Hey, I was reading about the data centric design, where the proposed architecture suggests that application code to be less, and more things happening from database systems only. The idea is database will evolve to serve greater needs which essentially means, applications will be heavy on business logic rather than doing mundane (generic) work. My mind immediately went to Pinot. Pinot is serving a greater need out of the box. Is this the trend folks here observe too? What do you think about the future?

Personally, I like to think not in terms of "trends", but what makes more sense to do in each situation. In some situations you have your application doing work with data that comes from, or is processed by, multiple external services, including multiple databases. In this case, it's the application's job to do at least the orchestration for the data processing to happen. In some other situations, however, we end up fetching lots of data from the database to be processed and even aggregated on the application side when it could be done on the database side, and in a fair amount of times this would be better done on the database side so that there's less I/O going on (less data being transferred). Doing aggregations on the database side is usually a great response times reducer.

Irrespective of Pinot. There will always be a constant battle between compute and storage being tightly couple vs loosely coupled. E.g push down compute to storage or pull data to compute.. both have their pros and cons.. In the end, it’s really a battle between how fast cpus, network and ssd evolve.

Agreed. And it's highly dependent on the situation.

I agree, it depends on the situation, I was just thinking about general patterns towards the future.