Introduction
AWS Glacier provides cost-effective archive storage for data you access once per year or less. This guide shows you how to migrate existing archives, configure vault policies, and optimize retrieval costs. You learn to balance preservation needs against budget constraints while maintaining compliance requirements.
Key Takeaways
- Glacier charges based on storage volume, requests, and data retrieval; pricing varies by retrieval type
- Vault lock policies enforce compliance retention rules automatically
- Multi-factor delete protection prevents accidental data loss
- Glacier Deep Archive offers the lowest cost tier at $0.00099 per GB monthly
- Data retrieval times range from milliseconds to 12 hours depending on tier
What is AWS Glacier
AWS Glacier is an Amazon Web Services storage class designed for rarely accessed data. According to AWS documentation, the service stores data for as little as $0.004 per GB monthly in the standard tier. Organizations use Glacier to replace expensive on-premises tape systems and reduce long-term data retention costs. The service automatically encrypts data at rest using AES-256 and distributes replicas across multiple facilities.
Why AWS Glacier Matters
Regulatory requirements force companies to retain financial records, medical images, and legal documents for years. According to SEC regulations, broker-dealers must preserve electronic communications for three years. Tape infrastructure costs $50-$200 per GB when including hardware, management, and facility expenses. Glacier reduces archive storage costs by 90% compared to traditional methods while providing better durability of 99.999999999%.
How AWS Glacier Works
Glacier organizes data in vaults that hold archives. You upload data in chunks up to 40TB per archive. The service uses a three-part retrieval model based on urgency and cost:
Retrieval Tiers and Formulas
Cost Formula: Total Cost = (Storage GB × Price) + (Uploads × Upload Cost) + (Retrieval GB × Retrieval Cost)
Instant Retrieval: $0.004/GB/month, retrieval in milliseconds, best for frequently accessed archives
Flexible Retrieval: $0.0036/GB/month, retrieval in 1-5 minutes (expedited), 3-5 hours (standard), 5-12 hours (bulk)
Deep Archive: $0.00099/GB/month, retrieval in 12-48 hours, designed for compliance archives
Vaults support configurable access policies and lifecycle rules that automatically move objects from S3 Standard to Glacier after specified days.
Used in Practice
A healthcare system implements Glacier Deep Archive for 10-year radiology retention requirements. The architecture uses S3 Lifecycle policies to transition DICOM files to Glacier after 90 days in Standard storage. Vault lock policies enforce the 10-year retention period, preventing deletion even by administrators. Retrieval costs average $0.02 per study, acceptable given the $0.001 monthly storage savings per GB.
Financial services firms use Glacier for transaction logs required by Dodd-Frank compliance. They configure cross-region replication to a secondary vault for disaster recovery. The automated vault inventory updates weekly, providing a catalog of all archived objects with checksums for integrity verification.
Risks and Limitations
Glacier retrieval costs surprise many users. A 100GB retrieval at standard speed costs approximately $0.09, but retrieval requests themselves incur charges even if the data fails integrity checks. You cannot attach Lambda triggers directly to Glacier events; notifications flow through SQS or SNS after retrieval completion.
Data availability presents another consideration. Expedited retrievals cost $0.03 per GB, ten times the standard rate. If your application requires sub-minute access, Glacier Instant Retrieval or S3 Standard better fit your needs. Archive counts per vault are unlimited, but vault access operations cost $0.05 per vault list request.
AWS Glacier vs S3 Standard vs S3 Intelligent-Tiering
S3 Standard charges $0.023 per GB monthly for frequently accessed data but offers immediate retrieval. Glacier Flexible Retrieval costs 84% less per month but requires hours for bulk retrievals. Intelligent-Tiering automatically moves objects between frequent and infrequent access tiers, charging $0.0125 monthly for infrequent access but adding monitoring fees of $0.0025 per 1,000 objects.
For active archives accessed monthly, Intelligent-Tiering makes sense. For compliance archives touched once yearly, Glacier Deep Archive delivers the lowest total cost. You sacrifice retrieval speed for storage savings—calculate your access patterns before choosing.
What to Watch
Monitor your retrieval patterns before archiving. Historical access logs in CloudWatch reveal whether “cold” data truly stays cold. Unexpected retrieval spikes indicate data that should remain in Standard storage. Set CloudWatch alarms for retrieval thresholds to catch runaway processes early.
Verify vault lock policies before applying them. Locked policies become immutable after the 24-hour validation period. Test retention policies on a dummy vault first. Consider using Glacier’s updateable vault lock for policies that require future modifications.
Frequently Asked Questions
How long does AWS Glacier retrieval take?
Glacier Flexible Retrieval completes standard requests within 3-5 hours and bulk requests within 5-12 hours. Expedited retrievals finish in 1-5 minutes but cost $0.03 per GB. Deep Archive retrieval takes 12-48 hours depending on data volume.
What happens if I delete data before the retention period ends?
Vault lock policies enforce retention periods. You cannot delete objects until the retention date passes. Attempted deletions before the retention date fail and generate error logs. Some compliance frameworks allow early deletion with special permissions.
Can I access Glacier data directly like regular files?
No. Glacier archives require initiation of a retrieval job first. Once the job completes, the data remains accessible for 24 hours via a temporary URL. Applications must handle asynchronous retrieval workflows rather than direct file access.
Is Glacier data encrypted?
Yes. AWS encrypts all data at rest using AES-256 encryption. You can manage your own encryption keys through AWS Key Management Service or use customer-provided keys for additional control.
How does Glacier pricing compare to S3 Standard over five years?
S3 Standard costs approximately $1.38 per GB annually, totaling $6.90 over five years. Glacier Deep Archive costs approximately $0.12 per GB annually, totaling $0.60 over five years. Glacier saves $6.30 per GB over five years but charges retrieval fees when you need access.
Can I restore Glacier data to S3 for editing?
Yes. Lifecycle policies can transition archived objects back to S3 Standard after retrieval. This enables temporary access for modifications before re-archiving. The round-trip involves standard retrieval fees plus standard storage rates for the time data spends in S3.
Leave a Reply