What Does EBS volume types Mean?
On This Page
Quick Definition
Amazon EBS volume types are different kinds of virtual hard drives you can attach to an EC2 instance in the AWS cloud. Each type is designed for a specific workload, like general-purpose applications, high-performance databases, or cost-effective data storage. Choosing the right type affects speed, cost, and reliability.
Commonly Confused With
Instance store volumes are temporary block storage physically attached to the host server of an EC2 instance. They offer very high performance but data is lost if the instance stops or terminates. EBS volumes persist independently and can be detached and reattached. Instance stores are often confused with EBS because both appear as local disks, but their durability and lifecycle are completely different.
Use instance store for cache or temporary data that can be regenerated. Use EBS for databases or file systems that need to survive instance termination.
EBS snapshots are point-in-time backups of EBS volumes stored in Amazon S3. Snapshots are not volume types; they are a feature that can be applied to any EBS volume type. Learners sometimes confuse snapshot performance with volume performance. For example, restoring a snapshot onto a gp2 volume may take longer than on an io1 volume because the volume type's performance during restoration depends on the volume type.
If you need to restore a large database snapshot quickly, choose io1 or io2 for the target volume for faster performance. For archival, restore as gp3 or even sc1 to save cost.
S3 is object storage, not block storage. EBS volumes appear as a block device attached to an EC2 instance, ideal for operating systems and databases. S3 is accessed over HTTP APIs and is better for storing large files, backups, and data lakes. The storage classes (S3 Standard, S3 Glacier, etc.) are analogous to EBS volume types but serve different architectural roles.
Use S3 for hosting static assets like images and videos for a website. Use EBS for the database that runs the website.
Some EC2 instance types (like i3 or i3en) offer physical NVMe SSDs directly attached to the server. These are instance store volumes, not EBS. They provide extremely high IOPS and low latency but are ephemeral. Learners might confuse them with EBS io2 volumes, which are durable and can be detached from the instance.
For high-performance computing requiring scratch space, use NVMe instance store. For persistent high-performance storage, use io2 EBS.
Must Know for Exams
EBS volume types are a core topic in AWS certification exams, including AWS Certified Solutions Architect (Associate and Professional), AWS Certified SysOps Administrator, and AWS Certified Developer. For the Solutions Architect Associate (SAA-C03) exam, understanding volume types is in Domain 2: Design for Performance and Cost Optimization. You will be asked to choose the most appropriate volume type based on a given scenario. For example, a NoSQL database that requires 50,000 IOPS with low latency would need io2 Block Express, not gp3.
In the SysOps Administrator exam (SOA-C02), you may see questions about modifying volume types, monitoring performance, and troubleshooting bottlenecks. Expect scenario-based questions where you must interpret CloudWatch metrics to identify an underperforming volume. For example, if a volume shows high QueueDepth and low IOPS, you might need to increase IOPS or switch to a higher-performing type.
The Developer exam (DVA-C02) focuses on integration with SDKs and APIs for volume lifecycle management. Questions around snapshot creation and volume type changes during backup operations are common. You might need to identify which volume type supports multi-attach (io1 and io2) or which can be used as a boot volume (SSD types only).
For the Professional Architect exam (SAP-C02), you will deal with complex, multi-volume architectures involving RAID configurations, EBS Multi-Attach, and performance optimization across large fleets. Questions might compare io2 Block Express vs. local NVMe instance storage for high-performance compute.
Across all exams, the key is to match workload characteristics (IOPS, throughput, latency, persistence, cost) with the correct volume type. Memorize the performance limits: gp3 max 16,000 IOPS and 1,000 MB/s; io2 Block Express up to 256,000 IOPS and 4,000 MB/s; st1 max 500 MB/s throughput; sc1 max 250 MB/s. Also remember that HDD volumes are not bootable and use burst credits. Trap questions often involve boot volumes or mixed workload scenarios.
Simple Meaning
Think of Amazon EBS volume types like choosing the right kind of storage for your home. If you need a place to store everyday items like books and clothes, a standard closet (gp3 volume) works perfectly. It is balanced, affordable, and fast enough for most tasks.
