Amazon DynamoDB in thirty-three slides: a table of items found by their keys, beside a relational database; the partition key, the sort key, partitions and hot keys, Query and Scan; global and local secondary indexes; read and write units, on-demand and provisioned capacity and what each costs; DAX, Streams and TTL; global tables and point-in-time recovery; then the purpose-built databases (Redshift, DocumentDB, Neptune, Keyspaces, Timestream, OpenSearch) and how to choose one from a workload. Every section ends with five exam questions.
The slides
The tool is thirty-three slides: an introduction, then six sections, listed on its left. Every section ends with five
exam questions, each with its answer and the tempting wrong answer.
- 0. Introduction
- Amazon DynamoDB: a table of partitions, each in three AZs, behind an HTTPS API.
- DynamoDB beside a relational database: one SQL join on RDS against one
GetItem.
- 1. Tables, items and keys
- Tables, items and attributes: the lab's table, and one item in DynamoDB's JSON.
- The partition key: the hash that picks a partition.
- The sort key: an item collection, and a Query over a range of it.
- Partitions and hot keys: 3,000 read and 1,000 write units a second per partition.
- Query and Scan: the lab's numbers, 0.5 read units against 2.0.
- 2. Secondary indexes
- Global secondary indexes, local secondary indexes, and GSI or LSI.
- 3. Capacity and its cost
- Read and write units: 4 KB, 1 KB, half for eventual consistency, double for transactions.
- On-demand capacity and provisioned capacity and auto scaling, with their us-east-1 prices.
- The cost of the two modes: steady traffic and a few busy hours a day, and the 29% break-even.
- 4. DAX, Streams and TTL
- DynamoDB Accelerator (DAX), DynamoDB Streams and time to live (TTL).
- 5. Global tables and backups
- Global tables: MREC and MRSC.
- Point-in-time recovery and backups: any second of the last 35 days, into a new table.
- 6. Choosing a database
- Purpose-built databases: the map.
- Amazon Redshift, Amazon DocumentDB, Amazon Neptune, Amazon Keyspaces, Amazon Timestream and Amazon OpenSearch Service: what each is, and when to choose it.
- Choosing a database by the workload: the exam's words, and the service each points to.
The buttons on the left and right edges move between slides. So do the arrow keys and a swipe.
The lab
Amazon DynamoDB Lab: the Session Table practices it.
The prices are us-east-1's, read from AWS's Price List API on 2026-11-16: $0.625 per million write request units,
$0.125 per million read request units, $0.00065 per write capacity unit-hour, $0.00013 per read capacity unit-hour.
The cost slide counts writes alone over 730 hours. The consumed capacity on Query and Scan is the lab's, worked out
from the Developer Guide's rules for 200 items of about 65 bytes.
Sources