See Inline...
----- Original Message -----
From: "Howard C. Berkowitz"
To:
Sent: Sunday, June 02, 2002 11:17 AM
Subject: Re: Another BGP attribute question [7:45619]
> At 7:00 AM -0400 6/2/02, Nigel Taylor wrote:
> >All,
> > I was reading the old RIPE(22nd meeting minutes) and was
wondering,
> >what
> >ever became of the BGP
> >proposal from Tony Bates and Enke Chen for the use of the Destination
> >Preference Attribute (DPA) for multi-homed sites.
>
> DPA keeps coming up, at least for end-to-end route selection. Its
> basic problem is that only ISPs with whom you have an economic
> relationship have any motivation to respect it. Geoff Huston's
> NOPEER is a simpler way to accomplish the same thing (probably
> coupled with class of service request communities).
Howard, thanks a lot for the info/insight of DPA and specifically pointing
me to the "NOPEER"
attribute draft. I was able to briefly read over the draft and I must say
this does seem
like a solution to the present problem. However, I was also doing some
reading of the
APNIC's (http://www.apnic.net/meetings/13/sigs/docs/irr-presentation.ppt)13
minutes
and it's noted some of the present problems with the IRRs. The one that
seems to apply
here would be the statement that, "About 50% of full routes are not
registered to public
IRRs.
I have a question? Do you see the "NOPEER" as having a directory class in
the RPSL
and if so in doing some recent reading of RPSL, and RPSLng, the enhancements
RPSL on the
same site wouldn't the "NOPEER" attribute be limited to representing what is
known in
the IRRs. With this being the case how effective can the attribute be, when
representing
at best 50% of the global BGP FIB.
Of course then there is the ever present security issues which seems to
being getting some
attention through the RPSS(rfc2725).
>
> >Based on our preivous thread with the known and unknown implications of
> >"inconsistant routes", I would think
> >this could've have been a step in the right direction.
> >
> >I did find a link where Enke Chen notes the use of the "LOCLA_PREF"
> attribute
> >by many providers, since the
> >lack of the DPA and rfc1998 also notes how the use of "communities" aid
in
> >this process.
>
> You can really solve LOTS of operational issues with creative use of
> communities. While RFC2547 was one driver for creating an extended
> community attribute, there are various ideas floating around for
> other applications thereof.
Do you care to mention some of the other ideas..floating aeround?
>
> >
> >Anyone has any thoughts or suggestions on this as it applies to the use
of
> >DPA
> >and where things stand on
> >global/ISP-based implementation of this attribute?
>
>
> As far as I know, it's never been implemented in operations. I'm
> reasonably certain that some versions of Bay RS could generate it,
> but I don't know of anyone that listens for it.
I remebered in reading Sam Halabi's book - Internet Routing architectures
(Pg. 118, 1st ed)
he noted cisco's lack of support for attributes 11(DPA). However, it is
noted as bieng MCI defined.
As you pointed out I've yet to come across anything that suggest anyone is
making use of the DPA
attribute.
>
> --
> "What Problem are you trying to solve?"
> ***send Cisco questions to the list, so all can benefit -- not
> directly to me***
>
****************************************************************************
****
> Howard C. Berkowitz [EMAIL PROTECTED]
> Chief Technology Officer, GettLab/Gett Communications
http://www.gettlabs.com
> Technical Director, CertificationZone.com http://www.certificationzone.com
> "retired" Certified Cisco Systems Instructor (CID) #93005
thanks
Nigel
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=45637&t=45619
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
