I find it amusing that decades-old techniques that have been in use in the financial trading industry finally made it into the mainstream. For example, proprietary systems like Kx Systems and OneTick have been providing stream processing, continuous query, and historical data store for decades. Similarly, open-source projects like Esper began many years ago (although it does not have persistent stores). I personally built something similar (with home-grown columnar data store) during my tenure at BGI/BlackRock many years ago.
I am not discrediting the importance of Kafka and the work that Confluence is doing. I do want people to know that this type of systems has been in existence for decades.