But if you need a high-security vault for your most valuable documents, you would pick a specialized safe (io2 Block Express). It is much faster, more durable, and can handle a huge number of operations per second, but it costs more. For things like old photo albums that you rarely look at, you might use a cheap plastic bin in the basement (sc1).
It holds a lot but is slow to access. EBS offers several such volume types: gp3 and gp2 for general use, io2 and io1 for high-performance databases, st1 for frequently accessed big data, and sc1 for infrequently accessed archives. Each type has different limits on how fast data can be read or written (IOPS) and how much data can flow per second (throughput).
You can also change the type later if your needs change, just like swapping storage bins around your house. Understanding these types helps you build efficient and cost-effective cloud applications, because picking the wrong one can either slow down your app or waste money on unused performance.
Full Technical Definition
Amazon Elastic Block Store (EBS) provides durable, block-level storage volumes for use with Amazon EC2 instances. EBS volume types are distinct categories defined by their performance characteristics, including IOPS (Input/Output Operations Per Second), throughput in MB/s, and latency. They fall into two main families: SSD-backed volumes optimized for transactional workloads and HDD-backed volumes optimized for streaming workloads. Each type is built on different underlying storage hardware and protocols, with specific APIs and durability guarantees.
The SSD family includes: gp2 and gp3 (General Purpose SSD) offer balanced price-performance for most workloads. gp3 allows independent scaling of IOPS and throughput up to 16,000 IOPS and 1,000 MB/s per volume. gp2 uses a credit-based burst model where IOPS are tied to volume size (baseline 3 IOPS per GB). io1 and io2 (Provisioned IOPS SSD) are designed for latency-sensitive enterprise workloads like databases. io2 provides 99.999% durability and up to 256,000 IOPS with io2 Block Express, which uses a scale-out architecture. io1 offers up to 64,000 IOPS and is being phased out in favor of io2.
The HDD family includes: st1 (Throughput Optimized HDD) is for frequently accessed, sequential workloads like big data and log processing, delivering up to 500 MB/s throughput per volume. sc1 (Cold HDD) is for less frequently accessed data, offering lower cost and up to 250 MB/s throughput. Both use burst credits and are limited to a maximum IOPS of 500 (st1) or 250 (sc1). HDD volumes cannot be used as boot volumes.
All EBS volumes are replicated within an Availability Zone for durability. They attach over the AWS Nitro System or Xen hypervisor via NVMe or SCSI protocols. Performance is measured per volume, with additional performance available through EBS-optimized instances and Elastic Fabric Adapter. Volume modifications can be done without downtime, including resizing, type changes, and performance adjustments. Monitoring via CloudWatch metrics is essential for right-sizing and cost optimization.
Real-Life Example
Imagine you are organizing a large community center. You need different storage spaces for different activities. First, there is the main storage closet (gp3) where you keep general supplies like chairs, tables, and basic equipment. It is accessible, fast enough for daily use, and you can adjust how quickly people can grab items (IOPS) and how much can be taken out at once (throughput). If you need a special high-security room for a high-value art exhibition (io2), you install a biometric lock and reinforced walls. This room has extremely fast access for security guards (high IOPS) and can handle many visitors per hour (high throughput). It costs more to build and maintain, but it is necessary for valuable items.
Then, consider the loading dock area where donations arrive on pallets (st1). You need a wide door and a conveyor belt to move boxes quickly in sequence. It can handle large volumes of big boxes but each box takes time to process (high throughput, low IOPS). For the rarely used storage room in the basement (sc1), you keep old event banners and supplies from years ago. It is cheap to rent, but if you need something, it takes a while to go down and find it. You would not use it for things you need every day.
Mapping this to IT: the general closet is your gp3 volume for web servers or dev environments. The high-security room is io2 for a critical database. The loading dock is st1 for log analytics or data warehouses. The basement is sc1 for backups or archives. Choosing the right storage type for each part of your application ensures you have good performance without overspending.
