> Have you tried?
Thank you John and Job. A long time ago I got one (not personal).
19092 Exa Networks Ltd
I could not recall it was allowed for individuals. Thank you for correcting my
misunderstanding.
Still my point stands:
The problem is that developers do not know what code they can use for
experimentation when they see TBD in a draft.
The proposed document does not prevent anyone to continue squatting and
polluting the DFZ with unknown attributes when implementing new drafts.
It does ask developers to go through a convoluted way to re-encode attributes
to experiment and them fully re-code the feature when it needs to go to
production.
Thank you but no thank you: as a developer, I just want to know that I can use
some attribute code safely.
AS operators will know what code is run on their kit, and can make sure that
there is no conflict of experimental features.
(And it would be trivial for WG chairs to make sure drafts do use different
codes).
Also this draft may introduce a new number of new challenges:
Once we have IANA defined sub-attributes, we may also see vendors “matching”
some other vendors experimental feature - which IMHO would be unadvisable. What
then ?
What need will vendors then have to standardise their own namespace ?
Just reserving a range such (let say 200-254) for experimentation and making
them non eBGP transitive would match what is already done with Private ASN,
preventing their propagation is MUCH simpler, and let people experiment as they
want.
(see included patch which would implement such draft/RFC)
Thomas
—
diff --git a/lib/exabgp/bgp/message/update/attribute/attributes.py
b/lib/exabgp/bgp/message/update/attribute/attributes.py
index cedd82b..0745117 100644
--- a/lib/exabgp/bgp/message/update/attribute/attributes.py
+++ b/lib/exabgp/bgp/message/update/attribute/attributes.py
@@ -227,6 +227,8 @@ class Attributes (dict):
alls = set(keys + default.keys() if with_default else [])
for code in sorted(alls):
+ if code in range(200,255) and local_asn != peer_asn:
+ continue
if code in (
Attribute.CODE.INTERNAL_SPLIT,
Attribute.CODE.INTERNAL_WATCHDOG,
_______________________________________________
GROW mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/grow
