Filters

[Elmar K. Bins]

dr at cluenet.de (Daniel Roesen) wrote:

> Now looking at the other side of the coin. With many people filtering
> more-specifics, the AS_PATHs for them become long at times. Especially
> when the advertisers are deep burried in 6BONE "structures". People
> connected to (or indirectly have upstream from) the large EU-US IPv6
> transits who do allow more-specific multihoming (C&W, Tiscali) usually
> still have very good AS_PATHs, "if it's seen, it has a good AS_PATH".
> 
> For an example look at DENIC's /48 PA more-specific here:
> http://www.sixxs.net/tools/grh/lg/?prefix=2001:608::/32
> 
> So wether PA more-specific multi-homing "works" highly depends on who
> the upstreams are. But factually it's no real multi-homing as DENIC
> _will_ lose connectivity to _many_ places when the aggregate /32 goes
> down.

Since you mentioned our /48 as an example, I'd like to comment a bit...

Since we're an end site (yes, we're a LIR, but that doesn't help
with IPv6), there's no way to obtain an allocation from any of the
RIRs. This does not greatly bother us, since we chose our transit
providers carefully. We are as happy as someone with an assignment
from a LIR allocation can be, but that is a product of both a good
v6 transit ISP _and_ good peers at the local exchange who for the
most part do accept the /48 unfiltered from us (there are inetnum6
and route objects anyway, so auto-generated filters should let
that prefix pass).

I am not sure what will happen the moment we add another v6 transit
provider; due to the nature of our business (critical to a lot of
ISPs in Germany), we will go multi-homed as soon as we can. We will
try reaching an appropriate agreement with all our transit ISPs,
not to filter the /48 and to properly propagate it to their peers
and customers. If this works out, we see no necessity of v6 PI
space for our home network (we have a different application we
need v6 PI lookalike for, but that's a different thing).

Like with any prefix at all, beyond your own sphere of influence,
nobody can predetermine who will filter the prefix or not. That's
the way the Internet works. We are quite confident that in the case
of someone filtering the /48, and it not showing up somewhere, or
showing up with a very long path, the allocation prefix will bring
the traffic close enough to us so that the /48 catches again :-)

In brief:
  - PA space multihoming depends on your transit/peering ISPs
  - our prefix is well-documented
  - we would prefer people not filtering it
  - if they do, we trust in the power of aggregation

Yours,
	Elmar.

--

"Begehe nur nicht den Fehler, Meinung durch Sachverstand zu substituieren."
                          (PLemken, <bu6o7e$e6v0p$2 at ID-31.news.uni-berlin.de>)

--------------------------------------------------------------[ ELMI-RIPE ]---