Loading MongoDB Atlas collection into ram for each network request

I have a fixed size dataset (it won’t grow).

Here are the figures:

STORAGE SIZE: 190.13MB
LOGICAL DATA SIZE: 253.71MB
TOTAL DOCUMENTS: 53452
INDEXES TOTAL SIZE: 1.44MB

I don’t like the fact that I am loading the whole dataset into RAM for each network request

const data = await collection.find().toArray();

I could cache the table contents with a global variable.
I would like to know what are the optimal solutions for this case?
FYI. I need the entire table for each request to do comparator logic.