Re: [pcp] Capability discovery

Re-,

Mentioning the two ways is OK.
Still the main point is to assess the benefits of this option/Opcode. If in your draft you can discuss how it can be useful for the authentication case, this would be a good input.

Cheers,
Med 

>-----Message d'origine-----
>De : Yoshihiro Ohba [mailto:yoshihiro.ohba@toshiba.co.jp] 
>Envoyé : vendredi 5 octobre 2012 08:57
>À : BOUCADAIR Mohamed OLNC/OLN
>Cc : pcp@ietf.org
>Objet : Re: [pcp] Capability discovery
>
>So there are two ways for capability discovery:
>
>- Use a PCP Opcode
>- Use a PCP Option
>
>If the WG has not decided which way to use for capability discovery,
>I would have to mention the two ways in the two PANA I-Ds I am 
>working on.
>
>Is this OK?
>
>Yoshihiro Ohba
>
>(2012/10/05 14:56), mohamed.boucadair@orange.com wrote:
>> Dear Yoshihiro,
>>
>> FYI, this document proposes an option to retrieve the 
>capabilities of the PCP-controlled device:
>> http://tools.ietf.org/html/draft-boucadair-pcp-capability-00
>>
>> The current version does only describe whether the 
>controlled device is NAT44, NAT46, NAT64, IPv4 FW, IPv6 FW, 
>Port Range Router.
>>
>> This feature is defined as an option but can be elected to 
>be defined as standalone OpCode. The capabilities set format 
>can be modified to support new capabilities: e.g., 
>authentication support, list of supported opcodes, etc.
>>
>> Saying that, IMHO we need further analyze what are the gains 
>and the impacts on the PCP server to support such feature:
>>
>> * Is it reducing exchanged messages?
>> * Does it harm if we try an opcode no matter if the PCP 
>server supports it or not?
>> * Reduce delay before establishing a session?
>> * Be a trigger to decide whether all available PCP Servers 
>need to be contacted in // or select only one?
>>
>>
>> Cheers,
>> Med
>>
>>> -----Message d'origine-----
>>> De : pcp-bounces@ietf.org [mailto:pcp-bounces@ietf.org] De la
>>> part de Yoshihiro Ohba
>>> Envoyé : vendredi 5 octobre 2012 05:04
>>> À : pcp@ietf.org
>>> Objet : [pcp] Capability discovery
>>>
>>> There has been a question on what to do when PCP client supports PCP
>>> authentication
>>> while PCP server does not, and vise versa. The same issue will
>>> exist for
>>> future PCP extensions.
>>>
>>> I would like to hear opinions whether defining a capability 
>discovery
>>> exchange
>>> in PCP base specification ever makes sense, where the capability
>>> discovery exchange is
>>> expected to happen prior to any other PCP opcodes.
>>>
>>> Best Regards,
>>> Yoshihiro Ohba
>>>
>>>
>>>
>>> _______________________________________________
>>> pcp mailing list
>>> pcp@ietf.org
>>> https://www.ietf.org/mailman/listinfo/pcp
>>>
>
>