Aisalkyn Aidarova

Posted on Oct 8

Solution architect associate level

#architecture #learning #aws #cloudcomputing

Amazon EBS (Elastic Block Store)

🧩 What It Is

EBS stands for Elastic Block Store — it’s like a network hard drive for your EC2 instance.
You can store data on it and keep it safe even after the instance is stopped or terminated.

💽 Key Features

Persistent Storage

Data remains even if the EC2 instance stops or is deleted.
You can detach an EBS volume from one instance and attach it to another.

Bound to an Availability Zone (AZ)

When you create an EBS volume in us-east-1a, it cannot be attached to an instance in us-east-1b.
You can move data to another AZ only by creating a snapshot and then restoring it in another zone.

Network-Based Drive

It’s not physically attached; EC2 connects to it through the network.
That can add a little latency, but it’s flexible and easy to manage.

Provisioning

You must specify:
- Size (in GB)
- Performance (IOPS) — Input/Output Operations Per Second
You’re billed for the size and performance you request.

🧠 Think of It Like

A network USB stick:

You can plug it into one computer (EC2 instance),
Unplug it, and plug it into another.
You can even have multiple USBs (EBS volumes) connected to one machine.

🗂️ Delete on Termination

When you create an EC2 instance:

The root EBS volume (where the OS is installed) has “Delete on Termination” enabled by default.
- So when the instance is terminated, that EBS is deleted too.
Additional volumes are not deleted by default (so your data stays safe).

✅ You can change this behavior:

Enable it → delete the volume when the instance is deleted.
Disable it → keep the volume even after the instance is gone.

Example use case:
You disable “delete on termination” if you want to keep logs or databases after deleting the instance.

🏗️ Example Setup

us-east-1a
 ├── EC2 instance #1
 │     ├── EBS Volume A (root, delete on termination = true)
 │     └── EBS Volume B (data, delete on termination = false)
 └── EC2 instance #2
       └── EBS Volume C (root)

Each instance can have multiple EBS volumes.
Each EBS volume can only be attached to one instance at a time.

Hands-On: Exploring and Managing EBS Volumes in AWS

🧩 1. Check Existing Volumes on an Instance

Go to the EC2 Console → Instances.
Click on your running instance.
Open the Storage tab.

You’ll see:
- Root device (usually 8 GB)
- Block device showing the attached EBS volume
  1. Click on the Volume ID to open it in the EBS → Volumes section.

Result:
You’ll see your current volume (8 GB, “in use”) attached to your EC2 instance.

💽 2. Create and Attach a Second Volume

In the left menu, choose Volumes → Create volume.
Select:

Volume type: gp2
Size: 2 GB
Availability Zone (AZ): must match your instance’s AZ (Check your instance’s AZ under the Networking tab — e.g. eu-west-1b)
1. Click Create volume.

After creation:

Status = “Available” (not yet attached)

Select the volume → Actions → Attach volume.
Choose your EC2 instance → Attach.

Result:
Your instance now has two EBS volumes (one 8 GB root + one 2 GB extra).

⚙️ 3. Verify Attachment

Return to your instance → Storage tab.
You’ll now see two block devices:
- 8 GB (root)
- 2 GB (new volume)

To use the new one, you would format and mount it from Linux,
but that’s advanced (out of scope for beginners).

🗺️ 4. Demonstrate AZ Limitation

Create another volume (again gp2, 2 GB).
Choose a different Availability Zone (e.g., eu-west-1a instead of eu-west-1b).

Now try to attach this new volume to your instance.

Result:
You cannot attach it, because:

EBS volumes are tied to a specific Availability Zone.

This is a critical exam and real-world point — volumes and instances must be in the same AZ.

🧹 5. Delete a Volume

If a volume is not needed:

Select it in Volumes.
Click Actions → Delete volume.
Confirm deletion.

It will disappear immediately — showing AWS’s flexibility and speed.

🔄 6. Understanding "Delete on Termination"

When you launch an instance, each storage device has a Delete on Termination setting.

Root volume: “Yes” by default → deleted when instance terminates.
Additional volumes: “No” by default → remain after termination.

You can check this under:

Instance → Storage → Block devices → scroll right to the “Delete on Termination” column.

If you terminate the instance:

Root (8 GB) volume → deleted.
Attached data (2 GB) volume → remains available.

Use Case:
Keep “Delete on Termination = No” if you want to preserve logs or databases after instance deletion.

