DynamoDB is a key-value and document database that you read by key, while a relational database joins normalized tables with SQL.
DynamoDB
Relational database
Schema
NoSQL, schema-less apart from the primary key
SQL, fixed schema
Scaling
Horizontal scaling across partitions
Vertical scaling, primarily
Consistency
Eventually consistent reads by default, with strongly consistent reads and ACID transactions on request
ACID consistency
Joins
No complex joins
Complex joins supported
Data model
Key-value and document model
Relational model
Operations
Managed service
Requires administration
Pricing
Pay per request or for provisioned capacity
Pay for instance time
Because only the primary key is fixed, each DynamoDB item can carry its own attributes. Efficient reads name a key value through GetItem or Query, so you design tables around known access patterns. Choose DynamoDB when those patterns are known and need steady low latency at any scale. A relational database fits better when queries are ad hoc or join many tables.
Rewriting in plainer words…
This answer doesn't lend itself to a diagram - it reads best . No credits were charged.
