When you launch a virtual server (EC2 instance), it needs storage for its operating system and data. In AWS, this is provided by Elastic Block Store (EBS), the foundational block storage for EC2 instances. Think of it as a virtual hard drive for your cloud server.
This article covers everything you need to know about Elastic Block Store (EBS).
What is Elastic Block Store (EBS)?
Elastic Block Store (EBS) provides persistent block storage. Block storage is used for operating systems, applications, and databases that require persistent, low-latency access.
- Persistence: The most important feature of EBS is that its data is persistent. Data remains intact even if the attached EC2 instance is stopped or restarted.
- Root Volume: The operating system must be installed on a volume, which is called the root volume. This volume must be an EBS volume because it needs to persist. The minimum size for a root volume is typically 8 GB to hold the OS.
- Network Attached: EBS volumes are network-attached devices. This makes them flexible; they can be created independently and attached to an existing instance later.
- Availability Zone (AZ) Restrictions: EBS volumes are zonal. A volume can only be attached to an EC2 instance that resides in the exact same Availability Zone.
EBS Volume Types: SSD vs. HDD
EBS performance is measured by IOPS (Input/Output Operations Per Second) and throughput (data flow rate). To meet different needs, EBS offers volumes based on Solid State Drive (SSD) and Hard Disk Drive (HDD) technology.
- SSD Volumes (Transactional): Optimized for transactional workloads and low latency.
- General Purpose SSD (GP2/GP3): The default choice, suitable for general use cases, boot volumes, and development environments. They can size up to 16 TB and offer a maximum of 16,000 IOPS.
- Provisioned IOPS SSD (IO1/IO2): Designed for heavy, latency-sensitive databases and high-performance computing. They support massive amounts of IOPS.
- HDD Volumes (Throughput): Cheaper and optimized for large, sequential workloads where throughput matters more than IOPS.
- Throughput Optimized HDD (ST1): User for log processing, data warehouses, and big data applications.
- Cold HDD (SC1): Offers the lowest storage cost for infrequently accessed archive data.
- Note: HDD volumes (ST1 and SC1) cannot be used as boot volumes (root volumes).
EBS Snapshots: Backups
Snapshots are point-in-time backups of EBS volumes. They are crucial for disaster recovery and are stored in Amazon S3.
The key feature of EBS snapshots is that they are incremental. The first snapshot copies all the data. Subsequent snapshots only back up the data blocks that have changed since the last snapshot. This saves space and money. If an older snapshot is deleted, its unique data is propagated to the next snapshot to ensure you always have a complete backup for a full restore.
Ephemeral Storage: Instance Store
It is important to contrast EBS with Instance Store. Instance Store provides temporary or ephemeral block storage.
- Non-Persistence: Data stored in an Instance Store volume is lost if the instance is stopped, terminated, or fails.
- Use Case: This storage is only available with specific instance types and is useful for temporary data like caches or buffers persistence is not required. It cannot be used for persistent data.
Conclusion: The "Hard Drive" of the Cloud
You now understand the two types of storage directly attached to virtual servers. Elastic Block Store (EBS) is the persistent, network-attached “hard drive” that stores your OS and critical data. You’ve learned how to choose between SSD and HDD volumes based on performance needs (IOPS vs. Throughput) and how to back it all up with incremental Snapshots.
We also contrasted this with the temporary, high-speed Instance Store. But this only covers storage for server. What about storing backups, media files, or shared data?
In the next article, we’ll explore standalone storage services, focusing on the “infinite” capacity of Amazon S3 for object storage and EFS for shared file storage.