✅ Final Verification

After instance termination:

Go to EC2 → Volumes.
Refresh.
The root 8 GB volume disappears.
The 2 GB volume remains, showing it survived the termination.

💡 Summary

Action	Result
Create instance	8 GB root volume auto-attached
Add new volume (same AZ)	Attach successfully
Add new volume (different AZ)	Cannot attach
Terminate instance	Root deleted, attached one kept
Delete volume manually	Removed instantly

Amazon EBS Snapshots

💡 What Is an EBS Snapshot?

An EBS Snapshot is a backup (point-in-time copy) of your EBS Volume.
It allows you to save the current state of your volume data — so you can restore it later if needed.

Think of it like taking a photo of your EBS volume at a moment in time.
You can create a new EBS volume from that photo anytime.

🧩 1. How It Works

Concept	Description
No Detachment Required	You can take a snapshot while the volume is still attached to an EC2 instance. (However, for consistent data like databases, it’s recommended to stop the instance or unmount the volume first.)
Cross-AZ or Cross-Region	Snapshots can be copied across Availability Zones or Regions, allowing you to restore EBS volumes anywhere.
Use Case	Move or duplicate a volume from one AZ to another using snapshots.

Example:

EBS Volume → Snapshot → Restore in another AZ
us-east-1a          copy → us-east-1b

This is how you “move” or “replicate” an EBS volume to another zone or region.

📦 2. Key Features of EBS Snapshots

(a) Snapshot Archive

Lets you move snapshots into a cheaper storage tier (up to 75% cost savings).
Restoring from the archive takes 24–72 hours.
Ideal for long-term backups that you rarely need to access.

Think of it like: moving old photos to cold storage — cheaper, but slower to retrieve.

(b) Recycle Bin for Snapshots

If you accidentally delete a snapshot, it goes to the Recycle Bin instead of disappearing immediately.
You can recover deleted snapshots from there.
Retention period can be set between 1 day and 1 year.

Use case: protect against accidental deletions or human error.

(c) Fast Snapshot Restore (FSR)

Makes the snapshot instantly usable (no initial latency).
AWS usually initializes snapshots gradually; with FSR, initialization happens immediately.
Great for large or critical volumes that must be restored fast.
Warning: It’s expensive — only use it when needed.

Example use case: restoring a 1 TB database snapshot for immediate production use.

⚙️ 3. Typical Workflow

Step	Action
1	Create snapshot of an existing EBS volume
2	(Optional) Copy it to another Region or AZ
3	(Optional) Move to Archive tier for cheaper storage
4	(Optional) Enable Fast Snapshot Restore
5	Restore snapshot into a new EBS volume when needed

🧠 Quick Summary

Feature	Purpose	Cost	Restore Time
Standard Snapshot	Normal backup	Normal	Instant
Snapshot Archive	Cheaper, long-term storage	75% cheaper	24–72 hours
Recycle Bin	Protects deleted snapshots	None	N/A
Fast Snapshot Restore	Instant recovery, no latency	Expensive	Instant

Hands-On: Working with EBS Snapshots in AWS

🧩 1. Take a Snapshot from an Existing EBS Volume

Go to EC2 → Volumes.
Select the 2 GB GP2 volume you created earlier.
Choose Actions → Create snapshot.
Add a description (e.g. DemoSnapshot) and click Create snapshot.
On the left menu, open Snapshots.

You’ll see the new snapshot with Status: Completed (100%) once ready.

✅ Result: You’ve successfully created a point-in-time backup of your volume.

🌍 2. Copy the Snapshot to Another AWS Region

This is useful for disaster recovery or multi-region backups.

Select your snapshot → Actions → Copy snapshot.
Choose any destination region (e.g. us-east-1, us-west-2, etc.).
Click Copy snapshot.

✅ The snapshot will now be replicated in the new region — a key step in cross-region backup strategies.

💽 3. Create a New Volume from the Snapshot

You can restore the snapshot into a new EBS volume — even in a different AZ.

In the Snapshots view → select the snapshot.
Click Actions → Create volume from snapshot.
Configure:

Volume type: gp2
Size: 2 GB
Availability Zone: choose a different one (e.g. eu-west-1b if the original was eu-west-1a)
Encryption: optional
1. Click Create volume.

Then go to EC2 → Volumes, and you’ll see:

The original 2 GB volume (source)
The new 2 GB volume, restored from a snapshot, in another AZ.

✅ This demonstrates how snapshots allow you to move data between AZs.

🗑️ 4. Protect Snapshots with the Recycle Bin

The Recycle Bin prevents accidental snapshot deletion.

Go to Recycle Bin → Create retention rule.
Name it DemoRetentionRule.
Select EBS Snapshots.
Apply to All resources.
Set Retention period to 1 day.
Leave Rule Lock unchecked (so you can edit/delete later).
Click Create retention rule.

✅ Now any deleted snapshots will first move to the Recycle Bin instead of being permanently removed.

🏷️ 5. Move a Snapshot to the Archive Tier

Go to EC2 → Snapshots.
Select the snapshot → Actions → Modify tier / Archive snapshot.
Confirm archiving.

💰 Benefits:

Up to 75 % cheaper storage.
Restore time: 24–72 hours. Use this for long-term, rarely-needed backups.

♻️ 6. Test Deletion and Recovery

In EC2 → Snapshots, select the snapshot and click Delete snapshot.

It disappears from the list.
1. Go to Recycle Bin → Resources → refresh.
Your snapshot appears there.
1. Select it → Actions → Recover resources.
2. Return to EC2 → Snapshots — it’s back!

✅ This proves the Recycle Bin works as expected — you can safely recover deleted snapshots.

🧠 Summary Table

Feature	Purpose	Key Benefit	Notes
Create Snapshot	Backup EBS volume	Point-in-time restore	Can be done while attached
Copy Snapshot	Move across Regions	Disaster recovery	Secure multi-region storage
Create Volume from Snapshot	Restore or clone data	Cross-AZ movement	Useful for scaling
Recycle Bin	Protect snapshots	Recover from deletion	Retention: 1 day–1 year
Archive Tier	Long-term, low-cost	Save 75 % storage cost	Restore: 24–72 hours

Amazon Machine Image (AMI)

🧩 What Is an AMI?

AMI stands for Amazon Machine Image — it’s the template used to launch an EC2 instance.

Think of it as a “golden image” that contains:

The operating system (Linux, Windows, etc.)
Software and packages (e.g., Nginx, Docker, Python)
System configuration (users, network settings, scripts)
Monitoring and security agents (e.g., CloudWatch Agent)

Every EC2 instance starts from an AMI — either one provided by AWS or one you create yourself.

⚙️ 1. Types of AMIs

Type	Description	Example
Public AMIs (AWS-provided)	Default, ready-to-use AMIs provided by AWS	`Amazon Linux 2`, `Ubuntu`, `Windows Server`
Custom AMIs (User-created)	You create these yourself after customizing an instance	Your own preinstalled web app, tools, users, etc.
Marketplace AMIs	Created and sold by third parties on the AWS Marketplace	Bitnami, Fortinet, Jenkins, or commercial software

✅ Note: You can even sell your own AMIs on the AWS Marketplace if you package software or configurations that others need.

🚀 2. Why Use a Custom AMI?

Benefit	Explanation
Faster Boot and Configuration	The software you pre-installed is already there — no need to reconfigure every new instance.
Consistency	Every new instance is identical — same setup, same packages, same tools.
Scaling and Automation	Perfect for Auto Scaling Groups — new instances can launch instantly using your pre-built AMI.
Disaster Recovery	Rebuild your system anywhere, anytime, using the saved AMI.

🧱 3. How an AMI Is Created (The Process)

Step 1: Launch an EC2 instance (e.g., in us-east-1a)
Step 2: Customize it — install software, configure settings, test it
Step 3: Stop the instance (to ensure file system consistency)
Step 4: Choose Actions → Create Image (AMI)
Step 5: AWS automatically creates:

A Custom AMI
A set of EBS Snapshots behind the scenes (These snapshots store the actual data that make up your AMI.)

Step 6: You can now:

Launch new EC2 instances from this AMI
Copy this AMI to other regions for global use

🌍 4. Cross-Region or Cross-AZ Usage

Once created, your AMI belongs to a specific region (e.g., us-east-1).

If you want to use it elsewhere:

You can copy the AMI to another region (e.g., us-west-2).
Then, launch instances from it in that new region or AZ.

Example:

Launch EC2 in us-east-1a → Create custom AMI → Copy AMI → Launch EC2 in us-east-1b

✅ This is how companies replicate server setups across multiple regions for high availability or disaster recovery.

🧠 Key Points to Remember

Concept	Description
AMI = EC2 template	Defines what your instance contains and how it runs
Backed by EBS Snapshots	Snapshots store the root volume data
Region-specific	Must be copied to use in other regions
Useful for scaling	Auto Scaling Groups depend on AMIs
Can be shared or sold	You can make AMIs public or sell them on AWS Marketplace

🖼️ Simple Analogy

An AMI is like a blueprint for a house.
You can build (launch) as many identical houses (instances) as you want from that same design.
And if you move to another city (region), you can copy the blueprint and build there too.

Hands-On: Creating and Using an AMI (Amazon Machine Image)

🧩 Goal

You’ll create an EC2 instance, install Apache (HTTPD), then save it as an AMI so that new instances can launch instantly without re-installing anything.

⚙️ 1. Launch a Base EC2 Instance

Go to EC2 → Launch instance.
Name it something like AMI-Demo.
Choose AMI: Amazon Linux 2.
Instance type: t2.micro.
Key pair: select any (e.g., easy2-draw).
Network settings:

Choose an existing security group (e.g., launch-wizard-1)
Make sure HTTP (port 80) is allowed.
1. Storage: leave defaults.
2. Advanced details → User data:

Paste the first four lines of your previous HTTPD setup script (installing Apache) — but omit the last line that creates the index.html file.

Example:

#!/bin/bash
yum update -y
yum install -y httpd
systemctl enable httpd
systemctl start httpd

Click Launch instance.

⏱️ 2. Wait for Initialization

The instance will begin running, but user data scripts take 1–2 minutes to complete.
Even if the instance status says “Running”, wait until the web server is installed.
After a few minutes, open the Public IPv4 address in your browser using http:// (not https).

✅ You should see the default Apache test page.

💾 3. Create a Custom AMI

Once your instance is ready:

Right-click (or select) the instance → Image and templates → Create image.
Name it demo-image.
Leave the default settings and click Create image.

Then go to the left menu → AMIs.

You’ll see your new AMI listed with Status: pending.
Wait until it changes to available.

✅ You now have your own custom AMI that contains:

Amazon Linux 2
Apache pre-installed and configured to start automatically

🚀 4. Launch a New Instance from Your Custom AMI

In the AMIs page, select your AMI (demo-image).
Click Launch instance from AMI.
Configure the instance as before:

Instance type: t2.micro
Key pair: any
Security group: launch-wizard-1
1. Advanced details → User data:

Now paste only the lines that create the webpage, since HTTPD is already installed.

Example:

#!/bin/bash
echo "Hello World from AMI" > /var/www/html/index.html
systemctl restart httpd

Launch the instance.

⚡ 5. Test the AMI Speed

Wait for the instance to show “Running”.
Copy its Public IP and open it in your browser: http://<Public-IP>

✅ You should see:
“Hello World from AMI”

Notice how this instance started much faster — because:

Apache was already pre-installed inside the AMI.

🧠 6. Why This Is Powerful

Advantage	Explanation
Faster Boot	No need to reinstall packages; everything is pre-baked.
Consistency	Every instance is identical — great for scaling.
Automation	Ideal for Auto Scaling Groups, CI/CD pipelines, or disaster recovery.
Custom Software	Include your own monitoring tools, security agents, or company apps.

🧹 7. Clean Up

When you’re done:

Terminate both EC2 instances.
Keep the AMI for future labs (optional).

🖼️ Process Summary Diagram

Step 1: Launch EC2 → Install Apache (via user data)
Step 2: Test instance
Step 3: Create AMI → "demo-image"
Step 4: Launch new EC2 from demo-image
Step 5: Add custom index.html
Result: Fast boot, Apache already configured ✅

Amazon EC2 Instance Store (Ephemeral Storage)

🧩 What It Is

An EC2 Instance Store is a temporary, high-performance storage that is physically attached to the host server running your EC2 instance.
Unlike EBS (Elastic Block Store), which is network-attached, the Instance Store is directly connected to the underlying hardware — giving it much higher I/O performance.

⚙️ 1. How It Works

When you launch certain EC2 instance types (like i3, d2, or m5d), they come with Instance Store volumes pre-attached.

These disks:

Are located on the same physical server as your EC2 instance.
Are faster than network storage (EBS) because data doesn’t travel over the network.
Are temporary (ephemeral) — meaning data is lost if:
- The instance is stopped, terminated, or
- The underlying hardware fails.

⚡ 2. Key Characteristics

Feature	EC2 Instance Store	EBS Volume
Type	Physical (local) disk	Network-attached drive
Performance	Extremely high IOPS (up to millions)	High but lower (e.g. 32,000 IOPS for gp2)
Persistence	Data lost when instance stops/terminates	Data persists until deleted
Use Case	Temporary data, caching, buffer, scratch space	Databases, logs, long-term files
Backup Responsibility	User must handle manually	AWS handles durability and replication
Attachment	Only at instance launch	Can attach/detach anytime
Availability	Only supported by certain instance types	Supported by almost all

🔥 3. Performance Advantage

Instance Store can reach millions of IOPS:
- Example: i3 instances → up to 3.3 million read IOPS and 1.4 million write IOPS.
Compare that with EBS gp2, which provides around 32,000 IOPS.

✅ So if you see “extremely high performance local storage” in an exam question —
think EC2 Instance Store.

🧠 4. Important Notes

Ephemeral nature:
If the instance stops or hardware fails → data disappears permanently.
You must backup or replicate data to persistent storage (e.g., EBS or S3).
Cannot attach after launch:
Instance Store volumes are predefined by instance type and cannot be added later.
Use cases:
- Cache / Buffer storage (e.g., Redis temp data)
- Scratch space for intermediate calculations
- Temporary logs or staging data
- High-performance workloads that don’t require durability

🧱 5. Real-World Example

Imagine you’re processing large amounts of temporary video data:

You launch an i3.large instance (includes NVMe-based Instance Store).
You store intermediate files locally for fast processing.
Once done, you move final outputs to S3 or EBS for persistence.
When you stop or terminate the instance — the local data is gone.

🧩 6. Summary

Feature	Description
Definition	Local, high-speed disk physically attached to the EC2 host
Durability	Data lost on stop/terminate
Speed	Very high (millions of IOPS)
Use Case	Temporary, transient workloads
Backup Needed?	Yes — user responsibility
Persistent Alternative	EBS Volume

💡 Simple Analogy

EBS is like a USB drive plugged over a network — slower but persistent.
Instance Store is like a built-in SSD inside the computer — lightning fast but erased when power goes off.

Amazon EBS Volume Types

🧩 Overview

Amazon EBS (Elastic Block Store) offers 6 volume types, grouped into SSD (Solid State Drive) and HDD (Hard Disk Drive) families — each optimized for a specific use case.

Category	Volume Type	Description
SSD (Performance-based)	`gp2`, `gp3`	General Purpose — balance price and performance
	`io1`, `io2` (Block Express)	Provisioned IOPS — mission-critical, consistent high I/O
HDD (Throughput-based)	`st1`	Throughput Optimized — frequent, large data workloads
	`sc1`	Cold HDD — infrequent access, lowest cost

⚙️ 1. General Purpose SSDs (gp2 & gp3)

gp2 – older generation

Balances cost and performance.
IOPS are tied to volume size:
- Baseline of 3 IOPS per GB
- Up to 16,000 IOPS max
Can burst up to 3,000 IOPS for short periods.
Use cases:
- Boot volumes
- Development and testing
- Web servers
- Virtual desktops

✅ Example:
A 100 GB gp2 volume = 100 × 3 = 300 IOPS.

To reach 16,000 IOPS, volume must be ~5,334 GB.

gp3 – newer generation

Successor to gp2 — cheaper and more flexible.
Provides baseline 3,000 IOPS and 125 MB/s throughput, regardless of size.
You can independently increase:
- IOPS → up to 16,000
- Throughput → up to 1,000 MB/s
Use cases:
- Same as gp2 but for modern workloads (default choice today)

✅ Key difference:
gp3 allows tuning performance independent of volume size — perfect for cost optimization.

⚡ 2. Provisioned IOPS SSDs (io1 & io2 Block Express)

Designed for mission-critical, low-latency, and high-throughput workloads.

io1

Size: 4 GB – 16 TB
Provision up to 64,000 IOPS on Nitro-based instances
Can set IOPS separately from storage size
Supports Multi-Attach (attach one EBS volume to multiple EC2 instances)
Use cases:
- Production databases (e.g., Oracle, MySQL, PostgreSQL)
- High-transaction systems

io2 / io2 Block Express

Next-generation version of io1 with better durability and higher performance.
Size: up to 64 TB
Latency: sub-millisecond
IOPS: up to 256,000
IOPS-to-GB ratio: up to 1,000:1
99.999% durability

Use cases:

Enterprise databases
Financial or healthcare systems needing consistent high I/O

✅ Note:
If you see “critical database workload” or “>16,000 IOPS needed” → choose io1 or io2.

