什么是大规模系统的最佳方法-DB过程或编程逻辑

Question

I am really caught into a puzzle.

We use some language (java / php /python) as mediator between the frontend and db. We use JPA / hibernate to smooth this data flow. If we check with any DBA, most of the data processing, calculations, job processing can be done directly at db side with help of procedure or triggers that too with good scalability. Same thing if handled at java / php side it takes rather longer time and again issue might come up with scalability as transaction handling is not much straight forward.

What I want to clear about my thought is : 1] is it really better / right approach to handle most possible logic with DB procedures ( like handling big chunk of processing impacting many tables ) ? 2] Can scalability be acheivied with DB which can't be with Coding? 3] Is there any better way to look at this DB and Coding relation ?

Thanks in advance.

Regards

Answer 1

Well, based on what you said, you got them right,

However, it is not always that you will have a correct answer, for example, I am working on a system with 6M+ records, you need to analyze the system and check the database performance, when you start, the performance will be good, but as the system scales and you started to have more and more records, here you will know if you can keep the triggers and everything in the database or you will need to hard code them, but from what I can see, you are good to have the calculations done in the database since it is shared between 3 systems.

So my recommendations will be, keep the calculations in the Database unless you have to move them due to performance or any factor.