AWS Solutions Architect Exam > AWS Solutions Architect Notes > : Professional Level > Cheat Sheet: Storage
Cheat Sheet: Storage
1. Amazon S3
1.1 Storage Classes
| Storage Class | Key Characteristics |
|---|---|
| S3 Standard | 11 9's durability, 4 9's availability, ≥3 AZs, millisecond latency, no retrieval fees |
| S3 Intelligent-Tiering | Auto moves objects between tiers, no retrieval fees, monitoring fee $0.0025 per 1000 objects |
| S3 Standard-IA | Lower storage cost, retrieval fee, 3 9's availability, ≥3 AZs, 30-day min storage |
| S3 One Zone-IA | Single AZ, 2.5 9's availability, 20% lower cost than Standard-IA, 30-day min storage |
| S3 Glacier Instant Retrieval | Millisecond retrieval, 90-day min storage, lowest cost for instant access archive |
| S3 Glacier Flexible Retrieval | Retrieval: Expedited (1-5 min), Standard (3-5 hrs), Bulk (5-12 hrs), 90-day min storage |
| S3 Glacier Deep Archive | Lowest cost, Standard (12 hrs), Bulk (48 hrs) retrieval, 180-day min storage |
1.2 Encryption
| Type | Details |
|---|---|
| SSE-S3 | S3-managed keys, AES-256, header: x-amz-server-side-encryption: AES256 |
| SSE-KMS | KMS-managed keys, audit trail via CloudTrail, header: x-amz-server-side-encryption: aws:kms |
| SSE-C | Customer-provided keys, HTTPS required, key provided in each request header |
| Client-Side | Encryption done before upload, customer manages entire encryption process |
| DSSE-KMS | Dual-layer encryption, KMS with two layers of encryption, compliance requirements |
1.3 Replication
| Feature | Details |
|---|---|
| Cross-Region Replication (CRR) | Replicate across regions, compliance, lower latency, disaster recovery |
| Same-Region Replication (SRR) | Log aggregation, live replication between prod/test, data sovereignty |
| Requirements | Versioning enabled on both buckets, IAM permissions, asynchronous replication |
| What Replicates | New objects after replication enabled, encrypted objects (SSE-S3, SSE-KMS with permissions), delete markers (optional) |
| What Doesn't Replicate | Objects before replication enabled, SSE-C encrypted, objects in Glacier/Deep Archive, deletions with version ID |
| Replication Time Control | 15-minute SLA for 99.99% of objects, metrics and notifications included |
| Batch Replication | Replicate existing objects, objects failed initial replication, on-demand replication |
1.4 Lifecycle Policies
- Transition actions: Move objects between storage classes based on age
- Expiration actions: Delete objects after specified time period
- Minimum transition wait: 30 days from Standard to Standard-IA or One Zone-IA
- Can transition from Glacier Instant to Glacier Flexible to Glacier Deep Archive
- Cannot transition from any Glacier tier back to Standard/IA tiers
- Applies to versioned objects: current and non-current versions managed separately
- Incomplete multipart uploads can be deleted after specified days
1.5 Performance Optimization
| Feature | Details |
|---|---|
| Request Rate | 3,500 PUT/COPY/POST/DELETE, 5,500 GET/HEAD requests per second per prefix |
| Multipart Upload | Recommended for files >100 MB, required for files >5 GB, parallel uploads, max 10,000 parts |
| S3 Transfer Acceleration | Uses CloudFront edge locations, 50-500% faster long-distance transfers, separate endpoint |
| Byte-Range Fetches | Parallel downloads, retrieve partial object, failure resilience |
| S3 Select | SQL queries to retrieve subset of data, reduces transfer and processing, works on CSV/JSON/Parquet |
1.6 Access Control
| Method | Use Case |
|---|---|
| Bucket Policies | JSON-based, cross-account access, public access, enforce encryption, conditions based on IP/VPC/tags |
| IAM Policies | User/role-based permissions, centralized management, works across all buckets |
| ACLs | Legacy, object-level access, grant basic permissions (READ, WRITE, FULL_CONTROL) |
| Pre-signed URLs | Temporary access, time-limited, inherits creator's permissions, SDK/CLI generated |
