A legacy service uses Azure Table Storage. New requirements include global distribution with low-latency reads and writes in multiple regions, automatic indexing of all properties, RU/s-based throughput control, and minimal code changes. What should you do?

The Azure Cosmos DB Table API is the ideal solution because it directly addresses all requirements with minimal code changes. It provides global distribution with low-latency reads and writes across multiple regions (by enabling multi-region writes), automatic indexing of all properties, and RU/s-based throughput control. The existing application can connect to Cosmos DB Table API with just a connection string update, as it's API-compatible with Azure Table Storage. The Data Migration Tool facilitates data transfer. Enabling RA-GRS on existing Azure Table Storage only provides read-only access in secondary regions, not low-latency writes. Migrating to Azure Cosmos DB SQL API would require significant code refactoring and a new data model. Azure SQL Database with JSON columns doesn't offer the same global distribution, automatic indexing, or RU/s control as Cosmos DB, and full-text search is not equivalent to automatic property indexing. Deploying multiple Azure Table Storage accounts with Traffic Manager would not provide automatic indexing or multi-region writes for a single logical table, and would complicate data consistency.