Why This Term Matters
In practical IT, selecting the correct EBS volume type directly impacts application performance, cost, and operational complexity. For example, a production database like MySQL or Oracle requires consistent, high IOPS for transaction processing. Using a gp3 volume might cause latency spikes during peak loads, while an io2 volume guarantees predictable performance. For a web server serving static content, gp3 is more cost-effective and offers sufficient performance.
Misconfiguration can lead to serious issues. Choosing a low-cost sc1 volume for a critical application can cause timeouts and poor user experience. Conversely, using io2 for a development environment that runs only a few hours a day wastes money. Many organizations use lifecycle policies to automatically transition volumes to different types as data ages, reducing storage costs by up to 60%.
Monitoring is key. EBS volumes report metrics like VolumeReadBytes, VolumeWriteBytes, and VolumeIdleTime. These help you identify if a volume is under-provisioned or over-provisioned. For instance, if your average IOPS is 1,000 but you provisioned io2 at 10,000 IOPS, you are paying for unused capacity. You can downsize or change volume type to save money.
EBS snapshots are tied to volume types. Snapshot restore performance is faster for io2 volumes compared to gp2. Understanding these nuances helps system architects design resilient, performant, and cost-optimized infrastructure. Cloud migration projects often fail because teams do not map on-premises storage requirements to the correct EBS volume type, resulting in performance degradation or budget overruns.
How It Appears in Exam Questions
Exam questions on EBS volume types appear in three main patterns: scenario-based selection, performance tuning, and cost optimization.
Scenario-based selection: You are given a description of an application workload and must choose the most appropriate EBS volume type. For example, A company runs a MySQL database with high transaction rates requiring consistent 40,000 IOPS. Which volume type should you choose? The correct answer is io2 or io2 Block Express. A distractor might be gp3, but gp3 supports only up to 16,000 IOPS on a single volume.
Performance tuning: Questions present a CloudWatch graph showing high VolumeQueueLength, low Average IOPS, and high latency. You are asked what action to take. Options might include increasing volume size, switching to a higher IOPS volume type, or enabling EBS optimization on the instance. The best answer is typically to change the volume type to a Provisioned IOPS SSD (io2) or increase the IOPS setting on a gp3 volume.
Cost optimization: A scenario describes a data warehouse that is used for batch reporting every night, requiring high throughput for sequential reads. The current gp2 volumes are expensive and underutilized. What is the most cost-effective volume type? The correct answer is Throughput Optimized HDD (st1), which provides high throughput at lower cost for sequential workloads. A common distractor is Cold HDD (sc1), but sc1 is for infrequent access, not nightly reporting.
Another pattern involves volume modifications. A developer needs to increase the IOPS of a running volume without downtime. Which volume types support this? All current generation EBS volumes support online modification. However, questions might test if users know that HDD volumes have fixed performance limits and cannot be scaled as high as SSD.
Finally, there are questions about multi-attach. Only io1 and io2 support attaching a single volume to multiple EC2 instances simultaneously. A scenario might describe a clustered database requiring shared storage, and the answer would be io1 or io2 with multi-attach enabled.
Practise EBS volume types Questions
Test your understanding with exam-style practice questions.
Example Scenario
A small startup is migrating its e-commerce platform to AWS. The platform has a customer-facing web application, a product catalog stored in a PostgreSQL database, and a logging system that collects clickstream data for analytics. The startup wants to optimize cost while ensuring good performance.
For the web application servers, they choose gp3 volumes because the workload is balanced, and the application can tolerate occasional latency spikes. They attach one gp3 volume per EC2 instance, sized at 100 GB with baseline IOPS sufficient for the expected traffic peak of 3,000 IOPS.
For the PostgreSQL database, they use an io2 volume provisioned with 10,000 IOPS and 99.999% durability because the database requires consistent low latency for transactions. They enable EBS-optimization on the EC2 instance and monitor VolumeLatency in CloudWatch. During a flash sale, the database IOPS spikes to 9,500, but the volume handles it without degradation.
For the clickstream logging system, they use an st1 volume attached to a dedicated instance running Apache Kafka. The volume is 500 GB with high throughput (400 MB/s) because the log data is written sequentially and read later for batch processing. This saves costs compared to using SSD volumes.
