Allow computation task with a reference to a computation template #292
Description
At the moment the whole computation template is send on each computation request.
By using the prepare-message when the frontends starts, a unique reference to the template can be send back from the Websocket-API. With this reference the computation template can be omitted in the computation message.
This increases scalability especially in the Ilias-usecase.
To be discussed:
- When is the computation template destroyed?
- Which tecnology is used to store the computation templates in the websocket-api? in-memory? db?