Core Architecture and Durability Model
S3 is region-scoped. When you create a bucket, you choose a Region. Within that Region, S3 automatically stores your data across multiple Availability Zones.
This is critical.
Durability and Availability
S3 is designed for:
- 99.999999999% durability (11 nines)
- High availability (varies by storage class)
Durability means data will not be lost. Availability means you can access it.
S3 achieves durability by:
- Replicating objects across multiple AZs
- Using distributed storage systems internally
- Continuously verifying data integrity
You do not configure replication between AZs. It is built-in.
Strong Consistency
S3 now provides strong read-after-write consistency.
Architecturally, this means:
- After you PUT an object, a GET immediately returns the latest version.
- No need for complex retry logic in most cases.
This simplifies system design compared to older eventually consistent systems.
Storage Classes
S3 offers multiple storage classes:
- Standard
- Intelligent-Tiering
- Standard-IA
- Glacier classes
Architecturally, storage classes allow cost optimization without changing application code.
You design lifecycle policies, not migration scripts.
Production insight:
- Use lifecycle rules early.
- Avoid manual cost management.
- Plan for growth.
In this section, I learned:
0 of 4 completed