2) ISP1 <-> BACKBONE1 <-> MAE <-> BACKBONE2 <-> ISP2
>
>and
>
>3) ISP1 <-> BACKBONE1 <-> Private Peer <-> BACKBONE2 <-> ISP2

I don't see how #1 isn't an apple to #2 and #3's oranges, unless your ISP
also provides the service of flying your clients and servers to the same
city so they don't have to go through a backbone.

miguel

On 12 Mar 1999 20:30:06 GMT, Miguel Cruz <mnc@diana.law.yale.edu> wrote:
>Eric Edwards <ese002@news9.exile.org> wrote:
>>IMHO, it's the difference between:
>>
>>1) ISP1 <-> MAE <-> ISP2
>>
>>and
>>
>>2) ISP1 <-> BACKBONE1 <-> MAE <-> BACKBONE2 <-> ISP2
>>
>>and
>>
>>3) ISP1 <-> BACKBONE1 <-> Private Peer <-> BACKBONE2 <-> ISP2
>
>I don't see how #1 isn't an apple to #2 and #3's oranges, unless your ISP
>also provides the service of flying your clients and servers to the same
>city so they don't have to go through a backbone.

Well, this *is* ba.internet.  Clients and servers *are* often in the
same city.  The exchanges work well in those cases.  It's a bit fuzzier
for large ISP's with private backbones (like Concentric).  I expect
backbones that wish to provide service for other ISP's to use private
peering.

-- 
Real courtesy requires human effort and understanding.  
Never let your machine or your habit send courtesy copies.

In article <7c7l1g$5ba$1@shell1.ncal.verio.com>, gerg@shell1.ncal.verio.com 
writes...
> For_Email@ddress.see.below_plea.se (Phil Koenig) writes:
> >In article <7c4uuc$5tl$1@shell1.ncal.verio.com>, gerg@shell1.ncal.verio.com 
> >writes...
> >> For_Email@ddress.see.below_plea.se (Phil Koenig) writes:
> >> >
> >> >Last I heard, T-1 etc. connections to major backbone providers 
> >> >were going UP in price, not down.
> >> >
> >> 
> >> Really?  Quest (or is it Qwest?) isn't offering high-speed
> >> connections for less than Alternet/Sprint/GTE, etc.?
> >> 
> >
> >Even if true, is Qwest the only company that any ISP ever buys
> >transit/connectivity from?  I daresay they are not yet a 
> >majority player in the market.
> >
> >Things may be different in a couple of years, but we're
> >talking about prices for DSL/cable customers NOW.
> >
> 
> Forgive me, but I understood the phrase "...were going up in price..."
> as indicating a trend that has been occurring *over time*, with the
> implication that the trend will continue through the future.
> 
> Qwest certainly isn't a majority player yet, but that doesn't mean
> they can't have an effect on the market, does it?
> 
>   -Greg


Sorry, didn't mean to imply that this trend is supposed to last
for the forseeable future.

I simply meant to point out that unlike so many other things
in the computer world (memory prices, disk prices, CPU prices,
etc.) you are not seeing similiar kinds of mass price erosion
in transit/backbone connectivity.. on the contrary the trend
(AFAIK) has been upward.

I have no doubt the prices will fall, they just aren't 
falling much YET.

And since they haven't yet fallen on the backbone side,
the prospect of people getting all this T-1 bandwidth 
for $50/mo. *reliably* and *consistently* just isn't 
extremely likely for a while.

Sure the early adopters are seeing some nice performance,
but (as happened w/ @home) that isn't bound to last
for the greater masses as more get hooked up. (until
there is a major paradigm shift on the backbone side)



-- 
Philip J. Koenig         The Electric Kahuna Organization          [see below]
-----------------Computers & Communications for the New Millenium-------------
References to my email address in this message have been modified to foil
address-collection robots.  If you wish to send email, use the following 
address by removing numbers and spaces:  pjkunet64 @  ekahuna27 . c o m

In article <slrn7ej45a.6bf.ese002@spica.exile.org>,
	ese002@news9.exile.org (Eric Edwards) writes:
> On 12 Mar 1999 20:30:06 GMT, Miguel Cruz <mnc@diana.law.yale.edu> wrote:
>>Eric Edwards <ese002@news9.exile.org> wrote:
>>>IMHO, it's the difference between:
>>>
>>>1) ISP1 <-> MAE <-> ISP2
>>>
>>>and
>>>
>>>2) ISP1 <-> BACKBONE1 <-> MAE <-> BACKBONE2 <-> ISP2
>>>
>>>and
>>>
>>>3) ISP1 <-> BACKBONE1 <-> Private Peer <-> BACKBONE2 <-> ISP2
>>
>>I don't see how #1 isn't an apple to #2 and #3's oranges, unless your ISP
>>also provides the service of flying your clients and servers to the same
>>city so they don't have to go through a backbone.
> 
> Well, this *is* ba.internet.  

