Authors,
I think you'll have some interesting discussion queued for the upcoming
session at IETF 125 about the new TLVs.
The one that catches my eye in particular is the sequence TLV. While I
absolutely understand the motivation for it, I wish to highlight two
underlying properties:
1. Implementations will likely just keep this as a local counter at
their PDU serialization layer and it will have zero additional meaning
in the implementation. To do so contrarily will be to need to keep
considerations about how this impacts internal state. Such an example
would be churn visible in a YANG model for BGP augmented with BMP state.
2. The most obvious use case is to make receiving implementations
happy. One could observe that given the likely property of 1, wouldn't
it be better to simply advise receiving implementations "keep track of
these PDUs by sequence locally?"
-- Jeff
On 3/2/26 11:35, [email protected] wrote:
Internet-Draft draft-ietf-grow-bmp-tlv-20.txt is now available. It is a work
item of the Global Routing Operations (GROW) WG of the IETF.
Title: BMP v4: Extended TLV Support for BGP Monitoring Protocol (BMP)
_______________________________________________
GROW mailing list -- [email protected]
To unsubscribe send an email to [email protected]
