The Attitude (was: the Internet Backbone)

Tue Apr 16 18:15:25 UTC 1996

>> From: Jeremy Porter <jerry at fc.net>
>> Bill, here is why you will never be taken seriously be
>> any ISP/NSP.  If networks can't find ways to cooperaate without
>> prodecting there own interests, we all might as well give up
>> now.
>>
>I'll try to remember not to take myself seriously....
>
>This is a most unusual argument: that we should cooperate without
>protecting our own interests; and if we cannot, we should give up.
>
>Are you sure that you know what you are writing?

You mis read, I quiet clearly wrote that, IF networks can not ind
ways to cooperate, while, protecting their own interest, which
in my limited view of the world, means that there MUST be
economic incentives to do the right thing.

>
>> THE ONLY WAY TO BUILD A SCALABLE INTERNET IS FOR PROVIDERS
>> TO BE MOTIVATED BY SELF INTEREST AND ECONOMIC FACTORS.
>>
>This is contrary to your above assertion.

Only via your misunderstanding.

>However, this does not match the historic record.  A scalable Internet
>doesn't even require the concept of "providers".  Nor does it require
>self interest.  There are certainly other models that could be (and have
>been) followed.

Quite, right, we could build networks based on the princicpals
of Marxism, or maybe Stalanism.

No one company, no one country, can build and make stable a
Internet that is growing at the rate our is.  I've got numbers
to back my point up, specificly with res[pect to bandwidht
demand and routing capability.  I can show, the COST is greater
that the investment capital that can be devoted to ths project
by ANY single entity.

If you can show me numbers based in fact, I'd like to see
otherwise.

>
>> This is exaclty an issue of customer education,  If customers
>> wanted good connectity they CAN find it.
>>
>Really?  Where?
>
>And who does the education?  Using what data (as the providers
>themselves have refused to release it)?
>
>May I suggest that this form of argument was advanced for rather a lot
>of other fields, such as food, drugs, comparing doctors and lawyers and
>their fees, and purchasing of automobiles and refrigerators.  As I
>remember, Consumers Union (I'm a lifetime member) was investigated at
>one time as a "communist conspiracy".
>
>This does not follow the subject of this thread, which was in regard to
>(lack of) cooperation with competitors.  I don't see how customer
>education will affect the physical interconnections of carriers, except
>as perhaps encouraging consumers to sue their providers for fraud....

I'd like to see that case one.  If you read your Internet connection
serivce contract closely, you wouldn't even venture this one forward.

I'm all for full disclosure of network metric s and performance.
It would show you just how bad MCI has been lately.

I suggest you and your communist friends, go raise the $10 million
needed to do the testing.  And heck, if they are a 501c3 corp,
I will even discount my normal consulting rates by 50% o design
the monitoring system your are asking for.

oh, wait, you expect the providers to do this themselves, when
they don't even have enough resources to manage their own networks.

>
>> Hm..  maybe Apple should have thought a little more
>> about their little T-1 link to a single provider when they
>> had about 3 T-1s worth of data to send.  I offer to sell
>> more connectivity to Apple, but they WEREN'T interested.
>
>I question your objectivity.  Sounds like you have a personal
>self-interest issue here.

Of course there is personal self interest, that is what drives
the Free market, but I guess, you wouldn't know that.

>By my measurements, there was no problem with the Apple T1(s).  The
>problem I documented was that Sprint didn't have enough bandwidth out of
>Texas (specifically to Chicago in my tests, but reportedly to other
>places as well) to ensure full T1 delivery for Apple while also
>supplying service to others in Texas.  That is, Sprint was
>underprovisioned.  Also, their links flapped a lot.

Well, gee, when a T-1 is operating at 80% saturation,
you WILL LOSE PACKETS, this WILL impact performance.
Eric could probably tell you what utilization it was running at.

>
>> They didn't care that there application/distribution model was
>> broken, and breaking the net, they didn't want to fork
>> out the extra bucks, or deal with the internal politics
>> to put the release out on the West coast where they
>> had a much higher bandwidth connection.
>>
>I disagree, as Apple DID care very much!  In fact, they eventually did
>put it out on other sites, including their west coast connection.

Ok, you are right, I was taking a unfair shot at Apple,
the more probably story is, that Apple didn't realize there would be
a problem until it hit them in the face.  This is fairly
common, I've seen it happen to sevearl of my customers.  Wont
be the first time or last.

>How would it have helped if they had a T3 in Texas, when Sprint couldn't
>even handled another T1's worth of traffic to the same site?

Where is your proof for this assertion?
And what has Sprint's capcity got to do with it, there are at least
4 other providers in Texas/Austin that could have provided the bandwidth
to them.

>How would it have helped the net if they had been multihomed?  You
>provide a separate link to Chicago, and load balancing between you and
>other providers?

Believe it or not, the Internet connects to other locations
that Chicago.  If they had been mutlihomed, it would have
reduced the traffic on the Sprint T-1 by 50% or so
depending on prefered paths through other providers.

You apparently don't understand how Internet routing works at all.

>
>> I haven't seen many reports of Sprint problems since
>> Sprint put its policies in places and they had time work.
>
>Only recently has Sprint been cooperating with the RA enough to actually
>get some Sprint flapping statistics.  I cannot help that you personally
>have not seen them....  Why not look at them yourself?
>
S1849  prefixes=72     flaps=1278
AS2697  prefixes=19     flaps=634
AS701   prefixes=14     flaps=501
AS5000  prefixes=17     flaps=384
AS4628  prefixes=21     flaps=336
AS1275  prefixes=17     flaps=306
AS1836  prefixes=9      flaps=289
AS813   prefixes=6      flaps=247

Don't see srint their myself, I admit, I have been wrong once before.

>> MCI on the other hand has been bleeding about 30% packets out of
>> San Francisco, on its OWN INTERAL network for several MONTHS.
>
>Too true.  I did not mean to pick on Sprint exclusively, except where
>Sean seemed to make a claim that somehow the Sprint network was somehow
>operating better then others, and that it was due to his wonderful
>sagacity in refusing to carry others traffic.

In not carrying non-customer's traffic?
You are just so WRONG.

>I have not (yet) heard the same non-cooperative attitude from MCI.

As far as I know, the only reason MCI hasn't adopted some of Sprint's
policies, is because they started later and are behind the curve.

-- 
Jeremy Porter, Freeside Communications, Inc.      jerry at fc.net
PO BOX 80315 Austin, Tx 78708  |  1-800-968-8750  |  512-339-6094
http://www.fc.net