We need to do kafka integration with DB2. What is the best framework for real time streaming among spark structured streaming and kafka streams ?
I'd generally stay away from custom solutions like Kafka Streaming because it's a very new product. It's bound to miss features or have some bugs or not work 100% like you want, leaving you little options better than waiting for the next release. I'd only chose it if the feature set is absolutely perfect for your needs AND that you understand your needs perfectly AND they will not change in the future.
Spark Structured Streaming GA from Spark 2.2.0 isn't currently supported on MapR, but our engineering is working on it very hard for compatibility with MapR ES. The API is really nice and, more importantly, lets your developers re-use their existing Spark knowledge. Spark community have demonstrated lots of commitment to streaming. This is the option I personally would prefer because of the flexibility & Sparks much bigger community and skill availability.
Retrieving data ...