A service that stands on top of a repository and intermediates access to it, as I understand.
This sounds a little bit like what I was proposing putting on top of the centralized datastore proposed by Nick, so that people don't just access the data directly.
This also seems a bit like:
In a large organisation, different teams can each publish their data to Kafka. Each team can independently develop and maintain stream processing jobs that consume streams and produce new streams. Since a stream can have any number of independent consumers, no coordination is required to set up a new consumer.