Cost-Optimized Architecture Design · Topic 6 of 8
Storage Tiering, Block Storage and Backup
Cut storage cost with S3 Intelligent-Tiering, lifecycle rules, Glacier classes, EFS lifecycle, gp3 volumes and AWS Backup or DLM retention policies.
- 6 min
- Intelligent-Tiering has no retrieval fees but charges monitoring per object, and objects under 128 KB stay in Frequent Access.
- Lifecycle rules fit predictable patterns and can expire data at end of retention.
- Minimum storage durations are 30 days for IA classes, 90 for Glacier Instant and Flexible Retrieval, and 180 for Deep Archive.
What it is
Storage cost grows quietly. Cold data sits in hot tiers, volumes pay for unused IOPS, and backups never expire.
This page covers the three fixes the exam tests: tier object and file data by access pattern, size block storage to real performance needs, and put retention limits on every backup.
How it works
S3 Intelligent-Tiering
Intelligent-Tiering moves objects between access tiers by observed access, for a small monthly monitoring and automation charge per object. There are no retrieval charges.
| Tier | When objects move there | Access |
|---|---|---|
| Frequent Access | Default on upload | Milliseconds |
| Infrequent Access | After 30 days without access | Milliseconds |
| Archive Instant Access | After 90 days without access | Milliseconds |
| Archive Access (optional) | After at least 90 days, configurable | Minutes to hours |
| Deep Archive Access (optional) | After at least 180 days, configurable | Within 12 hours |
Objects smaller than 128 KB are not monitored and always stay in the Frequent Access tier. Accessed objects move back to Frequent Access automatically. Activate the optional archive tiers only if the application can wait for a restore.
S3 Lifecycle rules
Lifecycle rules suit data with a known, predictable pattern, such as logs that are hot for a week and kept for seven years. Rules transition objects to colder classes and expire them at the end of retention.
| Class | Retrieval | Minimum storage duration |
|---|---|---|
| S3 Standard-IA and S3 One Zone-IA | Milliseconds, per GB retrieval charge | 30 days |
| S3 Glacier Instant Retrieval | Milliseconds | 90 days |
| S3 Glacier Flexible Retrieval | Minutes to hours | 90 days |
| S3 Glacier Deep Archive | Within 12 hours by standard retrieval | 180 days |
Deleting or moving an object before its minimum duration still bills the remaining days. Lifecycle transitions also charge per request, so moving millions of tiny objects can cost more than it saves.
For versioned buckets, add rules for noncurrent versions and incomplete multipart uploads. Old versions and abandoned uploads are common hidden costs.
Choosing between them
Use Intelligent-Tiering when access is unknown or changes over time. Use lifecycle rules when the pattern is predictable, because they reach the coldest classes on a fixed schedule with no monitoring fee.
Amazon EFS lifecycle
EFS lifecycle management moves files that have not been accessed for a set period from Standard to Infrequent Access, and then to Archive. A policy can move a file back to Standard on first access, which suits files that become hot again. Files in the colder classes carry a per GB access charge, so tier only data that really cools.
EBS volume types and IOPS
| Volume | Performance model | Fits |
|---|---|---|
| gp3 | Baseline 3,000 IOPS and 125 MiB/s at any size, IOPS and throughput provisioned separately | Default for most workloads |
| gp2 | 3 IOPS per GiB, small volumes burst to 3,000 IOPS | Legacy, usually replace with gp3 |
| io2 Block Express | Provisioned IOPS with high durability | Critical databases needing sustained high IOPS |
| st1 | Throughput optimized HDD | Large sequential reads and writes |
| sc1 | Cold HDD | Infrequently accessed sequential data |
gp2 ties IOPS to size, so teams often overprovision capacity just to get IOPS. gp3 separates them, which usually makes it cheaper per GB and lets you pay only for the IOPS you need.
Elastic Volumes changes type, IOPS and size in place without detaching the volume, so gp2 to gp3 is a live change. Elastic Volumes cannot shrink a volume, so reducing size means creating a smaller volume and copying data.
Backup retention
AWS Backup applies backup plans across services and accounts. A plan sets frequency, retention, cross Region or cross account copies, and for supported resources a move to cold storage. Backups in cold storage must stay there at least 90 days.
Amazon Data Lifecycle Manager automates EBS snapshots and EBS backed AMIs from tags. Policies keep a set count or age of snapshots, can copy them to another Region, and can move snapshots to the EBS Snapshots Archive tier for long term retention.
| Need | Tool |
|---|---|
| Central policy across many services and accounts | AWS Backup with organization backup policies |
| Tag driven EBS snapshot and AMI schedules | Data Lifecycle Manager |
| Long retention of rarely restored data | Cold storage or the snapshot archive tier |
Match retention to the requirement, not to habit. Cross Region copies add transfer and storage costs, so copy only what the DR plan needs.
How the exam asks about this
- "Access patterns for a data lake are unknown and change often." -> S3 Intelligent-Tiering.
- "A database runs on large gp2 volumes only to reach its IOPS target." -> Move to gp3 with Elastic Volumes and provision the IOPS needed.
- "Compliance requires monthly snapshots kept for seven years and rarely restored." -> Data Lifecycle Manager or AWS Backup with archive or cold storage tiers.
What to remember for the exam
- Intelligent-Tiering has no retrieval fees but charges monitoring per object, and objects under 128 KB stay in Frequent Access.
- Lifecycle rules fit predictable patterns and can expire data at end of retention.
- Minimum storage durations are 30 days for IA classes, 90 for Glacier Instant and Flexible Retrieval, and 180 for Deep Archive.
- gp3 gives 3,000 IOPS and 125 MiB/s baseline at any size, independent of capacity.
- Elastic Volumes can change type and grow volumes but cannot shrink them.
- AWS Backup cold storage requires at least 90 days in cold storage.
Do not move frequently restored backups or hot objects to archive tiers. Retrieval time and minimum duration charges can break the RTO and erase the saving.
Common mistakes
Transitioning tiny objects. Per object transition requests and minimum size billing in IA classes make small objects expensive to move. Aggregate them or leave them in Standard.
Keeping gp2 because it works. gp3 usually meets the same performance at a lower cost, and the change is online with Elastic Volumes.
Setting no expiry on backups. Without a retention rule, snapshots and recovery points grow every month. Every plan needs an end date that matches compliance.
Check yourself
When is S3 Intelligent-Tiering better than a lifecycle rule?
When access patterns are unknown or change, because it moves objects by actual access without retrieval fees.
How do you cut the cost of a large gp2 volume that exists only to reach an IOPS target?
Change it to gp3 with Elastic Volumes and provision only the IOPS and throughput it needs.
Which tool automates tag based EBS snapshot schedules and retention?
Amazon Data Lifecycle Manager, which also supports cross Region copies and the snapshot archive tier.
Topic 6 of 8 in Cost-Optimized Architecture Design
Test yourself on Cost-Optimized Architecture Design
The full AWS SAP-C03 question bank has 850 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by AWS. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.