Which DBMS to choose for a large amount of financial data?

Question

linux

It is necessary to choose a DBMS to store a large amount of financial information.
A close example is stock quotes: each entry in the main table contains a date, time to the nearest millisecond or better, a fractional value, and some overhead information. Millions of records will be added every day. Information will be read about once every 2 weeks, but will be recorded constantly.
The system will work under Linux, the main control software will be written in C ++ and Rust. Operational work with data will be carried out by other means, the question is about long-term storage.
What DBMS do you recommend?

Answer 1

yes, the choice is not great
from sql, this is mysql postgresql
sharding from nosql autosharding is convenient on mongodb

Answer 2

IMHO, each task should be solved by a specialist in solving similar problems. Advice is, of course, good, we still have a "country of advice", but in making a serious decision, which still has to be accompanied later, it is better to play it safe and even choosing a decision on advice, I would recommend checking it several more times and consulting again with competent people. Why am I a skeptic, but because, for example, answering a similar question in an abstract way, I will give one answer, and if the task has to be solved / accompanied by me myself, the answer may become completely different =)

Answer 3

Cyclic database (RRD)

- a database, the amount of stored data of which does not change over time, since the number of records is constant, in the process of saving data they are used cyclically.[

Answer 4

I would first decide whether the data is normalized or not? If not, then you can look towards NoSQL. There it is convenient to make schemes even for one record. And at the same time, there is an option to use the built-in analytics like elasticsearch.

Answer 5

First, come up with a database architecture, approximate sizes of tables, the most frequent actions with the database, their number, then you can advise something.