The current REST Proxy is implemented on top of Java. Having the REST proxy has opened up MapR Streams to a huge range of new use cases, and that's great.
However, MapR's greatest differentiators are built on unmatched natively implemented stack, from MapR FS to MapR-DB and MapR Streams.
Now that the first version is out the door, it would make a lot of sense to me to move the proxy over to a native implementation. Lets see what the true performance potential of MapR Streams really is!
Since we're already pushing out a C kafka library for MapR Streams, this is probably not the hardest project in the world to realize...