Recently researched three localized database and mysql to do comparison, research the main performance indicators are large data write speed, large data read speed and whether to support split table.
I. Test results
The test results were very different from what was expected
1, first of all, oceanBase community edition of this database official online description of the performance is very good, the feeling can kill mysql, but the actual test results of a single node, and mysql data is comparable to, or even a little worse.
2, I also like Damon database, because he imitated oracle to retain many of the characteristics of oracle, I thought his performance is certainly better than the free oceanBase community version and mysql community version, the results of the Damon database to start the speed is really okay, but with the increase in the amount of data is getting slower and slower a little bit out of the blue, to the ten million level insertion speed slowed down a lot! The speed of insertion is much slower at the 10 million level.
3, The Golden Silo database was expected to be on par with Damon, but it was a big letdown.
4, the write speed of the gold warehouse can not, read performance is much better than any other database.
5, charged actually than the free haha unexpected.
Note: I like oracle very much, but unfortunately not domestic, it is not tossed to test it, not to find the best performance of his more embarrassing ha ha.
Disclaimer: The results of this experiment are only for this experiment, can not illustrate the advantages and disadvantages of various databases, the database of this experiment are installed in accordance with the default configuration, did not do any optimization of any kind of data.
The python test code for these four databases will be posted in a subsequent blog (the actual tables and ip will be changed), interested partners can see how to operate these four databases using python.
II. Test methods
Environment description: single node for all kinds of data
1. Test framework diagram
2、Write performance
From the same mysql database in a table to read 14.8W of data (about 20 fields), after reading, respectively, to mysql, oceanBbase, daemon, gold warehouse four databases batch write (python program to achieve), ten million of data will be written to the process repeated eight times.
3, read performance
Read 10 million data to four databases: mysql, oceanBbase, Damon, and Gold Warehouse (python implementation).
Ranking of domestic popularity of domestic databases
/dbRank
Full test code to get:
(1) Login-Register:/
(2) Search: Performance Comparison of Domestic Databases oceanBbase, Damon, Jinchang and mysql Database