AWS Storage Solutions - SAA-C03¶
Amazon S3 (Simple Storage Service)¶
π Amazon S3 Documentation - Complete guide to S3 storage and features
Storage Classes¶
π S3 Storage Classes - Choose the right storage class for your data access patterns - S3 Standard: Frequently accessed data, low latency, high throughput - S3 Intelligent-Tiering: Automatic cost optimization, moves data between tiers - S3 Standard-IA: Infrequent access, lower cost, retrieval fee - S3 One Zone-IA: Single AZ, 99.5% availability, lower cost - S3 Glacier Instant Retrieval: Archive with millisecond retrieval - S3 Glacier Flexible Retrieval: Archive with 1-5 minute retrieval - S3 Glacier Deep Archive: Lowest cost, 12-hour retrieval
Key Features¶
- Versioning: Keep multiple versions of objects
- Lifecycle Policies: Automate transitions between storage classes
- Cross-Region Replication (CRR): Replicate objects across regions
- Same-Region Replication (SRR): Compliance and data sovereignty
- S3 Transfer Acceleration: Fast uploads using CloudFront edge locations
- Multipart Upload: For objects >100MB, required for >5GB
- S3 Select: Query data with SQL without retrieving entire object
π S3 Replication - Automatically replicate objects across AWS Regions
Security¶
- Bucket Policies: Resource-based policies (JSON)
- Access Control Lists (ACLs): Legacy access control
- Block Public Access: Account and bucket level protection
- Encryption at Rest:
- SSE-S3: S3-managed keys
- SSE-KMS: KMS-managed keys, audit trail
- SSE-C: Customer-provided keys
- Encryption in Transit: HTTPS/TLS
π S3 Security Best Practices - Implement security controls for S3 buckets and objects
Performance Optimization¶
- S3 Prefixes: 3,500 PUT/COPY/POST/DELETE, 5,500 GET/HEAD per prefix
- Multipart Upload: Parallel uploads for large files
- Transfer Acceleration: Use CloudFront edge locations
- Byte-Range Fetches: Download partial object
Amazon EBS (Elastic Block Store)¶
π Amazon EBS Documentation - Block storage for EC2 instances
Volume Types¶
π EBS Volume Types - Choose the right EBS volume for your workload SSD-Based (IOPS-focused): - gp3 (General Purpose SSD): 3,000-16,000 IOPS, 125-1,000 MB/s throughput - gp2 (General Purpose SSD): Baseline 3 IOPS/GB, burst to 3,000 IOPS - io2 Block Express: Up to 256,000 IOPS, 4,000 MB/s, 99.999% durability - io2: Up to 64,000 IOPS, 1,000 MB/s, 99.999% durability - io1: Up to 64,000 IOPS, 1,000 MB/s
HDD-Based (Throughput-focused): - st1 (Throughput Optimized): 500 IOPS, 500 MB/s, big data, log processing - sc1 (Cold HDD): 250 IOPS, 250 MB/s, infrequent access
Key Features¶
- Snapshots: Point-in-time backups, incremental, stored in S3
- Snapshot Archive: 75% cost reduction, 24-72 hour retrieval
- Fast Snapshot Restore (FSR): Instant restore, expensive
- Multi-Attach (io1/io2): Attach volume to multiple EC2 instances
- Encryption: At rest and in transit, uses KMS
- EBS-Optimized Instances: Dedicated bandwidth for EBS
π EBS Snapshots - Back up your EBS volumes with snapshots
Best Practices¶
- Use gp3 for most workloads (cheaper than gp2)
- io2 for databases requiring high IOPS
- Take regular snapshots for backup
- Delete old snapshots to save costs
- Enable encryption by default
Amazon EFS (Elastic File System)¶
π Amazon EFS Documentation - Scalable, elastic file storage for Linux workloads
Features¶
- Network File System (NFS): POSIX-compliant, shared storage
- Multi-AZ: Automatically replicated across AZs
- Auto-scaling: Grows and shrinks automatically
- Performance Modes:
- General Purpose: Low latency, web serving
- Max I/O: Higher latency, big data, media processing
Throughput Modes¶
- Bursting: Throughput scales with size (50 MB/s per TB)
- Provisioned: Set throughput independent of size
- Elastic: Automatically scales throughput (up to 3 GB/s reads, 1 GB/s writes)
Storage Classes¶
- Standard: Multi-AZ, frequently accessed
- Infrequent Access (IA): Lower cost, lifecycle policy
- One Zone: Single AZ, 47% cost savings
- One Zone-IA: Lowest cost
Use Cases¶
- Content management systems
- Web serving
- Home directories
- Shared development environments
- Container storage (ECS/EKS)
Amazon FSx¶
π Amazon FSx Documentation - Fully managed file systems built on popular file systems
FSx for Windows File Server¶
- SMB Protocol: Windows native file system
- Active Directory Integration: Windows authentication
