We are currently using Java async driver with Neo4j 4.0.
I had a look at your presentation here where performance results, notably for memory consumptions, are presented to be very better with new Reactive driver.
I would like to know how far is the performance gap between the async and reactive drivers for READ queries with a significant amount of data.
For what I have understood from the documentation, async driver is just supporting flow control whereas reactive driver is working with a Publisher-Subscriber API under the hood. Tell me if I am wrong, but I understand that async driver is just doing some pull with static fetchSize value for batching records, whereas reactive driver is using dynamic push-pull. Am I right ?
So, can we expect a significant difference of performance between async and reactive drivers, in favor of reactive one when the amount of data is big ?