Re: [Dime] AD review of draft-ietf-dime-drmp-03

[Steve Donovan <[e-mail]>]

All,

I've commented on Stephen's and Lionel's suggested changes below.

If there is agreement to my proposed changes outlined below then I will 
submit a new version of the document.

Regards,

Steve

On 3/8/16 9:41 AM, [e-mail] wrote:
> i will let Steve react but I can give my feeling :)
>
> The priority is set by the Diameter or Diameter server, not by agent.
>
> It is somehow describe in section 6 Theory of Operation
>
>     2.  Agents handing the request - Agents use the priority information
>         when making routing decisions.  This can include determining
>         which requests to route first, which requests to throttle and
>         where the request is routed.  For instance, requests with higher
>         priority might have a lower probability of being throttled.  The
>         mechanism for how the agent determines which requests are
>         throttled is implementation dependent and is outside the scope of
>         this document.  The agent also saves the transaction priority in
>         the transaction state, either as locally managed state or using
>         the Proxy-Info mechanism defined in [RFC6733].  This will be used
>         when handling the associated answer message for the transaction.
>
> Agents are just using this information if present. They are not modify it or include it if absent.
> It is said in section 8.  Normative Behavior
>
>        Note: This guidance on the handling of messages without a priority
>        does not result in a Diameter agent inserting a DRMP AVP into the
>        message.  Rather, it gives guidance on how that specific
>        transaction should be treated when its priority is compared with
>        other requests.  When a Diameter agent relays the request it will
>        not insert a DRMP AVP with a priority value of 10.
>
> It could be possible to clarify it as follow:
>
> in section 6, the end of the point 2 could be enhanced as follow:
>
>     2.  Agents *handling* the request - Agents use the priority information
>         when making routing decisions.  This can include determining
>         which requests to route first, which requests to throttle and
>         where the request is routed.  For instance, requests with higher
>         priority might have a lower probability of being throttled.  The
>         mechanism for how the agent determines which requests are
>         throttled is implementation dependent and is outside the scope of
>         this document.  The agent also saves the transaction priority in
>         the transaction state, either as locally managed state or using
>         the Proxy-Info mechanism defined in [RFC6733].  This will be used
>         when handling the associated answer message for the transaction.
>         *Agents are not supposed to modify or include priority information in
>         in forwarded requests or answers.*
SRD> I propose the following reworded last sentence: "Agents generally 
do no modify priority
information and agents generally do not add new priority information in 
forwarded requests or answers."

SRD> There is one scenario where I can see an agent does add add 
priority information, in a
transition period when not all endpoints support the DRMP mechanism and 
the agent
is used to insert priority information for the non supporting endpoint.

SRD> I also see one scenario where an agent might modify priority 
values.  This would
be an edge agent case where the priority values included by another 
operator's Diameter
network aren't trusted and new values are needed.

SRD> I propose that notes that address these scenarios  be added to the 
new normative requirements proposed below.
>
> The "not supposed" is used because it is difficult to use normative wording here.
>
> In section 8, a new requirement could be added, right after " Diameter agents MAY use routing priority information..."
>
>     Diameter agents SHOULD NOT modify or include the DRMP AVP when
>     relaying request and answer messages.
SRD> I propose the following:

    Diameter agents SHOULD NOT modify priority information when relaying
    request and answer messages.

       There might be scenarios where a Diameter agent does modify
       priority information.  For instance, an edge agent might need to
       modify the priority values set by an adjacent operator.

    Diameter agents SHOULD NOT add priority information when relaying
    request and answer messages.

       There might be scenarios where a Diameter endpoint does not
       support the DRMP mechanism and agents insert priority information
       for that non supporting endpoint.

>
> Just a proposal, waiting for Steve and WG comments.
>
> Regards,
>
> Lionel
>
>
>> -----Message d'origine-----
>> De : DiME [mailto:[e-mail]] De la part de Stephen Farrell
>> Envoyé : vendredi 4 mars 2016 18:07
>> À : [e-mail]
>> Objet : [Dime] AD review of draft-ietf-dime-drmp-03
>>
>>
>> Hiya,
>>
>> I just have one question I'd like to ask the wg about before I start IETF LC.
>>
>> You don't say if priorities are intended to be modified after they have been
>> set. In the security considerations you do say that this could be done
>> maliciously, and you do say that priorities need to be dropped if received
>> from a source not trusted for that, but you never say if it's considered ok or
>> not for e.g. an agent to change a priority for some local policy reason. Don't
>> you need to say that somewhere? (And apologies if you do say it somewhere
>> and I missed it:-)
>>
>> There are some nits below, you can handled these before or after IETF LC,
>> whichever is best.
>>
>> Cheers,
>> S.
>>
>>
>> - Section 5: URL and MME aren't expanded. Since you're just using it as an
>> example, I'd say expanding this will help any reader who's not a 3gpp
>> afficionado.
SRD> Change made.
>>
>> - Section 8, "The priority marking scheme SHOULD NOT require the Diameter
>> Agents to understand application specific AVPs."
>> Isn't that a bogus use of 2119 language since we're not expressing
>> requirements here? s/SHOULD NOT/does not/ would seem better.
SRD> Agreed, change made.
>>
>> - Section 8, People will ask "why default to 10?" I recall the WG discussed this
>> but iirc mostly didn't care too much but it might be nice to justify 10 if there's
>> a way to do it that doesn't amount to "just because" :-)
SRD> I'm open to wording suggestions here but the only real reason is 
that we needed a default and some thought it might be better to have the 
default allow for a few more higher-than-default values than 
lower-than-default values.  I'm not sure saying this adds much value.
>>
>> - Section 8, The "When setting and using..." paragraphs are quite verbose.
>> It'd be no harm to make that shorter, e.g. by just saying: "For all integers x,y
>> in [0,15] treat PRIORITY_<x> as lower priority than PRIOIRTY_<y> when y<x"
>> You could do something similar in 9.1.
SRD> The existing language was put in when we had 5 priority values.  
The above is certainly a more elegant way of specifying it.  Changed to 
the following:

    When setting and using priorities, for all integers x,y in [0,15]
    treat PRIORITY_<x> as lower priority than PRIOIRTY_<y> when y<x.

       Note: As a result PRIORITY_0 is the highest priority.

SRD> I'm not sure this can be done in section 9.1, as this is listing 
the enumerated values for the AVP.


>>
>> I-D nits:
>>
>>    == Unused Reference: 'RFC5226'
>>    == Unused Reference: 'RFC4412'
SRD> These references removed.
>
> _________________________________________________________________________________________________________________________
>
> Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
> pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
> a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
> Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.
>
> This message and its attachments may contain confidential or privileged information that may be protected by law;
> they should not be distributed, used or copied without authorisation.
> If you have received this email in error, please notify the sender and delete this message and its attachments.
> As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
> Thank you.
>