Standardize mechanism to specify a "Digital Shadow"

[mmccool]

A common pattern would be to have a TD for a real device and a very similar TD (or alternative forms in the same TD) for a "shadow" service that caches/mediates interactions with that Thing, aka a Digital Shadow.

It would be good to define a standard way to represent this pattern, e.g. perhaps a specific type (e.g. "Shadow") and a link with a specific relation type, and/or for the "multiple form" approach some semantic tags (maybe "Shadow" again) to annotate which forms point at the shadow and which at the real device.

[egekorkan]

I would ideally separate them since the shadow can have additional stuff on top and can behave differently. E.g. a shadow of a robot can do input checking thanks to its simulation and path planning capabilities and it would be weird to see different behaviour in two forms. Also, they are technically two different entities, so two TDs.

Also, I am not super sure about the "Digital Shadow" term as a digital twin can even act on the real Thing. For example, we can change the simulation and say/click replicate for the same actions to happen on the Thing and that shadow would have already additional capabilities that the Thing does not have.