The word "true" way "much like google cloud platform dataflow".
In reality, we started out from google cloud platfrom. Both dataproc spark and dataflow beam have been applicants for batch processing, eventually we picked dataflow due to the fact
1. Beam is simple to put in writing and nicely-abstracted
2.Dataflow is fee-effective for our sparse execution sample.
3.Dataproc calls for operation.
We've plan extend system to different cloud companies and we need to locate options offerings of dataflow. Kinesis statistics analytics with flink runner is supposed to be a terrific answer in AWS. But we nonetheless try to find dataflow like carrier.
Subscribe to get latest updates