SAA-C03
Deep Dive
The SAA-C03 exam tests whether you know how to pick the right storage type for the use case: block (EBS), shared file (EFS), object (S3), or managed high-performance file systems (FSx).
Contents
EBS (Elastic Block Store) volumes are block storage attached to EC2 instances. They persist independently of the instance lifecycle. They attach to a single instance at a time(except EBS Multi-Attach for io1/io2).
| Type | Technology | IOPS | Throughput | Use |
|---|---|---|---|---|
| gp3 (SSD) | NVMe SSD | 3,000-16,000 IOPS | 125-1,000 MB/s | Default for most workloads. OS, apps, dev/test. Low cost. |
| gp2 (SSD) | SSD | 100-16,000 IOPS | 128-250 MB/s | Legacy. Use gp3 instead (same price, more IOPS). |
| io2 Block Express (SSD) | NVMe SSD | Up to 256,000 IOPS | Up to 4,000 MB/s | Mission-critical DBs needing maximum performance. Multi-Attach. |
| io1 (SSD) | NVMe SSD | Up to 64,000 IOPS | Up to 1,000 MB/s | I/O-intensive DBs: Oracle, SAP HANA, MySQL. |
| st1 (HDD) | HDD | Up to 500 IOPS | Up to 500 MB/s | Big data, Kafka, logs, frequent sequential access. |
| sc1 (HDD) | HDD | Up to 250 IOPS | Up to 250 MB/s | Cold data, backups, infrequent access. Cheapest option. |
gp3 vs gp2 — always go gp3
gp3 has the same base price as gp2 but delivers a guaranteed 3,000 IOPS and 125 MB/s regardless of size. gp2 scales IOPS with size (3 IOPS/GB). For small volumes, gp3 is significantly better.
EBS Multi-Attach (io1/io2)
Lets you attach the same volume to multiple EC2 instances in the same AZ. Requires a cluster-aware file system (not standard ext4). For high-availability applications that need concurrent access (Oracle RAC, Teradata).
S3 scales automatically. Each prefix in S3 can handle at least 5,500 GET/HEAD and 3,500 PUT/COPY/POST/DELETE requests per second. Spreading objects across multiple prefixes multiplies throughput.
S3 Transfer Acceleration
Faster uploads using CloudFront's global network. The client uploads to the nearest Edge Location; AWS routes internally to the S3 region.
Multipart Upload
Splits large files into parts and uploads them in parallel. Recommended for files >100 MB, required for >5 GB.
S3 Byte-Range Fetches
Downloads specific parts of a file in parallel. Improves download speed for large files.
S3 Select and Glacier Select
Runs SQL queries directly against CSV/JSON/Parquet objects in S3 without downloading the whole file.
EFS (Elastic File System) is a fully managed NFS file system. Unlike EBS (one volume per instance), EFS can be mounted simultaneously on thousands of EC2 instancesacross multiple AZs.
Features
EFS signals in the exam
EFS vs EBS — the right choice
EBS = one instance (block). EFS = multiple simultaneous instances (file). If the scenario mentions "share data between instances" → EFS.
Amazon FSx for Windows File Server
A fully managed Windows file system. SMB, NTFS, Active Directory compatibility. For Windows apps that need shared file storage.
Exam
"Windows workloads", "SQL Server on EC2 needs shared storage", "Active Directory integration"
Amazon FSx for Lustre
A high-performance file system for HPC (High Performance Computing), ML and analytics. Native S3 integration — can automatically import data from S3.
Exam
"HPC", "high-speed ML training", "scientific simulations", "sub-ms access to S3 data" → FSx for Lustre
| Service | Type | Access | Best for |
|---|---|---|---|
| EBS gp3 | SSD block | One EC2 instance | OS, applications, general-purpose databases |
| EBS io2 | NVMe block | One instance (Multi-Attach optional) | Mission-critical high-performance DBs (Oracle, SAP) |
| EBS st1 | HDD block | One EC2 instance | Big data, Kafka, sequential access |
| EFS | File (NFS) | Thousands of simultaneous Linux instances | Shared file systems, CMS, home directories |
| FSx for Windows | File (SMB) | Windows instances | Windows apps, AD integration, SQL Server |
| FSx for Lustre | HPC file | HPC/ML clusters | ML training, scientific simulations |
| S3 | Object | Internet / AWS | Images, videos, backups, data lake, logs |
| S3 Glacier | Object (archival) | S3 API (minutes/hours latency) | Long-term retention, compliance |
Did you get this topic?
Test what you just learned
A company has a Linux web server with an Auto Scaling Group of up to 10 instances. They need to store user image files that all web servers can read and write simultaneously. Which storage service is most appropriate?
Put what you learned to the test with original SAA-C03 exam questions.
Start SAA-C03 mock examUnlimited mock exams with the Pro plan
Log in to track your progress.