simple performance comparison between mongodb and postgresql in feathers #2368
haniyasin
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I am not a fan of benchmarks just wanted to have a feeling about the behavior as a whole for the system (DB+Driver+ORM/ODM+ service)
I am researching which DB tech i should use for a planned project this will not and should not be how i will be choosing, it is only for assisting my and maybe your decision , i have researched Mongodb, Postgresql, ArangoDB, ArangoDB came by my mind due to being able to store graph info in addition to documents , and they have made a benchmark back in 2018 showing they had almost the best performance for variant DB technologies
Assumptions A:
i had a sample data (actual measurements) for some sensors around 39.5 Million records no joins, flat document
index made on the "address" column
Testing Query execution time concurrency is not being tested/stressed
All DB technologies are stored on the same local machine (VM )
no special tuning was made to any technology, except for the created indexes
specs for the bare metal is ESXi 6.5 U3
PowerEdge R620, 16 CPUs x Intel(R) Xeon(R) CPU E5-2660 0 @ 2.20GHz, 63.94 GB RAM
VM specs: Ubuntu server 20.04 , 16 vCPU, 32 GM RAM
first test made by querying using mongo shell , postgres PGadmin, and Arango browser interface , i eliminated ArangoDb not only because of the result but also for other reasons that are out of the scope of this experiment.
PG had better query execution time by a magnitude of around 2x
(not important test just a warm up)
since i am planning to use feathers my next step is to test the driver+ORM/ODM if applicable+ DB and how the results will vary
Assumptions B:
Test1: Mongodb Driver , feathers service d-2
Test2: Postgres -Sequelize , feathers service d-3
Test3: postgres - objection, feathers service d-4 (D4-had some duplicate data so total being queried is higher)
Limit is 1000 (being practical ++)
test is made by filtering a sensor address: [00:13:A2:00:41:BB:7C:83]!
test is made using chrome dev tools Network section
Data for any one interested https://drive.google.com/file/d/1OJiK7hASwXOzgSCI7lulydEgTX5R98WP/view?usp=sharing
Beta Was this translation helpful? Give feedback.
All reactions