And ... 

> Clients and servers *are* often in the same city.  

And not necessarily connected to the internet in the same way.

> The exchanges work well in those cases.  

The MAE's work pretty well.  Are they overloaded?  Sure they are.  That just 
indicates the demand for connectivity.  And that demand has occurred rapidly.

> It's a bit fuzzier for large ISP's with private backbones (like Concentric).
> I expect backbones that wish to provide service for other ISP's to use 
> private peering.

It might be fuzzier for you.  The main reason for the appearance of peering
points like the PacBell NAP's is to help avoid a meltdown at MAE West.  It's
overcrowded because everybody and their dog wants to peer there.  Because it's
such a popular place to peer and the area around the Greater Silicon Valley
is so densely occupied with connected sites it makes sense to spread out
the connectivity when it works well.  If it weren't for the offloading of
connectivity around MAE West the site would have melted years ago.  That has
little to do with being a 'Real "tier 1" backbone bandwidth provider'.

The idea that 'Real backbones" don't connect where the connecting is
happening is pretty strange.  Real backbones connect to a lot of places.

Paul
-- 
Use that connection!! Idle connections are the devil's playground!!
I, the Diabolical Ruler of the Universe, command it, OC-192 to every residence.
I'm a low resistance pure metal, so don't get me started.

On 13 Mar 1999 23:52:32 GMT, 
newhouse at rockhead dot com <newhouse@nospam.crap> wrote:
>In article <slrn7ej45a.6bf.ese002@spica.exile.org>,
>	ese002@news9.exile.org (Eric Edwards) writes:
>> It's a bit fuzzier for large ISP's with private backbones (like Concentric).
>> I expect backbones that wish to provide service for other ISP's to use 
>> private peering.
>
>It might be fuzzier for you. 

ISP1 -> private backbone -> NAP -> backbone2 -> ISP2

is clearly better than

ISP1 -> backbone1 -> private peer -> backbone2 -> ISP2

Really? 

Even the optimistic case:

ISP1 -> private backbone -> NAP -> ISP2

Could be a lose if the congestion at the NAP is bad enough.
 
>The idea that 'Real backbones" don't connect where the connecting is
>happening is pretty strange.  Real backbones connect to a lot of places.

Yes they do.  But they don't do *all* their peering at the NAP's.  A 
NAP peer is not as good as private peer.  It's better than no peer but 
if the there is enough traffic to justify it, it is better to use a 
private peer.  And bigger backbones are more likely to have enough cross
traffic to justify private peers among them.  A backbone provider that
does all of their peering at NAP's is at a severe disadvantage to larger
backbones who use private peering.

-- 
Real courtesy requires human effort and understanding.  
Never let your machine or your habit send courtesy copies.

In article <slrn7em7kf.9ob.ese002@spica.exile.org>,
	ese002@news9.exile.org (Eric Edwards) writes:
> On 13 Mar 1999 23:52:32 GMT, 
> newhouse at rockhead dot com <newhouse@nospam.crap> wrote:
>>In article <slrn7ej45a.6bf.ese002@spica.exile.org>,
>>	ese002@news9.exile.org (Eric Edwards) writes:
>>> It's a bit fuzzier for large ISP's with private backbones (like Concentric).
>>> I expect backbones that wish to provide service for other ISP's to use 
>>> private peering.
>>
>>It might be fuzzier for you. 
> 
> ISP1 -> private backbone -> NAP -> backbone2 -> ISP2
> 
> is clearly better than

Not at all clearly. You are asuming that because it is private it is
therefore better.  It could be really bad.
 
> ISP1 -> backbone1 -> private peer -> backbone2 -> ISP2
> 
> Really? 

Maybe but, not by definition.

> Even the optimistic case:
> 
> ISP1 -> private backbone -> NAP -> ISP2
> 
> Could be a lose if the congestion at the NAP is bad enough.

Could be a lose at the private peering points for the same reasons.

>>The idea that 'Real backbones" don't connect where the connecting is
>>happening is pretty strange.  Real backbones connect to a lot of places.
> 
> Yes they do.  But they don't do *all* their peering at the NAP's.  A 

You are the one who suggested real tier 1 providers peer privately, and that
it was somehow superior because it was private. And that they avoided the 
MAE's and NAP's.

> NAP peer is not as good as private peer.  

That is not a given and it really depends on what the point of the peering is.

> It's better than no peer but if the there is enough traffic to justify it, 
> it is better to use a private peer.  