After six months, the logging data older than 30 days is rarely accessed. The team uses a lifecycle policy to create snapshots of the st1 volume and restore them as sc1 volumes for archival. This reduces storage costs by 70%. They also right-size the gp3 volumes for the web tier, reducing unused IOPS. This scenario demonstrates how choosing the right EBS volume type for each workload component leads to a balanced, cost-effective architecture.
Common Mistakes
Using an HDD volume as a boot volume for an EC2 instance.
HDD volumes (st1 and sc1) are not supported as boot volumes because boot volumes must maintain a file system and handle many small random I/O operations, which HDDs cannot do efficiently.
Always use an SSD-backed volume (gp2, gp3, io1, or io2) for the root device of any EC2 instance.
Choosing gp2 instead of gp3 assuming it is cheaper.
gp3 often provides better performance at a lower cost because it decouples IOPS and throughput from volume size. gp3 baseline IOPS is 3,000 vs gp2's 100 IOPS baseline per GB, and gp3 pricing per GB is lower.
Always evaluate gp3 first for general-purpose workloads unless there is a specific reason to use gp2 (e.g., legacy compatibility).
Over-provisioning IOPS on a gp3 volume for a sequential workload.
Sequential workloads benefit from throughput, not IOPS. A high IOPS setting on a gp3 volume does not improve sequential performance and incurs extra cost.
Match the performance characteristic to the workload: use throughput-optimized settings for sequential access, and IOPS for random access. For heavy sequential, consider st1 instead.
Assuming all EBS volumes support multi-attach.
Only io1 and io2 volumes support multi-attach (attaching to multiple EC2 instances simultaneously). gp2, gp3, st1, and sc1 do not support this feature.
If you need a shared block storage volume across instances for a clustered database, use io1 or io2 with multi-attach enabled, or consider Amazon EFS or FSx for Lustre for shared file storage.
Failing to monitor EBS burst credits on gp2 and st1/sc1 volumes.
Burst credits allow temporary high performance, but once exhausted, performance drops to baseline. Many users experience sudden slowdowns during peak hours without understanding the credit model.
Monitor BurstBalance metric in CloudWatch. If credits are depleted frequently, consider switching to gp3 or io2 for consistent performance, or increase volume size to raise baseline performance.
Exam Trap — Don't Get Fooled
{"trap":"You are asked to recommend the most cost-effective storage for a database requiring 10,000 IOPS. Options include gp3 (max 16,000 IOPS), st1, sc1, and io1. Learners often choose gp3 because it is cheaper than io1, but they overlook that the database also needs low latency and consistent performance, which gp3 can provide, so actually gp3 is correct.
However, the trap is that st1 and sc1 are not SSD and cannot deliver 10,000 IOPS. The distractor is choosing st1 thinking it can handle any throughput. Or choosing io1 unnecessarily.
In this scenario, gp3 is the best cost-effective choice.","why_learners_choose_it":"Learners may pick st1 because they see 'throughput' and assume it works for databases, or they pick io1 because they think only Provisioned IOPS can reach 10,000 IOPS, forgetting that gp3 also supports up to 16,000 IOPS at lower cost.","how_to_avoid_it":"Remember that gp3 supports up to 16,000 IOPS and is cost-effective.
Use st1 only for sequential workloads, and only use io1/io2 when you need more than 16,000 IOPS or require specific features like multi-attach or maximum durability. Always check the latest limits because AWS updates them."
Step-by-Step Breakdown
Identify workload requirements
Determine the application's primary access pattern: is it random reads/writes (transactional) or sequential reads/writes (streaming)? Estimate required IOPS and throughput. Also check latency sensitivity and whether data must persist after instance termination. This step sets the direction for volume type selection.
Choose between SSD and HDD families
SSD volumes (gp2, gp3, io1, io2) are for transactional workloads with small I/O operations and high IOPS requirements. HDD volumes (st1, sc1) are for throughput-intensive sequential workloads. If the workload is a boot volume, you must use SSD. This step narrows the options significantly.
