Re: [Lime-oam-model] Design Team report

["Deepak Kumar (dekumar)" <[e-mail]>]

Hi Tom,

Good point. p2mp can't exist without multicast stack support but as we are
building rpc for tool in generic model I have used common tools (ping,
trace and defect indication (icmp error(s))) and technology specific Yang
can extend these generic tools and add new tools.

This is the reason I visualize about topology to be empty as from generic
model meaning generic model is not providing or assuming any topology
details or restrictions as this can vary from switch to switch also
depending on implementation. Ping/trace rpc in generic model should work
on Addressing itself without any topology details.

But as topology is defined as choice case, we can extend different
topology type easily. Do you see we need to provide any topology specific
restriction in generic model? If yes, then I have no issue extending cases
for different kind of topologies.

Thanks,
Deepak

On 3/20/15 5:31 AM, "Tom Taylor" <[e-mail]> wrote:

>As Greg said, that is the opposite of what I intended. It does raise an
>interesting question: is a service like p2mp an IP service or a service
>provided by a different layer?
>
>Tom
>
>On 19/03/2015 9:34 PM, Deepak Kumar (dekumar) wrote:
>> Hi Greg,
>>
>> I was referring to following model, where topology is defined as empty
>> in generic model and can be extended by technology specific OAM in their
>> own Yang model.
>>
>> Grouping topology {
>>      Choice topology {
>>        Case topo-null {
>>          Description
>>            "this is a placeholder when no topology is needed";
>>          Leaf topo-null {
>>             Type empty;
>>             Descroption
>>               "there is no topology define, it will be define in
>> technology specific."
>>          }
>>        }
>>      }
>> }
>>
>>
>> Grouping MEP {
>>
>>      Leaf mep-name {
>>      }
>>      Uses MEP-ID;
>>      Uses mp-address;
>>      Leaf interface {
>>      }
>>      Uses topology;
>> }
>>
>> Thanks,
>> Deepak
>>
>>
>>
>>
>> On 3/19/15 8:22 AM, "Gregory Mirsky" <[e-mail]
>> <mailto:[e-mail]>> wrote:
>>
>>     Hi Deepak,
>>     I thought I agree with Tom that topology of monitored object, e.g.
>>     p2p or mp2mp, is technology-independent OAM-wise but your view puts
>>     topology against ... topology. Perhaps you can illustrate or give
>>     definition to what you view as technology-independent topology and
>>     how it is different from the topology of monitored object.
>>
>>     Regards,
>>     Greg
>>
>>     -----Original Message-----
>>     From: Deepak Kumar (dekumar) [mailto:[e-mail]]
>>     Sent: Wednesday, March 18, 2015 7:21 PM
>>     To: Tom Taylor; Qin Wu; Gregory Mirsky; Ronald Bonica;
>>     [e-mail] <mailto:[e-mail]>
>>     Subject: Re: [Lime-oam-model] Design Team report
>>
>>     I agree with Tom, Topology is technology-independent and should
>>     belong to generic model and expand it in technology specific model
>>     with specific like p2p, p2mp, etc..
>>
>>     Thanks,
>>     Deepak
>>
>>     On 3/18/15 11:23 AM, "Tom Taylor" <[e-mail]
>>     <mailto:[e-mail]>> wrote:
>>
>>         Below with [PTT].
>>
>>         On 17/03/2015 10:57 PM, Qin Wu wrote:
>>
>>             Hi, Greg:
>>
>>         ...
>>
>>
>>             [Qin]: Do you proposed to add p2p and p2mp,bi-directional,
>>             uni-direcational as part of LIME generic model, or add them
>>             as part
>>             of technology-specific data model extensions?
>>
>>         [PTT] I would think topology is technology-independent as an
>>         abstract
>>         concept, and therefore belongs in the generic model. To bring
>>         that in
>>         line with reality, there would have to be interplay with the
>>         technology-specific models to indicate which topologies these
>>         models
>>         support.
>>
>>         ...
>>
>>
>>
>
>_______________________________________________
>Lime-oam-model mailing list
>[e-mail]
>https://www.ietf.org/mailman/listinfo/lime-oam-model