Module Review: Core Concepts

In this module, we explored the foundational architecture of DynamoDB. We learned how data is distributed across partitions using Consistent Hashing, how throughput is provisioned via RCUs and WCUs, and the trade-offs between Eventual and Strong Consistency.

Key Takeaways

  1. Architecture: DynamoDB uses Consistent Hashing to distribute data across physical partitions based on your Partition Key.
  2. Scaling: Scaling is horizontal. Partitions split automatically when they exceed 10GB storage or throughput limits (~3k RCU / 1k WCU).
  3. Throughput:
    • RCU: 1 read/sec for 4KB item (Strong). Eventual reads are half price (0.5 RCU).
    • WCU: 1 write/sec for 1KB item.
  4. Consistency:
    • Eventual (Default): Fast, high availability, potential for stale reads.
    • Strong: Guaranteed latest data, higher latency, costs 2x.
  5. Pricing:
    • Provisioned: Best for steady state.
    • On-Demand: Best for unpredictable spikes.
  6. Glossary Link: DynamoDB Glossary

Flashcards

Cheat Sheet

Mode Reads (RCU) Writes (WCU) Consistency
Provisioned /hour (Reserved) /hour (Reserved) Configurable
On-Demand /request /request Configurable
Operation Unit Cost (4KB Read / 1KB Write)
Eventual Read 0.5 RCU
Strong Read 1.0 RCU
Transactional Read 2.0 RCU
Standard Write 1.0 WCU
Transactional Write 2.0 WCU