So, MySQL néé Drizzle has always had a pluggable storage engine architecture.

Check up on InnoDB, Berkely DB, myISAM and the others.

These are all essentially BTree, and KeyValue stores, which when combined with the MySQL engine brings us SQL.

So… What about a pluggable MongoDB storage driver for Drizzle (since MySQL is Oracle’ised and dead).

This would give great scalability, sharding, on top of a well known layer (SQL) for people that cannot make the NoSQL shift, and in fact the combination of NoSQL and SQL might just make great sense for NoSQL people too.

Yes, Drizzle and SQL has schemas, but MongoDB doesn’t care, and neither did InnoDB and the others, really. The Drizzle MongoDb storage driver could take care of a schema and enforce it, or not. And imagine an “alter table” taking 100 milliseconds when defaulting to null across a few million rows…

Remember. I called it first 😉

** update **
Something like http://www.rabbitreplication.org/ is actually neat but sort of the reverse way I envisioned things, but useful nonetheless.