On 31. 08. 19 1:27, Wes Hardaker wrote:
> Shane Kerr <[email protected]> writes:
> 
>> While I thought the RCODE linkage was a bit clunky, the idea of having
>> some structure to the response codes was actually kind of nice, for
>> the same reason that the 1xx, 2xx, 3xx, 4xx, 5xx status codes were
>> nice. I think the draft is better without using RCODE, but maybe we
>> can pick numbers for EDE that are grouped in a similar way?
> 
> So, assuming we *can't* easily group them by rcode.  Well, we can, but
> the results may not match given discussions with implementers.
> 
> If you want to take a whack at suggesting appropriate ranges I'd love to
> see what you come up with.  As with all loaves of bread, do you slice
> them cross-wise, length-wise or diagonally?

Main question is not "what categories we need" but "what is the purpose/who is 
consumer of the information"? Once we have answer to this it will be easier to 
establish a categorization.

IMHO the most useful information is dichotomy:
a) the problem is local (= call network admin/ISP/telco)
b) the problem is remote (= call your bank because their internetbanking broke 
and _do not your bother ISP_).

Such simple distinction would allow client apps way more informative message 
than "We’re having trouble finding that site" + would make life easier for 
support people.

This distinction is especially important for DNSSEC validation failures - e.g. 
if RRSIGs are expired your ISP/resolver operator _should not_ take any action 
so it is desirable to "assign blame" to web site owner. 


Shane, what do you think about proposal at the end of message
https://mailarchive.ietf.org/arch/msg/dnsop/JHYQunp0L7C_vUcBb9fOi22QfpI
- specifically Near and Far bits?

Also we need to ask stub resolver developers what would be useful for them. 
Bernardo, what information would make life easier for stub/app developers?

-- 
Petr Špaček  @  CZ.NIC

_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to