I'm fine with this.
--Jafar
On 6/16/2015 10:17 AM, Alain Ritoux wrote:
Hi all
I'd really like that the chosen final name doesn't imply any hierarchy
between the RIBs; the hierarchy itself being left to the user.
I read at some point that the logical-table was too long, and I agree.
I just realized then that in my proposal, I forgot to add the "short"
version.
So here it is again:
CLI level:
# The command to pin the logical somewhere in the OS available
technology
logical-table XXX (... and here the OS specific sutff to describe
where it will be installed ...)
e.g.
- as of today:
logical-table 1 netns /var/run/VRF1
- possible evolutions:
logical-table 1 mrf 2 (mrf/table/whatever ...)
logical-table 1 netns /var/run/VRF1 mrf 2
# Short version when used as a parameter of other commands: ltid
e.g for static routes:
ip route 200.0.0.0/24 1.2.3.4 ltid 1
Code level:
s/vrfid/ltid (standing for logical table index)
Best regards
Alain
On 06/16/2015 04:16 PM, Jafar Al-Gharaibeh wrote:
Andrew,
According to the figure, you built your point view on the assumption
that LR is a network namespace, or at least associated with one
(one-to-one). While this is true for the VRF patch from 6wind, the
actual association is with one routing table within a network namespace.
That is what the term should describe. Multiple routing tables could
live in one namespace, you could use the term "Logical Router" to
describe that. Which means, a "Logical Router: LR" contains one or more
"Logical Routing and Forwarding: LRF" tables. For the same reason, MRF
is going to be confusing. We could call it SRF! for single RF table if
the L is unfit :)
Regards,
Jafar
_______________________________________________
Quagga-dev mailing list
[email protected]
https://lists.quagga.net/mailman/listinfo/quagga-dev
