Skip to main content

DynamoDB: Change Storage type to Standard-Infrequent Access

Updated over a week ago

Difficulty: Easy

Description

We identified DynamoDB tables in your account with high storage costs relative to access costs. The Standard-Infrequent Access (Standard-IA) table class offers 60% lower storage costs with a small increase in throughput costs (~24%), making it ideal for tables where storage dominates the bill.


Selection Criteria

  • The table uses the STANDARD storage class.

  • The projected net savings is at least 10% of the total annualized table cost.

  • Global Tables are excluded — Standard-IA is not supported for Global Tables.

⚠️ Warning — Conflict with Reserved Capacity recommendation

If the table is in Provisioned Capacity mode AND also qualifies for the Purchase Reserved Capacity recommendation, these two optimizations are mutually exclusive: Standard-IA tables cannot benefit from DynamoDB Reserved Capacity. Both recommendations must be surfaced simultaneously with their respective projected savings so the customer can make an informed choice before either is applied.


Expected Saving

20% to ~50% of the total annualized DynamoDB table cost, depending on the storage-to-throughput ratio.

How it's calculated:

We analyze tables still in STANDARD class from inventory and gather all costs from CUR split by usage type (based on AWS DynamoDB pricing):

| account_id | region | table_name | storage_ann_cost | read_capacity_ann_cost | write_capacity_ann_cost | total_ann_cost | 
+--------------+-----------+--------------------+------------------+------------------------+-------------------------+----------------+
| 123456789012 | us-east-1 | big-data-archive | 9872.45 | 21.34 | 18.76 | 12714.62 |
| 123456789012 | us-west-2 | historical-events | 7654.32 | 11.45 | 9.87

Example — big-data-archive table:

  • Storage Cost: $9,872.45/year

  • Total Throughput Cost: ~$40.10/year (read + write)

  • Storage-to-Throughput Ratio: 246.2 → strong candidate

Pricing change when switching to Standard-IA (us-east-1, as of 2025-09):

  • Storage: $0.25/GB-month → $0.10/GB-month (−60%)

  • Throughput (on-demand reads/writes): (+~24%)

Net savings calculation:

Storage Savings     = $9,872.45 × 0.60 = $5,923.47 
Throughput Increase = $40.10 × 0.25 = $10.03
Net Annual Savings = $5,923.47 − $10.03 ≈ $5,913.44 (46.5% of total cost)

Operational Impact

Downtime: None — the table class change is applied online with no impact on performance, availability, or durability.

Breaking change: None — no application code changes required. The DynamoDB API is identical for both table classes.

Constraint ⚠️: AWS allows a table class change at most twice per 30-day period.

Global Tables: Not eligible — Standard-IA is not supported for tables with Global Table replication enabled. These are automatically excluded.

Recovery: No data risk — the table class change is fully reversible (within the 2-changes-per-30-days limit).


References:

Did this answer your question?