Understanding the read consistency model

In traditional RDBMS, significant emphasis is placed on preventing various forms of data corruption through mechanisms such as data locks, waiting queries, and ensuring read consistency. These systems provide a fixed set of support for all use cases and workloads. However, in scenarios where only a small portion of data needs updating, database locks can still occur, impacting other incoming requests.

In contrast, NoSQL systems such as DynamoDB have evolved to adopt different measures tailored to specific use cases, focusing on flexibility and scalability.

For instance, in SQL Server databases, data updates could result in the locking of one or more pages of the table, even if the update only involved one specific data row on the page. Similarly, in Postgres, access-exclusive locks might be used by applications to lock the entire table during a particular read operation, causing any other concurrent operations on the table to wait until...