EC2 Instance Storage & EFS

EBS (Elastic Block Store) Overview

A durable, block level storage device you attach to a single EC2 instance. Think of it as a virtual hard drive that happens to be a network drive.

• AZ Locked: An EBS volume is bound to a specific Availability Zone. To move a volume to another AZ, take a snapshot and restore it there.
• Persistence:
  • The Root Volume (where the OS lives) defaults to Delete on Termination.
  • Additional data volumes default to Keep.
  • Both are controlled by the DeleteOnTermination attribute.

EBS Volume Types

The exam expects you to know SSD (performance / IOPS) vs HDD (throughput) cold.

Solid State Drives (SSD)

Optimized for transactional workloads measured in IOPS.

Hard Disk Drives (HDD)

Optimized for large streaming workloads measured in throughput. Cannot be used as a boot volume.

EBS Snapshots

• Mechanism: Incremental backups — only the blocks that changed since the last snapshot are saved.
• Storage: Stored in Amazon S3, but you cannot browse them in your own buckets.
• Scope: Regional. To use a snapshot in another region you must copy it.

AMI (Amazon Machine Image)

• Definition: A blueprint used to create EC2 instances (OS, app server, and applications).
• Scope: Region bound. To launch an AMI in another region, copy the AMI there first.
• Sources:
  • Public AMIs: AWS provided (Amazon Linux 2, Ubuntu, etc.).
  • Your own AMIs: Built from your customized instances.
  • AWS Marketplace: Paid/subscription AMIs from vendors (hardened images, firewall appliances).
• Lifecycle: Create (from instance or snapshot) → Register → Copy (cross region) → Deregister.

EC2 Instance Store (“Ephemeral Storage”)

Physical storage attached directly to the host machine running your instance.

• Performance: Extremely high I/O (millions of IOPS) and the lowest latency of any EC2 storage.
• Volatility — this is the exam critical part:

• Use Cases: Cache, buffers, scratch data, HPC, temporary content that is replicated somewhere else.
• Snapshots: You cannot snapshot an Instance Store volume.

EBS Multi-Attach

Allows a single EBS volume to be attached to multiple EC2 instances at the same time.

• Constraints:
  • Same AZ only.
  • io1 and io2 Block Express volumes only.
  • Nitro based instances only.
  • Requires a cluster aware file system (GFS2, OCFS2). Standard filesystems like ext4, XFS, or NTFS will corrupt data.
• Use Case: Clustered Linux applications (e.g., Teradata), higher availability during updates.

EBS Encryption

• Mechanism: AWS KMS with AES-256.
• Scope: Once encryption is on, it covers:
  1. Data at rest inside the volume.
  2. Data in transit between the instance and the volume.
  3. All snapshots created from the volume.
  4. All volumes created from those snapshots.
• Performance: Minimal latency impact.
• Encryption by default: Can be enabled at the account level, per region.

Amazon EFS (Elastic File System)

Managed NFS that can be mounted on many EC2 instances simultaneously.

• Protocol: NFSv4. Linux only (POSIX compliant).
• Scope: Multi-AZ — data is stored across multiple AZs in a region.
• Scaling: Storage capacity and throughput scale automatically, to petabyte scale.

Performance Modes:

• General Purpose: Web serving, CMS. The default.
• Max I/O: Big Data, parallel processing. Higher throughput but higher latency.

Comparison: EBS vs EFS vs Instance Store