CockroachDB offers two main modes for executing SQL queries: local and distributed. The local execution engine processes queries primarily on the node that receives them, operating sequentially but allowing for parallelized local processing of multiple data updates through SQL syntax extensions like RETURNING NOTHING. This approach is simple and trusted by developers for its correctness and speed when data is locally available. The distributed execution engine, on the other hand, delegates parts of a query to multiple nodes in the cluster, aiming to leverage proximity to data to expedite processing and reduce network traffic. This engine is designed to simplify development by providing a general-purpose distributed computing solution that removes the complexity of managing performance and consistency across multiple systems. While the distributed engine currently enhances SQL query performance for tasks such as filtering, joins, and aggregations, CockroachDB plans to expand its functionality further, supported by ongoing work on theoretical challenges related to distributed processing. Both execution modes can operate simultaneously, with distributed execution being the default for eligible queries, encouraging users to explore its benefits.