- Performance: Up to 2 GB/s throughput, millions of IOPS
- Use Cases: Windows applications, home directories, SQL Server
π FSx for Windows File Server - Fully managed Windows file storage
FSx for Lustre¶
- High Performance: Machine learning, HPC, video processing
- Throughput: 100s GB/s, millions of IOPS
- S3 Integration: Direct access to S3 objects
- Deployment Types:
- Scratch: Temporary, 200 MB/s per TiB
- Persistent: Long-term, 50/100/200 MB/s per TiB
π FSx for Lustre - High-performance file system for compute-intensive workloads
FSx for NetApp ONTAP¶
- NFS and SMB: Multi-protocol support
- Point-in-time Snapshots: Instant recovery
- Data Compression: Storage efficiency
FSx for OpenZFS¶
- High Performance: Up to 1 million IOPS
- Point-in-time Snapshots: Instant clones
- Use Cases: Linux workloads, databases
AWS Storage Gateway¶
π AWS Storage Gateway - Hybrid cloud storage integration
Gateway Types¶
File Gateway: - Store files as S3 objects - NFS and SMB protocols - Local cache for frequently accessed data - Use Case: Backup, archive, cloud-native app data
Volume Gateway: - Block storage using iSCSI - Cached Volumes: Primary data in S3, cache on-premises - Stored Volumes: Primary data on-premises, async backup to S3 - Use Case: Backup, disaster recovery
Tape Gateway: - Virtual tape library (VTL) - Backup software integration - Store tapes in S3 Glacier - Use Case: Replace physical tape infrastructure
Best Practices¶
- Use File Gateway for NFS/SMB file shares backed by S3
- Use Volume Gateway for block storage backup
- Tape Gateway for existing backup workflows
- Deploy gateway on-premises or in EC2
AWS DataSync¶
π AWS DataSync - Automated data transfer between on-premises and AWS
Features¶
- Fast Transfer: 10x faster than open-source tools
- Bandwidth Throttling: Control network usage
- Schedule: Hourly, daily, weekly transfers
- Data Validation: Verify integrity
- Filtering: Include/exclude patterns
Supported Locations¶
- Source: On-premises (NFS, SMB), AWS (S3, EFS, FSx)
- Destination: S3, EFS, FSx
- Use Cases: Migration, replication, archival
DataSync vs Storage Gateway¶
- DataSync: One-time or scheduled migrations
- Storage Gateway: Continuous hybrid storage access
AWS Snow Family¶
Snowcone¶
- Capacity: 8 TB HDD or 14 TB SSD
- Use Case: Edge computing, small data transfers
- Portable: 4.5 lbs, rugged
- DataSync Agent: Preinstalled
Snowball Edge¶
- Storage Optimized: 80 TB HDD
- Compute Optimized: 42 TB HDD + 7.68 TB SSD
- Compute with GPU: 42 TB + GPU
- Use Case: Large migrations (TB-PB), edge processing
Snowmobile¶
- Capacity: 100 PB per truck
- Use Case: Exabyte-scale datacenter migrations
- Security: GPS tracking, 24/7 monitoring
When to Use Snow¶
- Limited Bandwidth: >1 week to transfer over network
- High Transfer Costs: Network egress charges
- Disconnected Environments: Edge locations
- Rule of Thumb: >10 TB = consider Snow
Storage Comparison¶
Block vs File vs Object¶
Block Storage (EBS): - Low latency, high IOPS - Single instance attachment (except Multi-Attach) - Boot volumes, databases - Snapshots for backup
File Storage (EFS, FSx): - Shared access from multiple instances - POSIX permissions (EFS) - SMB/NFS protocols - Auto-scaling capacity
Object Storage (S3): - Unlimited storage - HTTP/HTTPS access - 99.999999999% durability - Static websites, backups, data lakes
Cost Optimization Strategies¶
- Use S3 Intelligent-Tiering for unknown access patterns
- Implement S3 Lifecycle Policies to move to cheaper tiers
- Delete incomplete multipart uploads
- Use EFS IA for infrequently accessed files
- Delete old EBS snapshots
- Use gp3 instead of gp2 (20% cheaper)
- Right-size EBS volumes
Exam Tips¶
Common Scenarios¶
- Shared file storage across EC2: EFS
- Windows file shares: FSx for Windows
- High-performance computing: FSx for Lustre
- Boot volumes: EBS (gp3 or gp2)
- Database storage: EBS (io2 for production)
- Static website: S3
- Long-term archive: S3 Glacier Deep Archive
- Hybrid cloud storage: Storage Gateway
- Large data migration: AWS DataSync or Snow Family
- Infrequent access: S3 IA, EFS IA, gp2 with burst credits
Key Differences¶
- EBS vs EFS: Single instance vs multi-instance shared
- S3 vs EBS: Object storage vs block storage
- DataSync vs Storage Gateway: Migration vs ongoing access
- Snowball vs DataSync: Offline vs online transfer