Which consistency option is not available when distributing data across multiple global regions?

Strong consistency is not available when distributing data across multiple global regions. This is primarily due to the inherent challenges of maintaining strong consistency in a geographically distributed system, where data is replicated across various locations.

Strong consistency ensures that any operation on the data reflects the most recent write across all nodes, making it impossible to have any delay in seeing changes made from different regions. In a distributed environment, especially across wide geographical areas, the network latency can lead to performance bottlenecks and increased complexity in coordination, making strong consistency less feasible.

In contrast, eventual consistency, bounded staleness, and session consistency are designed to allow for more flexible consistency models that are more manageable in a distributed setup. Eventual consistency allows different copies of the data to become consistent over time, proving beneficial for performance and availability. Bounded staleness permits a specified lag in data visibility, while session consistency provides a guarantee within the context of a single session, all of which suit the needs of applications operating across multiple regions where immediate synchronization of all data is not practical.