Select the specific volume type within the family
For SSD: gp3 is the default for general-purpose needs. Use io2 if you need more than 16,000 IOPS, consistent latency under load, multi-attach, or the highest durability (99.999%). For HDD: use st1 for frequently accessed data like logs or data warehouses, and sc1 for infrequently accessed archives. This step refines the choice based on performance and cost.
Configure volume size and performance parameters
For gp3, set baseline IOPS (default 3,000) and baseline throughput (125 MB/s) and scale independently. For io2, provision the exact IOPS needed (up to 256,000 with Block Express). For st1 and sc1, size affects baseline throughput. Ensure the volume size meets the kernel and application needs. Over-provisioning wastes money; under-provisioning causes performance issues.
Attach volume to EC2 instance and format
Attach the EBS volume to an instance using the AWS Console, CLI, or API. The volume appears as a device (e.g., /dev/sdf). Format it with a file system (ext4, xfs, etc.) and mount it. Ensure the instance is EBS-optimized or uses a Nitro instance to avoid network bottlenecks. This step translates the storage choice into a usable disk.
Monitor and optimize volume performance
Use CloudWatch metrics like VolumeReadOps, VolumeQueueLength, and BurstBalance (for gp2/st1/sc1). If performance degrades, consider modifying the volume to increase IOPS or throughput, or changing the volume type entirely. Also monitor costs and right-size periodically. This step ensures the volume continues to meet requirements efficiently.
Practical Mini-Lesson
In real-world cloud architecture, EBS volume types are not just a multiple-choice option. They influence everything from application design to operational monitoring. As a systems administrator or solutions architect, you must understand the interplay between volume type, instance type, and network performance.
First, always enable EBS optimization on your EC2 instances. This dedicates network bandwidth between the instance and EBS, preventing contention. For Nitro-based instances, EBS optimization is automatically enabled. Without it, your high-performance io2 volume may be throttled by the instance's network limits.
Second, understand the burst credit model for gp2, st1, and sc1. A gp2 volume starts with a bucket of burst credits (5.4 million IO credits for volumes up to 1 TB). As you use I/O, credits deplete. When the bucket is empty, performance drops to baseline (3 IOPS per GB). For a 100 GB gp2 volume, baseline is 300 IOPS. If your application suddenly needs 3,000 IOPS for an hour, you burn 10,800 credits. Once credits are gone, the volume becomes very slow. This is why many production environments migrate to gp3, which offers consistent baseline performance without credit management.
Third, be aware of volume encryption. When you create an EBS volume, you can enable encryption at rest. This is mandatory in many compliance frameworks. Encrypted volumes are created using AWS KMS keys. Performance impact is minimal, but you must ensure your IAM roles allow encryption actions.
Fourth, when resizing volumes, you can increase size and performance online without detaching. However, you may need to extend the file system to use the extra space. For Linux, this involves growpart and resize2fs. For Windows, use Disk Management. Failing to extend the file system means the extra capacity is wasted.
Finally, cost optimization is an ongoing process. Use AWS Compute Optimizer to get recommendations on EBS volume types. It analyzes utilization patterns and suggests downsizing or changing types. For example, if a gp3 volume uses only 20% of its provisioned IOPS for a week, Compute Optimizer might recommend switching to a smaller volume or changing to a lower performance tier. Automate this with AWS Lambda and CloudWatch Events to run monthly.
Professionals also use EBS Multi-Attach for clustered databases like SQL Server Failover Cluster Instance (FCI) or SAP HANA. This requires a cluster-aware file system (like NTFS or OCFS2) and the volume must be in the same Availability Zone. Many developers forget that multi-attach volumes cannot be used with instance store-backed instances.
EBS volume types are a foundational element of AWS storage. Mastering them means understanding workload patterns, performance limits, cost trade-offs, and monitoring. The best architects think in terms of 'the right tool for the right job' and continually refine storage choices as application requirements evolve.
Memory Tip
SSD for Database, HDD for Data Lake, gp3 is the default gateway, io2 for extreme IO, st1 for stream, sc1 for seldom.