| Access Points | Dedicated access points per application, simplified permission management, VPC-restricted access |
| Block Public Access | Account or bucket level, overrides other policies, four settings for blocking public access |
1.7 Event Notifications
- Triggers: PUT, POST, COPY, DELETE, restore from Glacier, Replication Time Control
- Destinations: SNS, SQS, Lambda, EventBridge
- EventBridge enables: advanced filtering, multiple destinations, 18+ services, archive/replay
- Standard notifications limited to 3 destination types, simple prefix/suffix filtering
- Versioned buckets: notifications triggered for each version ID
1.8 S3 Object Lock
| Feature | Details |
|---|---|
| Compliance Mode | Cannot be overridden by anyone including root, retention period cannot be shortened |
| Governance Mode | Special permissions can override, s3:BypassGovernanceRetention permission required |
| Retention Period | Fixed time period (days/years), protects from deletion and modification |
| Legal Hold | Indefinite protection, no expiration, s3:PutObjectLegalHold permission, independent of retention |
| Requirements | Versioning must be enabled, applied at object version level, cannot be disabled once enabled |
1.9 S3 Additional Features
| Feature | Details |
|---|---|
| S3 Inventory | Scheduled reports (daily/weekly), CSV/ORC/Parquet format, metadata about objects, encryption status audit |
| S3 Batch Operations | Bulk operations on billions of objects, invoke Lambda, restore from Glacier, ACL/tag modifications |
| S3 Analytics | Storage class analysis, recommendations for lifecycle policies, CSV export to S3, updated daily |
| S3 Storage Lens | Organization-wide visibility, 29+ metrics, free and advanced tiers, recommendations dashboard |
| Requester Pays | Requester pays transfer costs, bucket owner pays storage, requester must be authenticated |
2. Amazon EBS
2.1 Volume Types
| Type | Specifications |
|---|---|
| gp3 (General Purpose SSD) | 3,000-16,000 IOPS, 125-1,000 MB/s throughput, 1 GB-16 TB, cost-effective baseline performance |
| gp2 (General Purpose SSD) | 3 IOPS/GB baseline (min 100, max 16,000), burst to 3,000 IOPS, 1 GB-16 TB, throughput 128-250 MB/s |
| io2 Block Express | Up to 256,000 IOPS, 4,000 MB/s throughput, 4 GB-64 TB, 1000:1 IOPS:GB ratio, sub-millisecond latency |
| io2 (Provisioned IOPS SSD) | Up to 64,000 IOPS, 1,000 MB/s throughput, 4 GB-16 TB, 500:1 IOPS:GB ratio, 99.999% durability |
| io1 (Provisioned IOPS SSD) | Up to 64,000 IOPS, 1,000 MB/s throughput, 4 GB-16 TB, 50:1 IOPS:GB ratio, 99.9% durability |
| st1 (Throughput Optimized HDD) | 500 IOPS, 500 MB/s throughput, 125 GB-16 TB, big data/log processing, cannot be boot volume |
| sc1 (Cold HDD) | 250 IOPS, 250 MB/s throughput, 125 GB-16 TB, lowest cost, infrequent access, cannot be boot volume |
2.2 Multi-Attach
- Supported only on io1 and io2 volumes (not io2 Block Express)
- Attach single volume to up to 16 Nitro-based EC2 instances
- All instances must be in same AZ
- Each instance has full read/write permissions
- Requires cluster-aware file system (not XFS, EXT4)
- Use case: clustered applications requiring concurrent write access
2.3 Snapshots
| Feature | Details |
|---|---|
| Snapshot Storage | Stored in S3 (not visible), incremental, first snapshot is full copy, regional resource |
| Snapshot Copy | Copy across regions, copy and encrypt, disaster recovery, migration |
| Fast Snapshot Restore | Instant initialization, no latency on first use, charged per DSU-hour per AZ, max 50 FSR snapshots |
| Snapshot Archive | 75% cost reduction, 24-72 hours restore time, minimum 90-day archive period |
| Recycle Bin | Protect against accidental deletion, 1 day to 1 year retention, recover deleted snapshots |
| EBS Direct APIs | Read snapshots directly, create snapshots incrementally, no EC2 instance needed |
2.4 Encryption
- Uses KMS keys (AES-256), encrypted at rest and in transit between instance and volume
- Snapshots of encrypted volumes are encrypted, volumes from encrypted snapshots are encrypted
- Minimal impact on latency, handled transparently
- Encrypt unencrypted volume: create snapshot, copy snapshot with encryption, create volume from encrypted snapshot
- Default encryption: all new volumes/snapshots encrypted with default KMS key
- Encryption cannot be removed once enabled on volume
2.5 Volume Modifications
- Modify size, volume type, IOPS, throughput without detaching
- Size increases only (cannot decrease), minimum 6 hours between modifications
- After modification, volume enters "optimizing" state, can take hours to days
- File system extension required after size increase (OS-level operation)
- Monitor progress using CloudWatch VolumeModificationProgress metric
2.6 Performance and Monitoring
| Metric | Details |
|---|---|
| IOPS | Input/Output operations per second, smaller block sizes = more IOPS consumed |
| Throughput | MB/s, larger block sizes = more throughput consumed, independent configuration in gp3 |
| Queue Depth | Number of pending I/O requests, higher queue depth = better performance for high IOPS volumes |
| CloudWatch Metrics | VolumeReadBytes, VolumeWriteBytes, VolumeReadOps, VolumeWriteOps, BurstBalance (gp2/st1/sc1) |
| I/O Credits (gp2) | 5.4 million I/O credits at creation, replenish at baseline rate, depleted during bursts |
3. Amazon EFS
3.1 Fundamentals
- Managed NFS (Network File System v4.1), multi-AZ, elastic capacity
- Mount on multiple EC2 instances across AZs, Linux-only (POSIX-compliant)
- Uses security groups for access control, encryption at rest using KMS
- Pay-per-use pricing model, scales automatically to petabytes
- Performance: 10+ GB/s throughput, millions of IOPS, sub-millisecond latencies
3.2 Performance Modes
| Mode | Details |
|---|---|
| General Purpose | Latency-sensitive workloads (web servers, CMS), lowest latency per operation, default mode |
| Max I/O | Big data, media processing, higher aggregate throughput and IOPS, higher latency per operation, cannot change after creation |
3.3 Throughput Modes
| Mode | Details |
|---|---|
| Bursting | Scales with file system size, 50 MB/s per TB, bursts to 100 MB/s, uses credit system |
| Provisioned | Set throughput independent of storage size, pay for provisioned throughput, can exceed bursting limits |
| Elastic | Auto-scales throughput up/down based on workload, up to 3 GB/s reads, 1 GB/s writes, pay for actual throughput used |
3.4 Storage Classes
| Class | Details |
|---|---|
| Standard | Frequently accessed files, multi-AZ, highest availability and durability |
| Infrequent Access (IA) | Lower storage cost, retrieval fee, lifecycle policy auto-moves files, up to 92% cost savings |
| One Zone | Single AZ, 47% cost savings over Standard, less availability |
| One Zone-IA | Single AZ + IA pricing, 95% cost savings over Standard, development/backup use cases |
3.5 EFS Features
| Feature | Details |
|---|---|
| Lifecycle Management | Auto-move files to IA after N days (7, 14, 30, 60, 90 days), based on last access time |
| Access Points | Application-specific entry points, enforce user identity, root directory per application, simplified access management |
| Mount Targets | One per AZ, multiple EC2 instances share single mount target, requires ENI in VPC subnet |
| Encryption | At rest (KMS), in transit (TLS), must enable at creation for at-rest encryption |
| Backup | Integrates with AWS Backup, point-in-time recovery, incremental backups, cross-region backup |
3.6 EFS vs EBS
| Feature | EFS |
|---|---|
| Attachment | Multi-attach across hundreds of instances and AZs, shared file system |
| Availability | Multi-AZ by default, regional service, higher availability |
| Cost | 3x more expensive than gp2, pay-per-use, no capacity planning |
| Use Cases | Content management, web serving, data sharing, WordPress, shared application state |
| Protocol | NFSv4.1, POSIX-compliant, Linux-only |
4. Amazon FSx
4.1 FSx for Windows File Server
| Feature | Details |
|---|---|
| Protocol | SMB protocol, Windows-native shared file system, NTFS file system |
| Integration | Active Directory integration, user quotas, access control lists (ACLs) |
| Performance | SSD: 100s MB/s, millions of IOPS, sub-millisecond latency; HDD: 10s MB/s, 100,000s IOPS |
| Storage | SSD (hot data), HDD (warm data), 32 GB to 64 TB capacity |
| Deployment | Single-AZ or Multi-AZ, automatic failover, daily backups to S3 |
| Access | VPC, VPN, Direct Connect, on-premises via DFS Namespaces |
| Deduplication | Data deduplication enabled, saves 50-60% storage costs |
4.2 FSx for Lustre
| Feature | Details |
|---|---|
| Use Cases | HPC, machine learning, video processing, financial modeling, genomics |
| Performance | 100s GB/s throughput, millions of IOPS, sub-millisecond latency |
| S3 Integration | Read/write to S3 as file system, output results back to S3 |
| Deployment Types | Scratch (temporary, no replication, 200 MB/s per TB); Persistent (long-term, auto-replicated in AZ, 50/100/200 MB/s per TB) |
| Metadata | Lazy loading from S3, metadata stored on Lustre, instant file visibility |
| File System | POSIX-compliant, Linux-only, parallel distributed file system |
4.3 FSx for NetApp ONTAP
- Managed NetApp ONTAP on AWS, NFS, SMB, iSCSI protocols
- Move workloads from on-premises NetApp to AWS
- Works with Linux, Windows, MacOS, VMware Cloud on AWS
- Storage auto-scales, point-in-time clones, compression, deduplication
- Snapshots, replication, low-cost storage tiers
- Multi-AZ deployment for high availability
4.4 FSx for OpenZFS
- Managed OpenZFS file system on AWS, NFS protocol (v3, v4, v4.1, v4.2)
- Move workloads from on-premises ZFS to AWS
- Up to 1,000,000 IOPS, sub-millisecond latency
- Snapshots, compression, point-in-time cloning, low-cost storage
- Works with Linux, Windows, MacOS
- Single-AZ deployment
4.5 FSx Comparison
| FSx Type | Best For |
|---|---|
| Windows File Server | Windows applications, Active Directory, SMB, user quotas, NTFS |
| Lustre | HPC, ML, compute-intensive workloads, S3 integration, highest performance |
| NetApp ONTAP | NetApp migration, multi-protocol (NFS/SMB/iSCSI), enterprise features |
| OpenZFS | ZFS migration, NFS protocol, point-in-time cloning, Linux workloads |
5. AWS Storage Gateway
5.1 Gateway Types
| Type | Details |
|---|---|
| File Gateway (S3 File) | NFS/SMB protocol, files stored as objects in S3, local cache for low-latency, integrates with S3 lifecycle/versioning/encryption |
| FSx File Gateway | Native access to FSx for Windows File Server, local cache, Windows-native SMB, Active Directory integration |
| Volume Gateway (Cached) | iSCSI protocol, primary data in S3, frequently accessed data cached locally, low-latency access to hot data |
| Volume Gateway (Stored) | Entire dataset on-premises, asynchronous backup to S3 as EBS snapshots, low-latency access to entire dataset |
| Tape Gateway (VTL) | Virtual Tape Library, iSCSI protocol, backup to S3/Glacier using existing tape software, supports Veeam, Veritas, CommVault |
5.2 File Gateway Details
- Each file gateway can support up to 10 file shares
- Supports S3 Standard, S3 Standard-IA, S3 One Zone-IA, S3 Intelligent-Tiering
- Bucket access via IAM role per file gateway
- Integrates with Active Directory for user authentication (SMB only)
- File metadata stored as S3 object metadata
- Automated cache refresh from S3 based on lifecycle policies
5.3 Volume Gateway Details
| Feature | Cached Volumes |
|---|---|
| Storage | Primary data in S3, cache is 1 TB to 32 TB per volume, up to 32 volumes (1 PB total) |
| Snapshots | Point-in-time EBS snapshots stored in S3, incremental backups, can create EBS volumes from snapshots |
| Use Case | Primary data in AWS, on-premises applications need low-latency access to hot data |
5.4 Tape Gateway Details
- Virtual tapes: 100 GB to 5 TB each, up to 1,500 virtual tapes (1 PB total)
- Virtual Tape Library backed by S3, Virtual Tape Shelf backed by Glacier/Deep Archive
- Appears as iSCSI-based tape system to backup applications
- Archive tapes moved to Glacier: retrieval 3-5 hours (Flexible) or 12 hours (Deep Archive)
- Eliminates physical tape management and offsite storage costs
5.5 Deployment Options
| Deployment | Details |
|---|---|
| On-Premises VM | VMware ESXi, Microsoft Hyper-V, Linux KVM, hardware appliance option available |
| Hardware Appliance | Physical device from AWS, pre-configured, 1U/2U rackmount, for locations without virtualization |
| EC2 Instance | Deploy gateway as EC2 instance, disaster recovery, cloud-to-cloud migration |
6. AWS Transfer Family
6.1 Protocols
| Protocol | Details |
|---|---|
| SFTP (SSH File Transfer) | Port 22, secure file transfer over SSH, public key authentication |
| FTPS (FTP with SSL) | Port 21 (control), port 989/990 (data), FTP over TLS/SSL |
| FTP | Port 21, unencrypted, VPC only (not public internet), legacy application support |
| AS2 (Applicability Statement 2) | B2B data exchange, EDI messages, HTTP/HTTPS-based, message encryption and signing |
6.2 Storage Backends
- Amazon S3: store files as objects, integrate with S3 features (versioning, lifecycle, encryption)
- Amazon EFS: Linux file systems, NFS-based, shared file storage across multiple endpoints
- Each user can have home directory mapped to specific S3 bucket/prefix or EFS path
6.3 Authentication
| Method | Details |
|---|---|
| Service Managed | Store user credentials within Transfer Family, SSH keys stored in service |
| Custom Identity Provider | Integrate with Lambda, authenticate against external systems (LDAP, Active Directory, custom database) |
| AWS Directory Service | Microsoft AD integration, use existing corporate credentials |
6.4 Endpoint Types
| Type | Details |
|---|---|
| Public | Internet-accessible, AWS-provided hostname, cannot whitelist IP addresses |
| VPC (Internal) | Private IP in VPC, access via VPN/Direct Connect, internal applications |
| VPC (Internet-Facing) | Elastic IP attached, fixed IP addresses for whitelisting, internet-accessible via VPC |
6.5 Key Features
- Fully managed, pay per-protocol-enabled hour plus data transfer
- Multi-AZ deployment, auto-scaling, no infrastructure management
- Integrates with CloudWatch for monitoring, CloudTrail for auditing
- Supports managed workflows for post-upload processing (Lambda, ECS, Step Functions)
- Custom hostname using Route 53, certificate via ACM
- Up to 10,000 users per server, 3 protocols enabled per endpoint
7. AWS DataSync
7.1 Fundamentals
- Automated data transfer service for moving large amounts of data
- Agent-based (on-premises) or agentless (AWS-to-AWS)
- Replication can be scheduled hourly, daily, weekly
- Preserves file permissions, metadata, timestamps
- One agent task can use 10 Gbps, saturate network links
- Incremental transfers after initial full copy
7.2 Source and Destination Locations
| Location Type | Details |
|---|---|
| Sources | On-premises (NFS, SMB), AWS Snowcone, Amazon S3, Amazon EFS, Amazon FSx (Windows, Lustre, NetApp ONTAP, OpenZFS) |
| Destinations | Amazon S3 (all classes including Glacier), Amazon EFS, Amazon FSx (all types) |
| Cross-Region | Transfer between AWS regions, cross-account transfer supported |
7.3 DataSync Agent
- Required for on-premises to AWS transfers, deployed as VM (VMware, Hyper-V, KVM)
- Connects to AWS via internet, Direct Connect, or VPN using TLS encryption
- VPC endpoint (PrivateLink) for private connectivity to DataSync service
- Agent activates using activation key from AWS Management Console
- Multiple agents can be deployed for parallel transfers
7.4 Performance and Features
| Feature | Details |
|---|---|
| Bandwidth Throttling | Limit bandwidth used by DataSync, prevent network saturation |
| Data Validation | Verify data integrity using checksums, compare source and destination |
| Task Execution | Manual, scheduled, or event-driven, integrates with EventBridge |
| Filtering | Include/exclude patterns, transfer specific files or folders |
| Logging | CloudWatch Logs for transfer details, CloudWatch Metrics for monitoring |
7.5 Use Cases
- Data migration: on-premises to AWS, minimize downtime during migration
- Data replication: keep on-premises and AWS data in sync
- Data archival: move cold data to Glacier storage classes
- Data processing: transfer data to AWS for processing, return results
- Disaster recovery: replicate critical data to AWS continuously
8. AWS Snow Family
8.1 Device Comparison
| Device | Storage |
|---|---|
| Snowcone | 8 TB HDD or 14 TB SSD, 2 CPUs, 4 GB RAM, USB-C power or battery, weighs 4.5 lbs |
| Snowball Edge Storage Optimized | 80 TB usable (100 TB raw), 40 vCPUs, 80 GB RAM, 1 TB SSD for block volumes, 10 GbE/25 GbE networking |
| Snowball Edge Compute Optimized | 28 TB NVMe usable (42 TB raw), 104 vCPUs, 416 GB RAM, optional GPU, 100 GbE networking |
| Snowmobile | 100 PB per truck, exabyte-scale transfer, GPS tracking, 24/7 security, climate controlled, requires multiple-week on-site deployment |
8.2 DataSync on Snowcone
- DataSync agent pre-installed on Snowcone devices
- Transfer data to Snowcone using DataSync, then ship device to AWS
- Snowcone connects to on-premises systems via wired or wireless network
- Use when network connectivity is limited or expensive
8.3 Edge Computing Capabilities
| Feature | Details |
|---|---|
| EC2 Instances | Run EC2 instances on Snowball Edge, AMI-based deployment, process data at edge |
| Lambda Functions | Run Lambda at edge, process data locally before transfer to AWS |
| Use Cases | Pre-processing data, machine learning at edge, remote/disconnected locations (ships, mining, military) |
| Clustering | 5-10 Snowball Edge devices can cluster together, increased durability and storage |
8.4 OpsHub
- Graphical application to manage Snow Family devices
- Unlock devices, transfer files, launch EC2 instances, monitor metrics
- Works with Snowcone, Snowball Edge (both types)
- Installed on local computer (Windows, Mac, Linux)
- Replaces AWS CLI for Snow device management
8.5 Workflow and Pricing
| Stage | Details |
|---|---|
| Order | Request device via AWS Console, shipped in days |
| Load Data | Connect device, install OpsHub, copy data, encryption enabled by default |
| Ship Back | E-ink shipping label auto-updates, track via SNS/text/console |
| Import | AWS loads data to S3, device wiped using NIST standards |
| Pricing | On-demand (pay per use) or 1/3-year commitment, data transfer into S3 free, standard S3 pricing applies |
9. Storage Selection and Hybrid Architecture
9.1 Block Storage Selection
| Requirement | Solution |
|---|---|
| Single-instance boot volume | EBS gp3 (cost-effective) or io2 (performance) |
| Shared access across instances | EBS Multi-Attach (io1/io2) or EFS |
| High IOPS database (>64K IOPS) | EBS io2 Block Express (up to 256K IOPS) or instance store |
| Temporary high-performance storage | Instance store (ephemeral, highest IOPS) |
| Big data sequential I/O | EBS st1 (throughput optimized HDD) |
| Infrequent access, lowest cost | EBS sc1 (cold HDD) |
9.2 File Storage Selection
| Requirement | Solution |
|---|---|
| Linux shared file system | EFS (NFSv4.1, POSIX-compliant) |
| Windows shared file system | FSx for Windows File Server (SMB, Active Directory) |
| High-performance computing | FSx for Lustre (sub-ms latency, millions IOPS) |
| NetApp migration | FSx for NetApp ONTAP (multi-protocol) |
| ZFS migration | FSx for OpenZFS (NFS, snapshots, cloning) |
| On-premises NFS/SMB to S3 | File Gateway (S3 backend with local cache) |
