draft-mymb-sfc-nsh-allocation-timestamp has been presented to the ISE
for publication as an Informational RFC on the Independent Stream.
==Purpose==
The motivation for this document is to describe an SFC NSH metadata
type 1 encoding for a compact timestamp. This format has already been
implemented in silicon by several vendors and is presented in this
document to allow it to be understood when seen in the wild, and to
allow interoperable implementations.
== History==
This document was originally introduced to the SFC WG in January 2017.
It did not get much attention as, apart from the authors, there seemed
to be little enthusiasm for this function. The working group chairs
suggested that the authors pursue publication in the Independent Stream.
The document was first brought to the ISE in October 2020 at version -07
==Non-IETF Work==
With something so closely related to IETF work, it is important to make
a distinction between this and the output of the IETF. Asides from the
usual Independent Stream boilerplate, this document contains an explicit
statement in the Abstract and Introduction that the work has not been
standardised by the IETF.
==Document Status==
Consideration was given as to whether this should be positioned as
Experimental, but the function described has been implemented by more
than one silicon vendor making Informational more appropriate.
==Security Considerations==
The document makes some observations about the vulnerabilities
introduced by timestamping,and also notes how attacks on the system
time mechanisms might cause errors in time stamping.
No new security mechanisms are introduced with reliance being on the
security of existing protocols.
==IANA==
The document makes no request for IANA action.
==Implementation==
As noted in the document, the authors claim the existence of multiple
implementations in silicon. The authors were unwilling to commit names
to that statement.
==Reviews==
It was also important to get reviews from people closely involved in the
SFC working group. The ISE (who has worked a lot in the SFC WG) provided
a review and commissioned further reviews from Greg Mirsky,
Med Boucadair, and Carlos Pignataro (all very active in the working
group). The reviews resulted in a number of revisions for mainly minor
editorial points.
Details of the reviews are extensive, but can be retrieved on request.
==Debate===
One question arose related to whether the authors should also define a
timestamp extension for metadata type 2. Doing so would keep the
symmetry and help keep code points aligned. However, there is no
requirement from the SFC WG that both metadata types should be defined,
and no interest from anyone in having a type 2 definition. It was,
therefore, not included.