Open Stack

Isaku Yamahata writes:

> [...]
>> 2) Service Catalog - to accommodate multiple VNF services. Question: the
>> same problem exists with Trove which need a catalog for multiple concrete
>> DB implementations. Do you know which solution they will take for Juno?
>
> Regarding to Trove, I don't know.
> Any Trove developer, can you comment on it?
>

Trove doesn't require multiple entries in the service catalog to support
multiple datastore implementations. Trove internally has a concept of
datastore type (e.g. mysql, redis, etc) and version; the same Trove
service (at the same endpoint) is able to provision DB instances
according to the type and version specified in the instance-create
request.

Additionally, operators also have the option of setting a default type /
version so users don't need to explicitly specify this information as
part of the create request.

Thanks,
Nikhil