This archive is retained to ensure existing URLs remain functional. It will not contain any emails sent to this mailing list after July 1, 2024. For all messages, including those sent before and after this date, please visit the new location of the archive at https://mailman.ripe.net/archives/list/routing-wg@ripe.net/
[routing-wg]IPv6 Routing Recommendations - comments
- Previous message (by thread): [routing-wg]BGP Update Report
- Next message (by thread): [routing-wg]Comments on IPv6 Routing Recommendations
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Jerome Durand
jdurand at renater.fr
Wed Oct 7 11:11:19 CEST 2009
Hi all,
I would really push for allowing up to */40*
I see 2 practical reasons for that considering our IPv6 deployment example.
1°) We manage several networks
------------------------------
We are in charge of some networks in some french islands (oversea
territories) in addition to our main backbone AS2200, where IPv6 is
fully deployed. These networks have their own AS, own transits... For
IPv4 we use dedicated /24's, easy!
It's been years we are trying to deploy IPv6 there and we are facing
many policy problems (now shifted to this WG). This really prevents IPv6
deployments in our cutomer networks.
The solution here would tell me to use a /36 for each island. I would
then need to use 7 * /36's only for these small networks? I would say OK
but provide me first with a new /32 so I pick up these /36's in a new
prefix!
2°) We have multiple IPv6 transits on our main backbone
-------------------------------------------------------
As IPv6 has been deployed in our networks for many years now we thought
from the beginning it would be a good idea to have /40 prefix per french
region (and allocate /48 prefixes to our customers in this /40). We
thought about allocating a new /40 to a region when the first /40 was
fully used.
We also dedicated some address space to some infrastructure networks of
regional networks, for some projects...
Therefore we are already using many /36's of our network (and don't have
enough space for the 7 islands aforementionned BTW...)
For the moment we have 2 transits (one for north and one for the south)
to avoid useless waste of our network capacity. I would like to announce
southern regions with higher preference on southern transit and
vice-versa for northern transit. Actually I want to be able to do what
we have always done for IPv4...
Also I don't know where I will have my transit providers tomorrow and
how many I will have. I want to be able to have some granularity in the
way I split traffic among my transit providers. Considering a region
makes a lot of sense to me.
The /36 brings too many constraints to me:
- Regions next to eachother are not always in the same /36!
- We already used many /36's... and would require more address space
if this proposal is adopted.
- We don't want to renumber anything (already faced 2 renumbering
6bone -> /35 -> /32... enough please! I don't care so much renumbering
the backbone but our customers don't have time to lose renumbering their
IPv6 networks)
Please note I don't want to announce all the /40's (as I'm not
announcing all the /24's!!). Aggregation remains a MUST for sure and we
will probably announce few prefixes at the end. But I already see that
/36 brings too many limitations... unless I'm provided me with a /28 ;)
Happy to discuss that tomorrow!!
Thanks
Jerome
--
-------------------------------------------------------------
Jerome Durand
Responsable des services aux usagers
Services operations & support manager
Réseau National de Télécommunications
pour la Technologie, l'Enseignement et la Recherche
Tel: +33 (0) 1 53 94 20 40 | GIP RENATER
Fax: +33 (0) 1 53 94 20 41 | c/o ENSAM
E-mail: jdurand at renater.fr | 151 Boulevard de l'Hôpital
http://www.renater.fr | 75013 PARIS
--------------------------------------------------------------
- Previous message (by thread): [routing-wg]BGP Update Report
- Next message (by thread): [routing-wg]Comments on IPv6 Routing Recommendations
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]