Regarding the next dispute on the topic – “where to store business logic – in the DBMS or backend?”

I wonder if comparative tests were conducted? productivity information system – before and after transferring business logic from the DBMS level to the backend level?

Theoretically, it is probably possible to make synthetic tests. It would be very interesting to compare the performance. There should definitely be a difference. But I have not yet come across such works. Probably because the topic is purely academic, and R&D in the field of DBMS and information technology in general is not very good.

Yesterday, again about the sore subject, I came across a lecture by an associate professor on the topic of “DBMS Performance” – as usual – there is a lecture, but there is no definition of what “performance” is. The lecture talks about query plans, types of connections and access methods. What DBMS performance has to do with it is unclear.

But another question is more interesting: have calculations been made of the cost of ownership and maintenance of information systems with different levels of business logic storage, for example, in the most general, simplest case:

1. Business logic in DBMS: 1 DBMS server + 1 middle level server

2. Business logic in the backend: 1 DBMS server + 1 application server.

I am tormented by vague thoughts doubts:

1. Option 2 – not scalable.

2. Option 2 – will be significantly more expensive.

3. in case of significant growth of volumes and load, option 2 will have problems (this statement, by the way, is actually confirmed by practice – application server resources are increased, the DBMS server is not overloaded)

In general, as they say – the topic is waiting for its researcher. It's a pity that associate professors of the departments of applied mathematics or computer science are engaged in non-scientific work, but in voicing documentation.

And it's a pity that when we were reading systems theory, I was very carried away by C++ and did not pay due attention to fundamental disciplines. Now I have to remember and catch up.

P.S. It is unlikely that the situation will change, it is impossible to explain anything to the modern generation of developers and managers – they are in the trend of “bang, bang and into production”, but from a scientific point of view the topic is probably interesting, maybe someone will take it up.