Sandy, speaking as Joan Q private citizen, wg member only, in civilian
mufti:
With all due respect Sriram, I disagree.
On Thu, 2 Dec 2010, Sriram, Kotikalapudi wrote:
Changed the subject title to refer to the document WGLC.
Please see comments inline. Thanks.
Sriram
-----Original Message-----
From: Randy Bush [mailto:[email protected]]
Sent: Thursday, December 02, 2010 5:48 PM
I also request that it be an Informational RFC (rather than a
standards track RFC) just as draft-ietf-sidr-roa-validation already
is.
like hell. it specifies what running code MUST do in routers.
Mapping the updates to Valid, Invalid, etc. based on ROA objects is
an intermediate step to a final step of making a decision about path selection.
Both steps are internal to the router.
There needs to be a common standard understanding of what the ROA means,
in order for the signers to know what it means to produce a ROA.
If an ISP has leeway in the final step,
then why should anything prevent it from having some leeway in the first step
as well
(primarily during partial deployment)?
If the router has one degree of freedom in its action, is anything gained
by having two?
I am wondering what is lost if this is an Informational RFC?
Recognize that there is precedent for this - RFC 4271 standardizes the
router's *internal* best path decision process. I expect that the
reasoning is the same - the decision process is standard so that those who
are constructing BGP packets are operating with a standard understanding
of what those packets mean.
--Sandy, with no particular ceremonial vestments or badges of rank donned.
>> >> randy
_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr
