S3

S3 is a simple, infinitly scaling object storage service. Objects are divided into Buckets which must have a globally unique name, but they are a regional service. All objects have a key which is full path within the bucket, so we can have a mocked directory structure within a bucket. Objects can also have metadata, tags, versions, and encryption. There is also a setting to host a static website via S3. You can replicate S3 buckets across regions if versioning is enabled.

Storage Classes

• Standard - Frequent access
• Infrequent Access (IA) - Accessed less frequently, but requires rapid access
• One Zone-Infrequent Access - Limited to one AZ
• Glacier Instant Retrieval - MS retrieval, must store 90 days min
• Glacier Flexible Retrieval - Expedited (1 to 5 minutes), Standard (3 to 5 hours), Bulk (5 to 12 hours) – free, musy store for 90 days min
• Glacier Deep Archive - Standard (12 hours), Bulk (48 hours), must store for 180 dayts min
• Intelligent Tiering - Auto tiers storage for a monthly fee

Snow Family

Highly secure, portable devices to collect and process data at the edge and then migrate to the cloud. You can use AWS OpsHub to manage these devices. You pay for device usage and data out of the cloud, data in is free.

• Snowball Edge - Data transport for moving TBs or PBs of data. Can be optimized for storage or compute.
• Snowcone - Small portable computing, rugged and secure for harsh envs. 8TB for HDD and 14TB for SSD.
• Snowmobile - Literal truck. Can move 1EB (1000 PBs).

AWS Storage Gateway

A way to bridge on-prem data and cloud data in S3 and with other storage types (EBS).