Hello, I was wondering what is the overhead of passing data into a topology for processing via a DRPC call vs emitting data into the topology from a regular Spout?
I'd like to define a DRPC call, and use the call argument for injecting data into a Trident topology for further processing down the line, ultimately persisting data into a non transactional TridentState. Of course, with this approach the DRPC call argument would be considerably sized, since batches of data would be emitted via the DRPC argument. But I'm wondering about performance of this approach? Is this actually feasible? Any feedback is appreciated. Thanks, -- Danijel Schiavuzzi E: [email protected] W: www.schiavuzzi.com T: +385989035562 Skype: danijel.schiavuzzi
