Re: [dhcwg] rfc3679bis, or: about allocation of DHCPv4 options (was [IANA #1172829] Request for Early Allocation of DHCPv4 option (draft-ietf-dhc-v6only)

["Bernie Volz (volz)" <[e-mail]>]

I'm really not sure why this is generating so much discussion?

Including Jen as dhc wg is included, so just to make sure she see it.

I do believe some had a bad experience with option 160, however, it was related to Polycom usage. See https://tools.ietf.org/html/draft-ietf-capport-rfc7710bis-07 - it will ask to use option 114. We had a reference for that one and contacted Dieter at Apple to confirm it was not still in use and whether he thought it would create any conflict and the result was that there would be no issues as it never made it into the wild.

And, when we did the work to determine what was in use (~2004 if I recall), no one ever reported that 160 had been used for anything so there was no reason for IANA to hold off on allocating it (as we did recommend with some others that were in use). The general guidelines was indeed to avoid their use, but given the experience with 160 even that approach failed. So the bottom line is that any value has a small risk as people violated the rules; this was more obvious in the original private use space as vendors thought they could use a value.

For 108, while we have some information it was used for Swap Path, we sadly have no information about who was potentially using it to contact them. Also, a (perhaps now dated DHCP fingerprint file I have based on PRL values) has no one using 108 in the PRL.

So, can we be 100% certain there will not be an issues ... no. But that applies to any value IANA may pick. And, we've probably now done more to notify people of the use of 108 and so anyone that knows of a concern would hopefully have stepped forward? Or perhaps may in the next few days?

Note also that there have been other options allocated which have not run into problems as far as I am aware.

I don't think I would change the IANA recommendation, but I also see no reason to disallow folks from recommending values if they have some reason to do so.

- Bernie

-----Original Message-----
From: dhcwg <[e-mail]> On Behalf Of Michael Richardson
Sent: Tuesday, June 23, 2020 1:38 PM
To: [e-mail]; Bernie Volz (volz) <[e-mail]>
Subject: [dhcwg] rfc3679bis, or: about allocation of DHCPv4 options (was [IANA #1172829] Request for Early Allocation of DHCPv4 option (draft-ietf-dhc-v6only)


I hate that ML are not consistently named for the WG. resending with correct WG ML list name.

....

Bernie, is some of the desire to pick <127 a result of the experience with CAPPORT's use of 160?

Unfortunately, while I participated in the thread that Jen linked to justify why ietf-dhc-v6only should have 108, I can't seem to find much discussion about 108 itself.
(Much of it is not about why 108, and rather about the protocol as a whole)

So, aside from Tomek's original message there isn't much discussion about why we are going against IANA's normal process here.

Given this recommendation and the experience with CAPPORT, should we be changing advice to IANA in general?

Is there an need to repeat the delouse effort that resulted in RFC3679?
RFC3679 already says:

   IANA is requested to reassign these option codes after the list of
   option codes that have never been assigned or have previously been
   returned has been exhausted.

Maybe this is not well understood.

--
Michael Richardson <[e-mail]>, Sandelman Software Works  -= IPv6 IoT consulting =-