[6bone] In the summer time, we got cleaning to do... Where is
UUNET?
John Fraizer
tvo@EnterZone.Net
Fri, 2 Aug 2002 13:12:47 -0400 (EDT)
On Fri, 2 Aug 2002, Michel Py wrote:
> John,
>
> > John Fraizier wrote:
> > (We'll use your peering session as an example Michel)
> > Any comments appreciated.
>
> - I am happy to see the addition of "set community no-export additive".
>
> - I think you should drop seq 20 2000::/3 ge 16 le 16
>
> - As someone else mentioned, the "subTLA-only" name might not be
> appropriate.
>
> - There is overlap in the following two lines:
> neighbor 3ffe:1ced:ff02::2 prefix-list AS-ARNEILLPY in
> neighbor 3ffe:1ced:ff02::2 route-map AS-ARNEILLPY in
> The prefix-list is useless as the route-map uses it IMHO.
>
Michel, actually, if you look at our config for your peering session, the
prefix-list actually does do something. Since I use a "canned" route-map
for all peers, I have to use the neighbor 3ffe:1ced:ff02::2 prefix-list
AS-ARNEILLPY in on your peering session.
Notice in the route-map below, it would accept ANY prefix from you and
redistribute it as long as it passed prefix-list subTLA-only.
route-map AS-ARNEILLPY permit 10
match ipv6 address prefix-list subTLA-only
!
route-map AS-ARNEILLPY permit 20
match ipv6 address prefix-list AS-ARNEILLPY
set community no-export additive
Since I'm prefix-list filtering you in the neighbor statement, we will
only allow prefixes from you that pass prefix-list AS-ARNEILLPY to pass on
to the route-map. Once there, the match in stanza 20 sets the no-export
(in your case.)
This is simply to prevent you from accidently leaking transit routes to
me.
For peers that we do transit:transit peering with, they don't have a
"prefix-list [peername] in" filter and the prefix-list [peername] is
simply used to allow "more specifics" that won't pass the subTLA-only
filter. Those more specifics are tagged no-export then.
Do you see the reasoning now? This way, every "transit:transit" peer is
configured the same and every "customer:customer" peer is configured the
same way. The route-map is the same for every peer, period. It makes
life more simple.
>
> router bgp 13944
> neighbor 3ffe:1ced:ff02::2 remote-as 23169
> neighbor 3ffe:1ced:ff02::2 description ARNEILLPY
> no neighbor 3ffe:1ced:ff02::2 activate
> !
> address-family ipv6
> neighbor 3ffe:1ced:ff02::2 activate
> neighbor 3ffe:1ced:ff02::2 next-hop-self
> neighbor 3ffe:1ced:ff02::2 route-map AS-ARNEILLPY-IN in
> neighbor 3ffe:1ced:ff02::2 route-map AS-ARNEILLPY-OUT out
> exit-address-family
> !
> ipv6 prefix-list AS-ARNEILLPY seq 5 permit 3ffe:1ced:a002::/48
> !
> ipv6 prefix-list MYPREFIX seq 5 permit 3ffe:1ced::/32
> !
> ipv6 prefix-list STRICT seq 5 permit 3ffe::/18 ge 24 le 24
> ipv6 prefix-list STRICT seq 10 permit 3ffe:4000::/18 ge 32 le 32
> ipv6 prefix-list STRICT seq 15 permit 3ffe:8000::/20 ge 28 le 28
> ipv6 prefix-list STRICT seq 20 permit 2001::/16 ge 29 le 35
> ipv6 prefix-list STRICT seq 500 deny any
> !
> ipv6 prefix-list SIXTOFOUR seq 5 permit 2002::/16
> !
> route-map AS-ARNEILLPY-IN deny 10
> description deny the peer feeding me my own prefix
> match ipv6 address prefix-list MYPREFIX
> !
> route-map AS-ARNEILLPY-IN permit 20
> description generic filter
> match ipv6 address prefix-list STRICT
> !
> route-map AS-ARNEILLPY-IN permit 30
> description accept other prefixes, don't redistribute
> match ipv6 address prefix-list AS-ARNEILLPY
> set community no-export additive
> !
> !
> route-map AS-ARNEILLPY-OUT permit 10
> description generic filter
> match ipv6 address prefix-list STRICT
> !
> route-map AS-ARNEILLPY-OUT permit 20
> description feed _my_ 6to4 route only to the peer
> match ipv6 address prefix-list SIXTOFOUR
> match ** whatever condition that says the 6to4 route is yours **
> set community no-export additive
> !
>
>
> Michel.
That works fine except that it doesn't prevent you from accidently leaking
transit to us. We'd gladly accept any route you sent us that passed
prefix-list STRICT.
The one thing I noticed is the deny stanza, something that I forgot in
mine. I don't know how. I always do that on our v4 stuff. I'll have to
modify the route-maps to add that. Thanks for reminding me!
---
John Fraizer | High-Security Datacenter Services |
EnterZone, Inc | Dedicated circuits 64k - 155M OC3 |
http://www.enterzone.net/ | Virtual, Dedicated, Colocation |