Prefix-length FUD (was: Re: Opinions about InterNAP)

Wed Jun 13 20:26:30 UTC 2001

On Wed, 13 Jun 2001, Michael Martin wrote:

> The relevant portion of Tony's explanation (which is very concise) is the
> following --
>
> ++> but should see the route via P2 if P1 is accepting it. (Some
> ++> may either block the announcement or have anti-spoofing packet filters
> ++> at their borders that block the traffic itself).
>
> His explanation is very good but the statement that Seth made was that many
> providers DO block the /24 announcements.  Tony doesn't say anything
> specifically about this.

So the example is that you're numbered out of Provider1's CIDR block.
You're fearing that Provider1 will block announcements of more-specifics
from w/in their own blocks.  My anecdotal understanding (which I agree has
limited value) was that providers who filtered made *exceptions* in their
filtering policies for their own CIDR blocks.

*** What policies any other providers have is unimportant to my example
and things will work just fine no matter the case with those people. ***

At any rate, since you're a paying customer of P1, you at least have
some influence to exert to get them to make exceptions.

As for packet filters (vs. route filters), I doubt many ISPs would
implement such a thing as that filtering is typically done at the
customer edge.

Tony

> I remember plenty of threads on this topic but
> very few non-anecdotal facts about ISP filter policies.  Not being with an
> ISP I'm very curious if there is a good answer.  I'm not immediately
> impacted since Nortel has a class A to work with but I've run into this
> question from clients while doing consulting and just don't feel qualified
> to really answer it authoritatively. Anyone?
>
> ---
> Michael Martin
> Internet Design Engineer
> Internet Engineering
> Nortel Networks