Investigate possible in-memory datasink options: * Redis or some other in-memory databases * Loading stream right into Pandas/Polars