AWS EBS Volume Types: gp3, io2, and st1 Explained

AWS EBS volume types are categorized into SSD-backed (gp3, io2) for transactional workloads and HDD-backed (st1, sc1) for throughput-intensive tasks. Choose gp3 for balanced performance, io2 for critical low-latency databases, and st1 for large-scale data streaming, ensuring you balance IOPS and throughput against your budget.

Why does choosing the right EBS volume type matter for SAA-C03?

If you're prepping for the AWS Solutions Architect Associate (SAA-C03), you've likely noticed that storage is a recurring theme. Choosing the wrong EBS volume isn't just a technical mistake; in the real world, it's a financial disaster or a performance bottleneck. The exam will test your ability to map a specific business requirement—like 'low latency' or 'cost-effective big data storage'—to the correct volume type.

To nail these questions, you need to understand the trade-off between IOPS (Input/Output Operations Per Second) and Throughput (MB/s). IOPS is all about how many small read/write operations can happen per second, while throughput is about how much data is moved in a single operation. Getting this distinction right is the difference between a passing score and a retake. We see many students struggle here because they treat all 'SSD' volumes as the same, but the architectural implications of gp3 versus io2 are massive.

When should you choose gp3 over gp2?

For years, gp2 was the gold standard, but gp3 is now the recommended default for most workloads. The critical difference lies in how performance is scaled. In gp2, your IOPS are tied to the size of the volume (3 IOPS per GB). This meant if you needed more performance, you were forced to pay for more storage than you actually needed—a classic AWS 'tax.'

Gp3 changes the game by decoupling performance from storage capacity. With gp3, you get a baseline of 3,000 IOPS and 125 MiB/s throughput regardless of the volume size. If your application needs more, you can provision additional IOPS and throughput independently. From a cost perspective, gp3 is typically 20% cheaper per GB than gp2. When you see a scenario on the SAA-C03 asking for a cost-effective, general-purpose SSD for a boot volume or a small database, gp3 is almost always your best bet.

Is Provisioned IOPS (io2) worth the premium cost?

Provisioned IOPS SSDs (io1 and io2) are the 'heavy lifters' of the EBS world. You use these when your application cannot tolerate any latency spikes. While gp3 is great for most, io2 is designed for mission-critical, I/O-intensive workloads like large SQL Server or Oracle databases and NoSQL clusters.

Io2 offers significantly higher durability (99.9%) and can scale up to 256,000 IOPS per volume. On the exam, look for keywords like 'sub-millisecond latency,' 'extremely high I/O requirements,' or 'critical production database.' If the scenario describes a workload that would choke a general-purpose volume, io2 is the answer. Just remember: with great power comes a great price tag. You're paying for a guaranteed level of performance that doesn't rely on 'burst credits,' making it the most predictable—and most expensive—option.

When do HDD volumes like st1 and sc1 make sense?

Not everything needs an SSD. When you're dealing with massive amounts of data that are accessed sequentially rather than randomly, Hard Disk Drives (HDD) are the way to go. Throughput Optimized HDD (st1) is built for frequently accessed, throughput-intensive workloads. Think of it as the ideal choice for Big Data, Log processing, and Data Warehousing. It's designed to move large blocks of data quickly.

Then there's Cold HDD (sc1), which is the cheapest EBS volume type available. Use this for infrequently accessed data where cost is the primary driver. A critical 'gotcha' for the SAA-C03 exam: HDD volumes (st1 and sc1) cannot be used as boot volumes. If a question asks you to launch an EC2 instance with an st1 root volume, that's a trick—it's impossible. You must use an SSD-backed volume for the root and attach HDDs as secondary data volumes.

How do IOPS and Throughput impact database performance?

This is where many candidates get tripped up. To simplify: think of IOPS as the number of trips a delivery driver makes, and Throughput as the size of the truck. If you have a transactional database (OLTP) where thousands of users are making tiny updates to their profiles, you need high IOPS. The driver needs to make many quick trips.

Conversely, if you're running a data warehouse (OLAP) that scans millions of rows to generate a monthly report, you need high Throughput. You don't care about the number of trips; you want the biggest truck possible to move the most data per second. When analyzing SAA-C03 scenarios, identify if the workload is 'random' (IOPS) or 'sequential' (Throughput). This distinction will lead you directly to the correct volume type, whether it's the balanced gp3 or the high-throughput st1.

Can you modify EBS volumes without downtime?

One of the most powerful features of EBS is 'Elastic Volumes.' In the past, changing a volume size or type required creating a new snapshot and restoring it to a new volume—a tedious process. Now, you can modify your volume type (e.g., moving from gp3 to io2) or increase the volume size on-the-fly while the volume is still attached to a running EC2 instance.

However, there are a few rules you must remember for the exam. First, you can increase the size of a volume, but you cannot decrease it. To shrink a volume, you'd have to create a smaller volume and migrate the data manually. Second, there is a cooldown period—typically 6 hours—between modifications. This means you can't just spam changes to your volume. Understanding this flexibility is key to designing resilient architectures that can scale as application demand grows without requiring maintenance windows.

How can you master EBS for the SAA-C03 exam?

Reading the documentation is a start, but the SAA-C03 is a scenario-based exam. You won't be asked 'What is gp3?'; you'll be asked 'Which storage solution provides the best cost-to-performance ratio for a 500GB database with 10,000 IOPS?' That's where practice becomes the differentiator. You need to train your brain to recognize the patterns in these questions.

At Cert Sensei, we provide 1,000 expert-curated AWS Solutions Architect Associate practice questions specifically designed to mimic the actual exam. Instead of just giving you a correct letter, we provide detailed expert reasoning for every answer, explaining why the other options were wrong. Our performance analytics with domain-level tracking allow you to see exactly where you're weak—whether it's EBS volume types or VPC peering—so you can stop wasting time on what you already know and focus on the gaps in your knowledge.

❓ Frequently Asked Questions