💽 3. HDD Volumes (st1 & sc1)

These are magnetic disks optimized for throughput (MB/s), not IOPS.

st1 – Throughput Optimized HDD

Size: up to 16 TB
Max throughput: 500 MB/s
Max IOPS: 500
Cannot be boot volumes
Use cases:
- Big data analytics
- Data warehouses
- Log processing
- Streaming workloads

sc1 – Cold HDD

Size: up to 16 TB
Max throughput: 250 MB/s
Max IOPS: 250
Lowest cost option
Use cases:
- Archival data
- Backups or rarely accessed datasets

✅ Exam tip:
If the question says “lowest cost” + “infrequently accessed data,” pick sc1.

🧠 4. Boot Volume Rules

Only the following can be used as boot/root volumes:

gp2
gp3
io1
io2

❌ st1 and sc1 cannot be used as boot volumes.

🧾 5. Key Metrics

Metric	Meaning
IOPS (I/O Operations per Second)	Measures how many read/write operations per second the volume can handle.
Throughput (MB/s)	Measures data transfer speed.
Latency	Time delay between request and response.

🧩 6. Summary Table

Type	Category	Boot?	Max Size	Max IOPS	Max Throughput	Best For
gp3	SSD	✅	16 TB	16 000	1 000 MB/s	General purpose, cost-efficient
gp2	SSD	✅	16 TB	16 000	250 MB/s	Older general purpose
io1	SSD	✅	16 TB	64 000	1 000 MB/s	Databases needing consistent IOPS
io2 (Block Express)	SSD	✅	64 TB	256 000	4 000 MB/s	Mission-critical, sub-ms latency
st1	HDD	❌	16 TB	500	500 MB/s	Big data, frequent access
sc1	HDD	❌	16 TB	250	250 MB/s	Cold storage, infrequent access

💡 7. Exam Tip Summary

Scenario	Best Volume Type
Boot volume, web server, dev/test	gp3
Database needing steady performance	io1 or io2
High throughput, frequent access	st1
Archival, rarely accessed data	sc1
Over 32 000 IOPS required	io1/io2 with Nitro instance
Need low cost + SSD speed	gp3

🖼️ Simple Analogy

Think of EBS like car tires:

gp3 = all-purpose tires (balanced, reliable)

io1/io2 = racing tires (maximum performance)

st1 = heavy-duty truck tires (for big loads)

sc1 = spare tire (cheapest, only for emergencies)

EBS Multi-Attach

🧩 What It Is

EBS Multi-Attach allows you to attach a single EBS volume to multiple EC2 instances simultaneously — within the same Availability Zone (AZ).

All attached instances can read and write to the same volume at the same time.
This feature is available only for io1 and io2 (Provisioned IOPS SSD) volume types.

⚙️ 1. Key Characteristics

Feature	Description
Supported Volume Types	`io1` and `io2` only
Availability Zone Scope	Works only within the same AZ
Max Number of Instances	Up to 16 EC2 instances can attach the same volume
Access Type	Each instance gets full read/write access
Performance	High throughput, low latency (since `io1/io2` are used)

💡 2. Why Use Multi-Attach

Use Multi-Attach when multiple servers need to access the same shared storage simultaneously, without creating copies of the data.

Typical use cases:

Clustered Linux applications (e.g., Oracle RAC, Teradata)
High-availability (HA) clusters
Applications managing concurrent writes at the software level
Distributed caching or compute grids that require shared block storage

🚫 3. Limitations

Single AZ only:
You cannot attach the same EBS volume across multiple Availability Zones.
Cluster-aware file system required:

Normal file systems (like ext4, XFS) will corrupt data if multiple servers write at once.
You must use cluster-aware file systems, e.g.:
- OCFS2 (Oracle Cluster File System)
- GFS2 (Red Hat Global File System)
- IBM GPFS
These are designed to handle synchronized access and metadata locking.

