Amazon S3 Glacier Specialized2012年〜
An extremely low-cost storage service for archiving long-term data
What It Does
Amazon S3 Glacier is an archive storage service for storing data that needs long-term retention but is infrequently accessed, at extremely low cost. It safely stores data required by regulations, historical log files, backup data, and more. Retrieval takes minutes to hours, but storage costs are roughly one-tenth of S3 Standard.
Use Cases
It is used for regulatory compliance data retention, archiving medical records and financial transaction records, storing master copies of media files, long-term database backups, and disaster recovery data storage - managing any data that needs safe long-term preservation.
Everyday Analogy
Think of it like a bank safe deposit box. You store valuables (important data) you don't use daily in a bank safe deposit box (Glacier) for safekeeping. Retrieving them requires going to the bank and completing paperwork (retrieval takes time), but the storage fee is much cheaper than a home safe (S3 Standard).
What Is S3 Glacier?
Amazon S3 Glacier is an archive-specific storage service from AWS. It is integrated as one of S3's storage classes, and you can use S3 lifecycle policies to automatically transition data to Glacier. With 99.999999999% (eleven nines) durability and automatic replication across multiple facilities, the risk of data loss is extremely low even for long-term storage.
Storage Class Types
Glacier has three storage classes. S3 Glacier Instant Retrieval provides millisecond retrieval, suited for data accessed about once per quarter. S3 Glacier Flexible Retrieval provides retrieval in minutes to hours, suited for data accessed once or twice per year. S3 Glacier Deep Archive is the cheapest class, with retrieval taking 12+ hours but offering the lowest storage costs.
Lifecycle Management
Setting up S3 lifecycle policies lets you automatically transition storage classes based on data age. For example, you can set rules to move data to S3 Standard-IA after 30 days, Glacier Flexible Retrieval after 90 days, and Glacier Deep Archive after 365 days. This optimizes costs without manually moving data. For a deeper understanding of lifecycle management, related books (Amazon) are a useful resource.
Getting Started
The easiest way to start using Glacier is by setting a lifecycle policy on an S3 bucket. Select a bucket in the S3 console, go to the Management tab, and create a lifecycle rule. Specify the destination storage class and the number of days, and objects matching the conditions are automatically transitioned to Glacier. You can also upload data directly to Glacier.
Things to Watch Out For
- Data retrieval takes time, so use S3 Standard for data that needs immediate access
- Retrieval from Glacier incurs retrieval fees, so it is not suited for frequently accessed data
- Deleting data before the minimum storage duration (90 days for Flexible Retrieval, 180 days for Deep Archive) incurs early deletion charges
