Cloud-Native Storage Tiering: A Working Cost-vs-Latency Map
Storage tiers are mostly a cost-vs-latency tradeoff. Picking by gut wastes money; picking by access pattern saves 40-70%.
The four-tier model
Hot: queryable in milliseconds, expensive ($23/TB/mo S3 Standard).
Warm: queryable in seconds, cheaper ($12/TB/mo S3 Standard-IA).
Cold: queryable in minutes-to-hours, very cheap ($4/TB/mo S3 Glacier).
Archive: queryable in hours, near-free ($1/TB/mo S3 Glacier Deep Archive).
Per-cloud equivalents
- AWS: S3 Standard / Standard-IA / Glacier Instant / Glacier Deep Archive.
- GCP: Cloud Storage Standard / Nearline / Coldline / Archive.
- Azure: Blob Hot / Cool / Cold / Archive.
- Names differ; the tier model is identical.
Access-pattern analysis
Pull last-90-day access logs for the bucket. Files accessed weekly: hot. Monthly: warm. Quarterly or less: cold. Never: archive (or candidate for deletion).
Most buckets discover 60-80% of data is cold-or-colder. The tiering decision is mechanical once you have the data.
Lifecycle policies
Set lifecycle policies that auto-tier based on age. After 30 days → warm. After 90 days → cold. After 365 days → archive (or delete with policy approval).
The savings compound monthly. The policy runs forever.
Antipatterns
- Manual tiering. Always falls behind.
- Tiering without retrieval-cost awareness. Cold tiers charge per GB retrieved; check the math.
- Tiering everything aggressively. Hot data on cold tier kills application performance.
What to do this week
Three moves. (1) Pick the most exposed instance of the pattern in your environment. (2) Apply the lightest fix and measure for one week. (3) Schedule a quarterly review so the discipline does not rot.