Highlight the Native DB overhead compared to direct access to the underlying database (redb).
The benchmarks ignore:
native_model
overhead.- Serialization overhead used by
native_model
likebincode
,postcard
etc. - The fact that
redb
can write the data using zero-copy.
If you want to run the benchmarks, you need to install just, nushell and run:
just bench
We can see that the overhead is very low. These result are obtained with the version 0.5.3
of native_db
.