Re: [ippm] WGLC for draft-ietf-ippm-ioam-yang

[t petch <[e-mail]>]

On 15/08/2022 02:30, Tianran Zhou wrote:
> Hi Tom,
>
> Thanks again. Please see in line.
>
> Tianran
>
> -----------------------------
> Well yes, but RFC9197 (needs adding to the I-D references and to the
> YANG) has a MUST for e.g. support of the default namespace so having a lower case 'must' to me downgrades that requirement whereever it appears.  It is common for YANG modules to have RFC8174 language with the disclaimer therefrom next to the Copyright of the module.
>
> ZTR> OK. I will update.
>
> In passing, the spelling of 'namespace' is inconsistent.
> ZTR> Thanks.
>
> I wonder if there is a security exposure such as a buffer overflow when a string is 18446744073709551615 characters long.
> ZTR> I believe we do not need such a long name. I think various devices may have different constraints. Maybe we can make it 300 characters long? Any suggestion?

That value is the one in RFC7950 that implementations are expected to 
support!  I always think that there should be a lower constraint but 
have no logic to support any given figure in this case. A profile seems 
to be internal to the node, name unique within the node and so likely to 
be a quite short string unless it contains a node identifier or some 
such.  Whatever value you put in, someone (IESG!) will say why, to which 
the answer is why not?.

Tom Petch

> Tom Petch
>
>> Tom Petch
>>
>>>
>>> The last call will go until Wednesday, August 31.
>>>
>>> Best,
>>> Tommy & Marcus
>>>
>>>
> .
>