Tips and tricks AWS Solutions Architect Associate #13
Access & Security
-
Pre-signed URL: Temporary access (default 7 days).
-
Bucket Policy: Controls access inside S3 (resource-based).
-
Encryption: S3 encrypts every object by default, but no audit/rotation built-in.
Data Access Features
-
Byte-Range Fetch: Retrieve partial object (useful for large files).
-
S3 Select: Query data within objects (CSV, JSON, Parquet).
Costs & Storage Classes
-
Cost hierarchy: S3 < EFS < EBS.
-
Standard: General purpose.
-
Intelligent Tiering: Moves objects automatically.
-
Infrequent Access (IA): Backup use.
-
One Zone-IA: Cheaper, 1 AZ only.
-
Glacier Instant: Retrieval < 6 min.
-
Glacier Flexible: Variable retrieval times.
-
Glacier Deep Archive: Cheapest, retrieval ~12h.
Management & Analysis
-
Storage Lens: Usage and activity metrics.
-
Versioning: Enabled → keeps all versions (non-null versions remain even after delete).
-
Transitions possible (e.g., from IA → Glacier).
-
Data Movement
-
Copy/Sync:
-
s3 sync(local ↔ S3 or S3 ↔ S3). -
Batch replication (S3 → S3, cross-region).
-
-
Transfer Acceleration: Faster uploads to a bucket (not bucket-to-bucket).
-
Multipart Upload: Required > 5 GB.
-
Snowball → S3 → Glacier: For large offline transfers.
Limits & Performance
-
API rates: ~3500 PUT/POST/s, 5500 GET/s per prefix.
-
Bucket operations:
list bucket,delete bucket/*. -
Key Prefixing: Use prefixes to optimize performance.
-
Object URL:
https://<bucket>.s3.<region>.amazonaws.com/<key>.
⚡ In short:
-
S3 = scalable, secure, tiered storage, cheaper than EFS/EBS.
-
Access via pre-signed URLs, policies, or APIs.
-
Optimize costs using storage classes and transitions.
-
Optimize performance with prefixes, multipart uploads, and transfer acceleration.