Driver to store streaming API

[Toshbrown]

When performing a bulk import or dealing with high data rate data sources using the REST API is inefficient.

We support streaming data for stores --> app (Reading) but not for Driver --> Store (Writing)

[yousefamar]

Perhaps one approach would be for stores to recognise "keep-alive" headers and maintain persistent connections (as opposed to another WS API).

[cgreenhalgh]

IMO allowing multiple/bulk add in a single request is more important than making individual requests more efficient. This is a clear win for bulk import, but also allows high data-rate sources to do their own chunking of data.

[Toshbrown]

This will be addressed by the new store #104