This section explains how driver internals work. The intended audience is:
driver developers and contributors;
framework authors, or architects who want to write advanced customizations and integrations.
Most of this material will involve “internal” packages; see API conventions for more explanations.
We recommend reading about the common infrastructure first. Then the documentation goes from lowest to highest level:
Native protocol layer: binary encoding of the TCP payloads;
Netty pipeline: networking and low-level stream management;
Request execution: higher-level handling of user requests and responses;
Administrative tasks: everything else (cluster state and metadata).
If you’re reading this on GitHub, the
.nav file in each directory contains a suggested order.