Re: [p2pi] ALTO Information Export Service

[Laird Popkin <laird at pando.com>]

I suspect that we're in "violent agreement".

Yes, ALTO can't "find good routes". But there is a relationship between ALTO and routing in an indirect sense - the goal of ALTO is to help applications "find peers that the ISP can probably find better routes to than random peers".

- Laird Popkin, CTO, Pando Networks
  mobile: 646/465-0570

----- Original Message -----
From: "John Leslie" <john at jlc.net>
To: "Reinaldo Penno" <rpenno at juniper.net>
Cc: "Laird Popkin" <laird at pando.com>, p2pi at ietf.org
Sent: Wednesday, October 29, 2008 1:56:30 PM (GMT-0500) America/New_York
Subject: Re: [p2pi] ALTO Information Export Service

Reinaldo Penno <rpenno at juniper.net> wrote:
> On 10/29/08 9:47 AM, "Laird Popkin" <laird at pando.com> wrote:
> 
>> I don't understand why you say that "policy stated in ASNs is actively
>> misleading" though. Keep in mind that ALTO helps p2p networks find good
>> endpoints, not good routes - that's done by BGP, OSPF, etc., when
>> actually exchanging data.

   So far as I know, ALTO is completely unrelated to "finding good routes"
(which, BTW, is an issue possibly worth pursuing for multihomed users).

   While we hope that ALTO will help "find good endpoints", we're talking
about a (small) piece of ALTO only loosely related to that -- ISP policy.

>> selecting endpoints that have cheap routes between them makes it likely
>> that those cheap routes are actually used, but that's more a matter of
>> the lower layers working properly

   ... which is mostly the responsibility of the ISP...

>> than anything ALTO allows ISPs to directly express.

   It's premature, IMHO, to talk of what ALTO "allows".

> If you say "prefer this ASN" (since you are implicitly or explicitly for
> that matter) assuming that prefix 1.1.1/24 is hosted in that ASN but is
> actually not, then is misleading, yes?

   Any case where the publisher of information assumes one meaning while
the user assumes a different one, I would call "misleading".

   Reinaldo raises an issue -- whether an ISP might want to express an
opinion on whether CIDR blocks served by a particular AS are better or
worse than "average". I must admit (as an ISP) that I would not.

   But perhaps some ISP might -- and if so, they should be able to list
list opinions about ASNs separately from their own routing preferences.

--
John Leslie <john at jlc.net>
_______________________________________________
p2pi mailing list
p2pi at ietf.org
https://www.ietf.org/mailman/listinfo/p2pi