With that qualification I agree.  If you are exchange a large amount of 
traffic with a subset of providers it can make sense to setup a private
peering arangement with them. Wny bog down the MAE's and NAP's with a lot
of well know traffic patterns.  It makes sense and it relieves the more
popular peering points of a lot of traffic that just causes congestion.
That doesn't make it a superior scheme, except in the econimic sense. It makes 
it useful and sensible under certain circumstances.

> And bigger backbones are more likely to have enough cross traffic to justify 
> private peers among them.  

And that can be relative.  Larger local ISP's could find it benefical to
set up local "private" peering. As a number of Denver ISP's found out.

> A backbone provider that does all of their peering at NAP's is at a severe 
> disadvantage to larger backbones who use private peering.

That's not at all clear as a general statement.  Most private peering is
driven by economics.  They still peer at MAE's & NAP's because it makes 
sense to do so.

If the two largest backbones only peered at a private peering point that
means they only peer with each other.  Holy PSINet Batman, that would make a 
lot of sense.  How long do you think it would take customers to abandon those
carriers?

Paul
-- 
Use that connection!! Idle connections are the devil's playground!!
I, the Diabolical Ruler of the Universe, command it, OC-192 to every residence.
I'm a low resistance pure metal, so don't get me started.

In article <7cbtge$6ne$1@news.ycc.yale.edu>,
Miguel Cruz <mnc@diana.law.yale.edu> wrote:
>Eric Edwards <ese002@news9.exile.org> wrote:
>>IMHO, it's the difference between:
>>
>>1) ISP1 <-> MAE <-> ISP2
>
>I don't see how #1 isn't an apple to #2 and #3's oranges, unless your
>ISP also provides the service of flying your clients and servers to
>the same city so they don't have to go through a backbone.

Let's try a real life example...

lair bash[100] t www.cygnus.co.jp
traceroute to godzilla.cygnus.co.jp (203.216.65.129), 30 hops max, 40 byte packets
 4  bb1-fe0-0-100bt.rdc1.sfba.home.net (24.0.0.2)  22.423 ms  12.454 ms
 5  bb2.mae-w.home.net (172.16.4.74)  16.154 ms  11.908 ms
 6  mae-west.idc.ad.jp (198.32.136.117)  10.409 ms  13.88 ms
 7  POS12-0-0.pat1.idc.ad.jp (158.205.224.74)  15.528 ms  22.436 ms
 8  POS2-0-0.tok2.idc.ad.jp (158.205.224.69)  117.174 ms  117.271 ms
 9  FastEthernet10-1.tok1.idc.ad.jp (158.205.224.129)  119.96 ms  117.687 ms
10  gol-tokyo.idc.ad.jp (158.205.226.66)  120.75 ms  119.334 ms
11  alpha-s4-0.gol.net (203.216.2.29)  133.704 ms  140.968 ms
12  oscar.bravo.gol.net (203.216.71.60)  137.724 ms  147.477 ms
13  lima3.gol.net (203.216.36.3)  144.454 ms  146.739 ms
14  tokyo-gw.cygnus.co.jp (203.216.65.190)  172.075 ms  165.382 ms
15  godzilla.cygnus.co.jp (203.216.65.129)  201.253 ms  168.145 ms

No sprint, no uunet, no qwest, no backbone (well, technically
idc.ad.jp appears to be backboning in jp land for gol.net), with MAE
West in the middle.  This is close to case 1.  But dude, I'm in the
Bay Area, and the other guy appears to be in jp.  I didn't think we
were in the same city.  There is a special type of telcom circuit
called a WAN circuit, it can go to a wide area, not limited to just
the same city.  Some people at MAE West actually have these special
types of circuits called WAN circuits.

Oh, and I thought 117 ms to get to .jp was neat in and of it itself.
Last time I did this, it was at around 400 ms.

In article <F8ru77.BB9@kithrup.com>, Mike Stump <mrs@kithrup.com> wrote:
> No sprint, no uunet, no qwest, no backbone (well, technically
> idc.ad.jp appears to be backboning in jp land for gol.net), with MAE
> West in the middle.  This is close to case 1.  But dude, I'm in the
> Bay Area, and the other guy appears to be in jp.  I didn't think we
> were in the same city.  There is a special type of telcom circuit
> called a WAN circuit, it can go to a wide area, not limited to just
> the same city.  Some people at MAE West actually have these special
> types of circuits called WAN circuits.

I can't figure out whether you're talking down to me or just genuinely
amazed by this. A WAN circuit? 

It went across some part of @home's infrastructure (we don't know exactly
what because you deleted that part of the traceroute) to MAE-West, then
across IDC's 155M line to Tokyo, then along their backbone to Global Online.
What is the revelation?

> Oh, and I thought 117 ms to get to .jp was neat in and of it itself.
> Last time I did this, it was at around 400 ms.

This particular miracle is not about backbones, it's about giant undersea
fiber lines.

miguel