S3 Standard - Infrequent Access Used Where Intelligent Tiering Would Be Cheaper

Paul Marcelin

CER:

CER-0307

Service Category

Storage

Cloud Provider

AWS

Service Name

AWS S3

Inefficiency Type

Suboptimal Pricing Model

Explanation

Organizations often use the Standard - Infrequent Access (Standard-IA) storage class based on documentation and code that predate 2021 updates to the Intelligent Tiering storage class. Intelligent Tiering became suitable as an initial S3 storage class even for objects that are small and/or will be deleted early. It also gained a heavily-discounted access tier. Older internal runbooks, lifecycle policies (including ones specified in infrastructure-as-code templates), scripts, programs, and public examples may still default to Standard-IA, inflating storage costs.

This inefficiency report compares Standard-IA with Intelligent Tiering. It is not intended to cover other storage classes. S3 storage is billed per gibibyte or GiB (powers of 2) rather than per gigabyte or GB (powers of 10), which matters for small objects and also for large volumes of storage.

Relative to the Standard storage class, the Standard-IA storage class offers a moderate, constant storage price discount but imposes a minimum billable object size of 128 KiB, a minimum storage duration of 30 days, and a per-GiB retrieval charge.

In contrast, AWS updated the Intelligent Tiering storage class in September, 2021, eliminating the minimum storage duration and exempting small objects from a monthly per-object monitoring and automation charge. Intelligent Tiering never had retrieval charges. In November, 2021, AWS added the heavily-discounted Archive Instant Access tier.

For objects stored beyond a few months, Intelligent Tiering's progressive storage price discounts surpass Standard-IA's constant discount. Storage savings accumulate each month. Objects in the Intelligent Tiering storage class automatically move through progressively cheaper access tiers unless the objects are accessed. Intelligent Tiering also avoids Standard-IA's minimum billable object size and minimum storage duration penalties.

Relevant Billing Model

S3 has several billing dimensions, but storage volume (per GiB-month) usually dominates. Storage class determines pricing in each dimension.

S3 Standard-IA offers a reduced per-GiB-month storage price compared to S3 Standard, but imposes constraints: a minimum billable object size (small objects are billed as if they were 128 KiB), a minimum storage duration (even if objects are transitioned or deleted early, storage is billed for 30 days), and a per-GiB retrieval fee (billed each time an object is accessed).
S3 Intelligent Tiering imposes a per-object, per-month monitoring and automation charge for objects 128 KiB and larger, but has no minimum billable object size and no minimum storage duration. Small objects, exempt from the monitoring and automation charge, are always stored in the Frequent Access tier, at the same storage price as if they were in the Standard storage class.
Within the Intelligent Tiering storage class, objects automatically progress (without lifecycle transition charges) through access tiers, from the Frequent Access tier (same storage price as Standard) to the Infrequent Access tier (same storage price as Standard-IA) after 30 days without access, to the Archive Instant Access tier (heavily discounted) after 60 more days (90 days total) without access. If accessed, an object returns to the Frequent Access tier and the countdown restarts. Retrieval does affect cost, even though there is no retrieval charge.

Waste can arise with Standard-IA because Intelligent Tiering's progressive storage price discounts overtake Standard-IA's constant discount, and/or because Intelligent Tiering avoids minimum billable object size and minimum storage duration penalties.

Detection

Identify S3 lifecycle policies that transition objects to Standard-IA, and replication rules that copy objects to Standard-IA, and evaluate whether Intelligent Tiering would save money based on actual access patterns.
Review infrastructure-as-code templates and deployment scripts (including ones that specify lifecycle policies or replication rules), as well as operational scripts and programs that specify an initial storage class, for references to Standard-IA. Pay particular attention to work completed before 2022.
Assess the share of objects in Standard-IA that are significantly smaller than 128 KiB, as these objects incur inflated storage charges, due to the minimum billable object size.
Review retrieval and early deletion charges for objects in Standard-IA, to identify cases where the Intelligent Tiering storage class would be less expensive on balance.
Review internal documentation that recommends the Standard-IA S3 storage class.

Remediation

For new S3 buckets, configure Intelligent Tiering rather than Standard-IA as the initial storage class in scripts, code, and replication rules. In lifecycle transitions from Standard, specify Intelligent Tiering rather than Standard-IA.
Update internal documentation, runbooks, and reference architectures to reflect the 2021 updates to the Intelligent Tiering storage class. AWS eliminated the minimum storage duration, exempted small objects from the monitoring and automation fee, and added the heavily discounted Archive Instant Access tier.
For existing Standard-IA objects, model the impact of switching to Intelligent Tiering by weighing one-time per-object transition charges (unfavorable) against expected per-GiB storage price discounts (favorable, and typically large) plus per-object monitoring and automation charges (unfavorable, but typically small) over time.
If modeling predicts net savings, transition existing objects from Standard-IA to Intelligent Tiering using lifecycle transition rules. Keep in mind that the objects will enter the Intelligent Tiering storage class in its Frequent Access tier (more expensive than Standard-IA, but temporary) and progress to cheaper tiers over time (if not accessed).
Keep using Standard-IA only for workloads with highly predictable, well-understood object size, access, and longevity patterns, for which lifecycle transition and expiration rules can be precisely tuned and retrieval charges, controlled.
Establish periodic reviews of storage class selections across the organization to prevent regression to outdated defaults.

Relevant Documentation

Submit Feedback