When running a federated query that joins a large Iceberg table in watsonx.data with an external PostgreSQL table; what technique improves performance?

Presto's pushdown capabilities send filter predicates to the JDBC connector which translates them into WHERE clauses executed in PostgreSQL; minimizing the amount of data transferred over the network to Presto. The incorrect options ("Loading the entire PostgreSQL database into RAM on the Presto coordinator", "Converting the PostgreSQL table to Avro format before the query runs", "Replicating the Iceberg table into PostgreSQL for co-location") are distractors that don't fully capture the concept described.