Basic Information
| Field | Details |
|---|---|
| English name | DynamoDB Accelerator (DAX) |
| Full name | DynamoDB Accelerator (DAX) |
| Chinese description | DynamoDB 专用内存缓存 |
| Japanese description | DynamoDB 専用インメモリキャッシュ |
| Exam frequency | ⭐⭐⭐⭐ |
| Often confused with | ElastiCache / RDS Read Replica |
In one sentence
DAX sits between an application and DynamoDB and caches repeated key-based reads; DynamoDB remains the durable source.
Core purpose
- Applications use the DAX client; hits return directly and misses read DynamoDB then populate cache.
- It fits read-heavy, repeated-key, and bursty workloads.
- A multi-AZ DAX cluster can improve cache-layer availability.
Exam focus
- DynamoDB-only cache and microsecond reads point to DAX.
- Strongly consistent reads, write-heavy workloads, or low hit rates do not fit DAX well.
- For RDS, arbitrary APIs, or sessions, usually consider ElastiCache.
Common misconceptions
- DAX does not replace DynamoDB.
- Microsecond latency does not guarantee that every request is a cache hit.
Key takeaway
DynamoDB + repeated eventually consistent reads + microseconds → DAX.
Related services
DynamoDB, ElastiCache, Global Tables, Lambda.