Benchmarks in #14 are still about 40 times slower in comparison with the direct use of CassandraCSharpDriver on my testing machine. It seems to me that EF should not cause this much overhead but needs to be something with the configuration.
E.g., in other benchmarks, I use the following Cluster configuration to eliminate drivers/Cassandra protocols safeguards.
Is there a way to access
Cassandra.Builder
or modify low-level CassandraCSharpDriver's connection properties?It would be beneficial for #12 and further benchmarking (https://github.com/simpleidserver/EFCore.Cassandra/pull/14) to play with it.
Benchmarks in #14 are still about 40 times slower in comparison with the direct use of CassandraCSharpDriver on my testing machine. It seems to me that EF should not cause this much overhead but needs to be something with the configuration.
E.g., in other benchmarks, I use the following Cluster configuration to eliminate drivers/Cassandra protocols safeguards.