Covered in These Exams
Current Exam Context
Current exam versions that test this topic — use these objectives when studying.
Related Glossary Terms
A 2-in-1 laptop is a portable computer that can switch between a traditional laptop form and a tablet form, usually by detaching or rotating the keyboard.
The 24-pin motherboard connector is the main power cable that connects the computer's power supply unit (PSU) to the motherboard, supplying electricity to the motherboard and its components.
Two-factor authentication (2FA) is a security method that requires two different types of proof before granting access to an account or system.
A 3D printer is a device that creates physical objects by depositing layers of material based on a digital model.
5G is the fifth generation of cellular network technology, designed to deliver faster speeds, lower latency, and support for many more connected devices than previous generations.
The 8-pin CPU connector is a power cable from the power supply that delivers dedicated electricity to the processor on a computer's motherboard.
802.1Q is the networking standard that allows multiple virtual LANs (VLANs) to share a single physical network link by tagging Ethernet frames with VLAN identification information.
802.1X is a network access control standard that authenticates devices before they are allowed to connect to a wired or wireless network.
Frequently Asked Questions
Can I change an EBS volume type after I create it?
Yes, you can modify the volume type of an existing EBS volume, as well as its size and IOPS/throughput settings, without detaching it from the instance. The volume remains available during the modification process.
What is the difference between gp2 and gp3 volumes?
gp3 volumes provide baseline performance that is independent of volume size, offering 3,000 IOPS and 125 MB/s throughput for any size. gp2 performance scales with size (3 IOPS per GB) and uses a burst credit model. gp3 is generally more cost-effective and predictable.
Which EBS volume types are bootable?
Only SSD-backed volume types (gp2, gp3, io1, io2) can be used as boot volumes for EC2 instances. HDD-backed types (st1, sc1) are not supported as root devices because they are optimized for sequential I/O.
What is the maximum IOPS I can get from a single EBS volume?
With io2 Block Express, a single volume can achieve up to 256,000 IOPS. gp3 supports up to 16,000 IOPS. st1 and sc1 are limited to a maximum of 500 and 250 IOPS respectively, but they are designed for throughput, not IOPS.
Can I attach an EBS volume to multiple EC2 instances at the same time?
Yes, but only io1 and io2 volumes support the multi-attach feature. You can attach the same volume to up to 16 instances within the same Availability Zone. This is used for clustered databases like SQL Server FCI or SAP HANA.
What happens if I run out of burst credits on a gp2 volume?
When burst credits are exhausted, the volume's performance drops to its baseline IOPS (3 IOPS per GB). This can cause significant performance degradation until credits reset. Switching to gp3 eliminates this credit management issue.
Summary
Amazon EBS volume types are a critical concept in cloud storage and a frequent topic in AWS certification exams. The main types include General Purpose SSD (gp2, gp3), Provisioned IOPS SSD (io1, io2), Throughput Optimized HDD (st1), and Cold HDD (sc1). Each is designed for specific workload patterns: SSD for transactional, latency-sensitive applications; HDD for sequential, throughput-heavy tasks.
Choosing the correct volume type involves balancing performance, cost, and durability. gp3 is the recommended default for most workloads due to its predictable pricing and independent scaling of IOPS and throughput. For mission-critical databases requiring consistent high IOPS and low latency, io2 with multi-attach capability is the best choice. For big data and log processing, st1 offers cost-effective high throughput. For archival data, sc1 provides the lowest storage cost.
Exam takers must memorize the performance limits and use cases for each type. Common mistakes include using HDD as a boot volume, confusing instance store with EBS, and ignoring burst credits. Practical application involves continuous monitoring with CloudWatch and using AWS Compute Optimizer for recommendations.
Ultimately, mastery of EBS volume types enables you to design resilient, performant, and cost-efficient cloud architectures. Whether you are preparing for the AWS Certified Solutions Architect, SysOps Administrator, or Developer exams, understanding these storage choices is essential. Use the memory tip: SSD for Database, HDD for Data Lake, and let that guide your choices.