It an educated guess that should prevent a undetected rollover occurring. 

-- 
Mark Andrews

> On 2 Dec 2023, at 08:29, Warren Kumari <[email protected]> wrote:
> 
> 
> 
> 
> 
> 
>> On Fri, Dec 01, 2023 at 4:03 PM, Mark Andrews <[email protected]> wrote:
>> It’s stopping the serial changing too fast. 
> 
> 
> 
> Well, yeah, obviously, but what is "too fast"? Why is 2^16 OK but 2^20 or 
> 2^30 or 2^18.365 not?
> 
> W
> 
> 
>> 
>> -- 
>> Mark Andrews
>> 
>>>> On 2 Dec 2023, at 06:43, Warren Kumari <[email protected]> wrote:
>>>> 
>>> 
>>> Dear DNSOP (and Wes),
>>> 
>>> I was wading through my mailbox and realized that I hadn't seen any 
>>> discussion of this.
>>> 
>>> 
>>> I'm quite sure that 2^16 is not a typo (there is quite a lot of text around 
>>> this section), but I cannot really figure out / remember what exactly the 
>>> threat model here is. 
>>> 
>>> Here are the relevant paragraphs:
>>> Sec 2.1.1.1.  The SOA Serial Field:
>>> "Although Section 3.2 of [RFC1982] describes how to properly implement
>>>    a less-than comparison operation with SOA serial numbers that may
>>>    wrap beyond the 32-bit value in both the SOA record and the CSYNC
>>>    record, it is important that a child using the soaminimum flag must
>>>    not increment its SOA serial number value more than 2^16 within the
>>>    period of time that a parent might wait between polling the child for
>>>    the CSYNC record."
>>> 
>>> Sec 5.  Security Considerations
>>> "To ensure that an older CSYNC record making use of the soaminimum
>>>    flag cannot be replayed to revert values, the SOA serial number MUST
>>>    NOT be incremented by more than 2^16 during the lifetime of the
>>>    signature window of the associated RRSIGs signing the SOA and CSYNC
>>>    records.  Note that this is independent of whether or not the
>>>    increment causes the 2^32 bit serial number field to wrap."
>>> 
>>> 
>>> I can (mostly) understand why the SOA must not fully wrap (2^32) or 
>>> probably even 1/2 wrap (2^31), but what bad thing would happen if it 
>>> incremented by e.g 2^24? 
>>> 
>>> It might just be that 2^16 was sufficiently far from 2^32 that it was 
>>> viewed as "conservative even with much slop", but that feels somewhat like 
>>> a cop-out…
>>> 
>>> Can someone help me understand?
>>> W
>>> 
>>> 
>>> 
>>>> On Thu, Nov 09, 2023 at 1:45 PM, Bob Harold <[email protected]> wrote:
>>>> https://datatracker.ietf.org/doc/html/rfc7477#section-5
>>>> section 5.  Security Considerations
>>>> last paragraph
>>>> 
>>>> "the SOA serial number MUST NOT be incremented by more than 2^16"
>>>> 
>>>> 2^16 is a very small fraction of the 2^32 serial number space.  It seems 
>>>> that half of the 2^32 would be sufficient, which is 2^31 (not 2^16).  Is 
>>>> that a typo, or is there a reason for the small range?
>>>> 
>>>> -- 
>>>> Bob Harold
>>>> 
>>>> _______________________________________________ 
>>>> DNSOP mailing list 
>>>> [email protected] 
>>>> https://www.ietf.org/mailman/listinfo/dnsop
>>>> 
>>> 
>>> 
>>> 
>>> _______________________________________________
>>> DNSOP mailing list
>>> [email protected]
>>> https://www.ietf.org/mailman/listinfo/dnsop
> 
> 
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to