How to implement a new driver?

CUBRID client applications can communicate directly over TCP with CUBRID by implementing the CUBRID communication protocol. Examples of drivers implementing this protocol are CCI (cubrid C driver), JDBC and the ADO.NET driver.

If your situation allows it, rather than implementing a new driver from scratch, it is recommended that you create a wrapper over one of the exiting drivers. For example, the php and python drivers are built over the CCI driver. The advantage of this approach is that drivers that are already released are maintained by the CUBRID team (CCI for example) and you can add support to new versions of CUBRID to your driver without having to understand how the internals of CUBRID have changed in the new release.

node.js is written in C++ i think, so the best way to write a CUBRID driver for their data API is to create a wrapper over the CCI driver.