Re: enterprise multihoming with ISP consortiums

[Michael Richardson <mcr@sandelman.ottawa.on.ca>]

-----BEGIN PGP SIGNED MESSAGE-----


>>>>> "Darrell" == Darrell Root <droot@cisco.com> writes:
    Darrell> I'm open to suggestions regarding how to do this.  One
    Darrell> possibility is to try to align ISP consortiums with major
    Darrell> peering points (having one or more "MAE-West-centric"
    Darrell> consortiums instead of allowing arbitrary consortium
    Darrell> combinations). In effect this is similar to geographic
    Darrell> based allocation, but the "geographic" basis is derived
    Darrell> from peering points instead of ICBM addresses ;-)

  Yes.

  Your proposal is essential identical to geo-pi, except that with geo-pi,
the "consortium" is geographic, and has a local monopoly on PI addresses.
  
  My concern with your proposal is that getting ISPs to cooperate is
difficult. It is unlikely that tier-1s would join many consortia.
  The tier-1-wannabees in Canada (they are all tier-2s, but few of them 
will admit that) are near impossible to get to cooperate, because they 
think that this will reveal that they aren't a tier-1. 

  This would work very well for tier-3s. However, the problem is that,
once the consortium of tier-3s is formed, I suspect that, if things go well,
they will simple gobble each other up, resulting in me being "multihomed"
to a tier-2 ISP. (Further, some tier-2 ISPs think that "multihoming" means
getting a connection to their Ottawa POP and one to their Montreal POP).

  Having said all of this, it doesn't mean that I think that the world has
to work this way. I strongly believe that IXs makes a HUGE amount economic
sense outside of the metropolitan areas of US. Inside the US, it seems that a
combination of oversupply of fiber and market conditions has lead many to
conclude that there is no business model for IXs. I can't argue this point,
as I don't quite understand the situation there.

    Darrell> 3.2.6) Cooperation between transit providers:
    Darrell> This proposal is in complete violation of the
    Darrell> requirement that multihoming solutions not require
    Darrell> ISP cooperation.  The question is whether the
    Darrell> economic incentive of enterprises wanting to pay
    Darrell> for multihoming will be sufficient to convince
    Darrell> ISP's to work together to provide the service.

  Precisely.

]       ON HUMILITY: to err is human. To moo, bovine.           |  firewalls  [
]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
] mcr@sandelman.ottawa.on.ca http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another Debian GNU/Linux using, kernel hacking, security guy"); [

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)
Comment: Finger me for keys

iQCVAwUBPdfY9IqHRg3pndX9AQGujQP+MNsDpuKalxJBEXMZY7KEVDxXkYYEv/fX
29TuBWKucZLDLBO7LO+9+yjHDou5UYTOXFM/L+i2+26SLfE9vYcJy1FdsrdguDAl
jPjF68tRCK2OSpEbG6mNJ+X5r4yOaxXxkNexzVs0efjyND6yLCXQ4uGw71QNibo4
LdI/NVJ3rPk=
=saTk
-----END PGP SIGNATURE-----