...

1. ConfigurableProperties vs ServiceProfile
   1. Chuyi introduced her cognition for CCVPN usecase, in CCVPN case, there are sdwan_topology and other attributes that are related to network configuration, which are more close to the conception of configurableProperties in ServiceDescriptor.
   2. In Borislav said that in the slicing usecase, there has the distinction between CFS and RFS, some are known by customers like SLA or 3GPP serviceProfile which belongs to CFS, some are internally, like slice subnet, which belongs to RFS. In . But in ONAP mode, there is no need to distinguish, just need a collection of these properties. In this case, latency, bandwidth etc., can be in ServiceProfile and ServiceInstanceProfile.
   3. ConfigurableProperties is for describing input parameters which will be  finally configurabled during instantiation, currently, the conception for ServiceProfile tend to be describing user level requirements, including SLA and else, A suggestion according to use cases.
   4. Each service can have its own profile, there should be a specific place to carry those attributes and should support extended new versions, thus this will not require other part of  service model changes each time. The version changes should also be reflected to AAI schema.
   5. ServiceProfile don't need to hold common attributes, every model will holds the specific attributes that are needed. 
   6. A question is if sharing a service, what capabilities the service need to provide, what is needed for this service? These information can be reflected into serviceProfile. 
   7. Suggestion is to merge the two conceptions into one, and change a name for it to avoid confusion.
   8. Another suggestion is we can analysis the ccvpn use case template CCVPN Service Design, and then assist the construction.
2. For the slicing usecase,InstanceSharingNumber,
   1. Need two parameters, one is for the max ability that one service can carry if it can be shared, another is for the real shared time in instance side.
   2.  The remained question is where to put the former parameter, in design time or in run time? 

Recordings

audio_only.m4a

zoom_0.mp4