It would also be good to see examples of native I/O being noticeably more performant. The CassandraIO and PubsubIO are examples of corresponding Beam sources missing rather than Beam sources being slow.
I think it is better to look into why Beam IO can not be performant. I think in most cases it can be addressed, there will always be small number of exceptions that work best in a runner. If the performance gap is because of a specific runner design or implementation, then it should not ideally influence Beam API too much. I suspect 'useNative()' would not be that clean to implement. What a Beam source need to provide to runner might vary a lot between the runners. Can we see the API between the source and the runner to support useNative()? On Thu, Apr 28, 2016 at 5:41 AM, Jean-Baptiste Onofré <j...@nanthrax.net> wrote: > * Beam doesn't provide the IO yet (for instance, spark cassandra connector > is available whereas we don't have yet any CassandraIO (I'm working on it > anyway ;))