9.3 Object Storage Selection
| Requirement | Solution |
|---|---|
| Frequent access | S3 Standard |
| Unknown access patterns | S3 Intelligent-Tiering |
| Infrequent access, millisecond retrieval | S3 Standard-IA or S3 One Zone-IA |
| Archive, instant retrieval | S3 Glacier Instant Retrieval |
| Archive, retrieval in hours | S3 Glacier Flexible Retrieval |
| Long-term archive, lowest cost | S3 Glacier Deep Archive |
9.4 Hybrid Storage Patterns
| Pattern | Solution |
|---|---|
| On-premises file shares to S3 | File Gateway (NFS/SMB to S3) |
| On-premises block volumes to S3 | Volume Gateway cached/stored volumes |
| On-premises tape backup to S3/Glacier | Tape Gateway (VTL to S3/Glacier) |
| Large-scale data migration | DataSync (scheduled sync) or Snow Family (offline transfer) |
| SFTP/FTPS to S3 | Transfer Family (managed FTP servers) |
| Low-latency access to FSx for Windows | FSx File Gateway |
9.5 Data Transfer Decision Tree
| Scenario | Recommended Solution |
|---|---|
| Good network, ongoing sync | DataSync (scheduled replication) |
| Limited network, one-time transfer <10> | Snowcone |
| Limited network, one-time transfer 10-80 TB | Snowball Edge |
| Limited network, one-time transfer >80 TB | Multiple Snowball Edge or Snowmobile (>10 PB) |
| SFTP/FTP access to S3 | Transfer Family |
| Continuous hybrid access | Storage Gateway (File/Volume/Tape) |
9.6 Multi-Region Strategies
- S3 Cross-Region Replication: async replication, compliance, disaster recovery, 15-min SLA with RTC
- EBS snapshots: copy snapshots across regions, can encrypt during copy
- EFS: no native replication, use DataSync or Backup for cross-region
- FSx: no native replication, use Backup or third-party tools
- DataSync: schedule cross-region transfers for any supported storage
The document Cheat Sheet: Storage is a part of the AWS Solutions Architect Course AWS Solutions Architect: Professional Level.
All you need of AWS Solutions Architect at this link: AWS Solutions Architect
About this Document
Apr 21, 2026 Last updated
Related Exams
Document Description: Cheat Sheet: Storage for AWS Solutions Architect 2026 is part of AWS Solutions Architect: Professional Level preparation. The notes and questions for Cheat Sheet: Storage have been prepared according to the AWS Solutions Architect exam syllabus. Information about Cheat Sheet: Storage covers topics like and Cheat Sheet: Storage Example, for AWS Solutions Architect 2026 Exam. Find important definitions, questions, notes, meanings, examples, exercises and tests below for Cheat Sheet: Storage.
Introduction of Cheat Sheet: Storage in English is available as part of our AWS Solutions Architect: Professional Level for AWS Solutions Architect & Cheat Sheet: Storage in Hindi for AWS Solutions Architect: Professional Level course. Download more important topics related with notes, lectures and mock test series for AWS Solutions Architect Exam by signing up for free. AWS Solutions Architect: Cheat Sheet: Storage
Description
Cheat Sheet: Storage of AWS Solutions Architect to help you remember important concepts with short tricks. Start learning for AWS Solutions Architect exam & improve retention with EduRev.
Information about Cheat Sheet: Storage
In this doc you can find the meaning of Cheat Sheet: Storage defined & explained in the simplest way possible. Besides explaining types of Cheat Sheet: Storage theory, EduRev gives you an ample number of questions to practice Cheat Sheet: Storage tests, examples and also practice AWS Solutions Architect tests
Related Searches
Summary, Semester Notes, ppt, Free, Cheat Sheet: Storage, Cheat Sheet: Storage, Extra Questions, Important questions, Viva Questions, video lectures, practice quizzes, shortcuts and tricks, Cheat Sheet: Storage, pdf , Exam, Sample Paper, Objective type Questions, MCQs, study material, mock tests for examination, past year papers, Previous Year Questions with Solutions;