Compressed Data Transfer (CDT)

During CDT, the row data between remote DA agent (excluding local DA agent) and DA server gets compressed thereby improving overall comparison time in distributed environments that have high network latency.

Compressed data includes:

All row data transmitted during the initial row or key comparison
All retries of row or key comparison
Verified differences of row or key comparison

CDT is optional for row comparison and not supported for schema comparison. Hashes do not compress well; the initial comparison and the in-flight data comparisons do not see much benefit in using CDT when the columns use the "column_hash" or "row_hash" option. However, literal data compresses well. CDT is beneficial to the initial comparison and in-flight data comparisons when the columns use the "literal" column option. The verify differences phase always benefits from using CDT.

Consider CDT when:

Local area network (LAN) or wide area network (WAN) is a bottleneck
Performing literal comparisons
You expect hundreds or thousands (or more) of differences
Your primary key column is large (as key columns are never hashed)

Do not choose CDT when:

You use a local DA agent.
LAN or WAN performance is not an issue.
You always use hashing and either never use the “verify differences” option or you use it, but expect few or no differences.

To use this option, set compress_data_transfer to true.