Cloud Storage Options Comparison Cheat Sheet
Compares object, block, and file storage across AWS, Azure, and GCP, with guidance on when to use each storage class or type.
2 PagesBeginnerFeb 10, 2026
Storage Types
The three fundamental categories of cloud storage.
- Object Storage- Flat namespace of key/value blobs with HTTP access; e.g. S3, Azure Blob, GCS. Best for unstructured data
- Block Storage- Raw storage volumes attached to a VM; e.g. EBS, Azure Managed Disks. Best for databases, boot volumes
- File Storage- Shared filesystem accessed via NFS/SMB; e.g. EFS, Azure Files. Best for shared app data across instances
AWS S3 Storage Classes
Object storage tiers trading cost against access latency/frequency.
- S3 Standard- Frequently accessed data, millisecond access, highest cost per GB
- S3 Intelligent-Tiering- Automatically moves objects between tiers based on access patterns
- S3 Standard-IA- Infrequent access, lower storage cost but has a retrieval fee
- S3 One Zone-IA- Like Standard-IA but stored in a single AZ, cheaper, less durable
- S3 Glacier Instant Retrieval- Archive tier with millisecond retrieval, for rarely accessed data
- S3 Glacier Deep Archive- Lowest cost tier, retrieval takes hours, for long-term compliance archives
Common S3 CLI Commands
Uploading, syncing, and changing storage class via the AWS CLI.
bash
aws s3 cp file.txt s3://my-bucket/ # upload a fileaws s3 sync ./local-dir s3://my-bucket/dir/ # sync a directoryaws s3 ls s3://my-bucket/ # list objectsaws s3 cp s3://my-bucket/file.txt . \ --storage-class GLACIER # change storage class on copy
S3 Lifecycle Configuration (JSON)
Automate tiering and expiration rules instead of managing storage class transitions manually.
json
{ "Rules": [ { "ID": "tier-and-expire-logs", "Filter": { "Prefix": "logs/" }, "Status": "Enabled", "Transitions": [ { "Days": 30, "StorageClass": "STANDARD_IA" }, { "Days": 90, "StorageClass": "GLACIER" } ], "Expiration": { "Days": 365 }, "NoncurrentVersionTransitions": [ { "NoncurrentDays": 30, "StorageClass": "GLACIER" } ], "AbortIncompleteMultipartUpload": { "DaysAfterInitiation": 7 } } ]}
Multipart Upload for Large Objects
Objects over ~100MB should use multipart upload for resumability and parallel throughput; the CLI does this automatically above its threshold.
bash
# CLI auto-multiparts above the configured threshold (default 8MB)aws configure set default.s3.multipart_threshold 64MBaws configure set default.s3.multipart_chunksize 16MBaws s3 cp big-dataset.tar.gz s3://my-bucket/datasets/ \ --storage-class INTELLIGENT_TIERING# Low-level API flow (what the CLI does under the hood):# 1. create-multipart-upload -> returns UploadId# 2. upload-part (x N, each returns an ETag) # 3. complete-multipart-upload with the list of ETags# Abandoned uploads still incur storage cost until aborted or expired# by a lifecycle rule (AbortIncompleteMultipartUpload).
Provisioned IOPS Block Volumes
gp3 decouples IOPS/throughput from volume size, letting you tune performance independently of capacity.
bash
aws ec2 create-volume \ --volume-type gp3 \ --size 100 \ --iops 8000 \ --throughput 500 \ --availability-zone us-east-1a# io2 Block Express for sub-millisecond latency, sustained high IOPSaws ec2 create-volume \ --volume-type io2 \ --size 500 \ --iops 64000 \ --availability-zone us-east-1a# gp3 IOPS/throughput are billed separately from GB-month storage;# io2 durability is 99.999% vs gp3's 99.8-99.9%
EFS Throughput & Performance Modes
EFS throughput scales with stored data in Bursting mode, or can be provisioned independently for spiky/small datasets.
bash
# Bursting: throughput credits accrue proportional to storage size (default)aws efs create-file-system \ --performance-mode generalPurpose \ --throughput-mode bursting# Elastic: auto-scales with workload, best for unpredictable access patternsaws efs create-file-system \ --performance-mode generalPurpose \ --throughput-mode elastic# Provisioned: fixed MiB/s independent of storage size, for small-but-hot datasetsaws efs create-file-system \ --performance-mode maxIO \ --throughput-mode provisioned \ --provisioned-throughput-in-mibps 256
Cross-Cloud & Advanced Storage Terms
Concepts that matter for multi-cloud comparisons and production tuning beyond the basic type/class breakdown.
- Azure Access Tiers (Hot/Cool/Archive)- Blob Storage equivalent of S3 storage classes; Archive requires rehydration (hours) before read access, like Glacier
- GCS Storage Classes (Standard/Nearline/Coldline/Archive)- Google's tiering by access frequency; minimum storage duration penalties apply below Standard (30/90/365 days)
- S3 Object Lock (WORM)- Write-once-read-many compliance mode preventing deletion/overwrite for a retention period, used for regulatory holds
- Consistency Model- S3 is strongly read-after-write consistent for all operations since Dec 2020; older docs describing eventual consistency are outdated
- Cross-Region Replication (CRR)- Asynchronous, versioning-dependent replication of objects to a bucket in another region for DR or latency
- Presigned URLs- Time-limited, credential-scoped URLs granting temporary object access without making the bucket public
- Requester Pays- Bucket setting that shifts data transfer/request costs to the requester instead of the bucket owner, common for public datasets
- Storage Gateway- Hybrid appliance (file/volume/tape modes) that caches on-prem access to cloud object storage
Pro Tip
Use S3 Lifecycle rules to automatically transition objects to cheaper tiers over time (e.g. Standard -> IA after 30 days -> Glacier after 90) instead of manually managing storage classes.
Was this cheat sheet helpful?
Previous
Cloud Networking (VPC) Basics Cheat Sheet
Next
Disaster Recovery & Backup Strategies Cheat Sheet
Explore Topics
#CloudStorageOptionsComparison#CloudStorageOptionsComparisonCheatSheet#CloudComputing#Beginner#StorageTypes#AWSS3StorageClasses#CommonS3CLICommands#S3LifecycleConfigurationJSON#OOP#CheatSheet#SkillVeris
