We do have a case for a topology that has multiple "topology types": we are 
defining a topology-type "sr-topoloyg" which is derived from 
"l3-unicast-igp-topology". We are also defining a topology of type "SR TE 
topology", which type is both "sr-topology" and "te-topology", hence 
multiple-inheritance is needed. 

Thanks,

- Xufeng

> -----Original Message-----
> From: Robert Varga [mailto:[email protected]]
> Sent: Tuesday, July 12, 2016 12:54 PM
> To: Alexander Clemm (alex) <[email protected]>; Zhangxian (Xian)
> <[email protected]>; [email protected]
> Cc: [email protected]
> Subject: Re: [i2rs] network-type: container vs. identity?
> 
> On 07/12/2016 06:29 PM, Alexander Clemm (alex) wrote:
> > Per Robert's point, if for some reason you have multiple traits and want to
> "compose" a network to encompass multiple types, this is straightforward to do
> with the current pattern, not so straightforward with identities.
> 
> There is also an issue of consistency within the models. While it may not be
> obvious to have more 'topology types', it certainly makes sense for nodes 
> within
> a topology -- for example in ODL we tag topology nodes with capabilities based
> on the southbound protocol (like PCEP, OpenFlow, NETCONF).
> 
> Bye,
> Robert

_______________________________________________
i2rs mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2rs

Reply via email to