Re: [stir] error handling 03

[Chris Wendt <[e-mail]>]

Thanks Paul,  

> On Aug 5, 2022, at 12:57 PM, Paul Kyzivat <[e-mail]> wrote:
> 
> Chris,
> 
> This text pretty much clears up my concerns.
> But I think it would also be helpful to change the "Protocol Cause" field in the registry from "Status code" to something different than for SIP and more specific to STIR. E.g. "STIR Error Code".

Yes, i like that suggestion

> 
> Also, where do you plan on inserting that paragraph? I'm guessing you intend it after the first paragraph of section 5, before the example.

yes, sorry, that is exactly where i plan to put the text.

> 
> 	Thanks,
> 	Paul
> 
> On 8/5/22 11:42 AM, Chris Wendt wrote:
>> Hi All,
>> As a follow up from the 114 stir meeting discussion, we went through both Christer’s question and Paul’s questions.  My action item from the meeting was to clarify two of Paul’s comments, specifically:
>>>>>> 
>> Also, I think this draft should be more explicit in stating that the new STIR protocol it defines permits multiple uses, and perhaps constraints on how. (E.g., in what ways the multiple uses must differ, or how to resolve ambiguities among them. I *think* the response codes defined in RFC 8224 are mutually exclusive for a single passport, so perhaps constrain to a single cause per ppi.)
>> <<<<
>> >>>>
>> The Protocol Cause for the STIR Protocol Value is specified as "Status code". This is the same as is defined for the SIP protocol. The values used for STIR as a subset of the SIP Protocol Causes. It isn't clear what values are valid for use here. I guess you intend the values specified in section 6.2.2 of RFC 8224. But defining that way isn't future proof. What if a subsequent extension/update to RFC 8224 defines some additional codes?
>> <<<<
>> Based on discussion in the meeting, I plan to cover both of these comments with the following:
>> As implied and defined in {{RFC8224}}, error codes associated with STIR targeted at authentication services that produced a specific identity header represent a single error occurring with the verification and processing of that identity header. Therefore the association of a "ppi" parameter with a Reason header using "STIR" protocol MUST only identify a single cause code in the context of a call dialog defined in {{RFC8224}} or in future documents defining STIR related errors.
>> Would like to get some initial feedback on that, I plan to create an -03 version early next week depending on comments.
>> Thanks!
>> -Chris
>> _______________________________________________
>> stir mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/stir
> 
> _______________________________________________
> stir mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/stir