Root volumes not supported:
Multi-Attach volumes can only be data volumes, not boot/root volumes.
No cross-region or cross-AZ attachment:
Each Multi-Attach configuration exists inside a single AZ.

🧱 4. How It Works (Conceptually)

Example setup:

us-east-1a
 ├── EC2 instance A
 ├── EC2 instance B
 ├── EC2 instance C
 └── io2 volume (Multi-Attach enabled)

All three EC2 instances share and access the same io2 volume simultaneously.
Each instance sees the same data blocks.
Application logic or the cluster file system ensures synchronization.

🧠 5. Exam Tips

Question Type	Correct Answer / Concept
“Attach one EBS volume to multiple EC2 instances for high availability”	Use io1/io2 Multi-Attach
“Share EBS volume across AZs”	❌ Not possible (use S3, EFS, or FSx instead)
“How many instances can share one volume?”	Up to 16
“Which file system is needed?”	Cluster-aware (OCFS2, GFS2, etc.)
“Which types support Multi-Attach?”	io1 and io2 only

⚙️ 6. Comparison Summary

Feature	gp2/gp3	io1/io2 (Multi-Attach Enabled)
Attach to multiple instances?	❌ No	✅ Yes
AZ restriction	N/A	Same AZ only
Use case	Regular single-instance workloads	Clustered, HA applications
File system	ext4/XFS	Cluster-aware (OCFS2, GFS2)
Performance	Good	Very high

💡 Quick Analogy

Think of Multi-Attach like multiple chefs sharing one kitchen —
they can all cook (read/write), but only if they follow a system (cluster-aware file system)
so they don’t bump into each other and spoil the recipe (data corruption).

EBS Encryption

🧩 What It Does

When you create an encrypted EBS volume, AWS automatically encrypts:

Data at rest – stored inside the volume
Data in transit – moving between the EC2 instance and the EBS volume
Snapshots – created from that encrypted volume
New volumes – restored from encrypted snapshots

Everything connected to an encrypted volume remains encrypted — end to end.

Encryption and decryption happen transparently (you don’t need to change your application or OS).

🔐 1. Encryption Mechanism

Concept	Details
Algorithm	AES-256 encryption
Key Management	AWS KMS (Key Management Service)
Latency Impact	Almost none (handled by hardware acceleration)
Default Encryption	Can be enabled account-wide for all new volumes

✅ Best practice: Always enable EBS encryption — it’s free, secure, and fast.

⚙️ 2. Creating an Encrypted EBS Volume

When launching or creating a volume:

Go to EC2 → Volumes → Create volume
Check “Encrypt this volume”
Choose a KMS key (default or customer-managed)
Click Create volume

Your new volume will display:

State: available
Encryption: encrypted

🔁 3. How to Encrypt an Existing (Unencrypted) EBS Volume

You cannot directly encrypt an existing unencrypted EBS volume.
Instead, you must go through snapshots.

Step-by-step process

Create a snapshot of your unencrypted volume.

Result → Snapshot is also unencrypted.

Copy the snapshot, and during the copy:

Enable Encryption
Select a KMS key
Keep the same region or copy to another region

Once the encrypted snapshot is ready,
Create a new EBS volume from it.
The new volume is now encrypted.
You can attach it to your original EC2 instance.

✅ You’ve successfully converted an unencrypted volume into an encrypted one!

🧩 4. Shortcut Option

Instead of creating a copy, AWS also allows on-the-fly encryption when creating a volume directly from an unencrypted snapshot.

Steps:

Go to your unencrypted snapshot → Actions → Create volume from snapshot
In the creation dialog:
- Enable Encryption
- Choose a KMS key
Click Create volume

Now the new volume is encrypted automatically — a quicker path to secure existing data.

🧠 5. Exam + Practical Notes

Concept	Remember
Encryption is automatic	All linked snapshots and future volumes inherit encryption
KMS keys are used	Either AWS-managed or customer-managed
Cross-region copies	Encryption can be applied during copy
Cannot remove encryption	Once encrypted, a volume or snapshot cannot be unencrypted
Performance	Minimal impact — fully hardware-accelerated
Billing	No extra cost for encryption; only normal EBS usage costs

💡 6. Example Workflow Summary

Unencrypted Volume
   ↓ (Create Snapshot)
Unencrypted Snapshot
   ↓ (Copy with encryption enabled)
Encrypted Snapshot
   ↓ (Create Volume)
Encrypted Volume
   ↓
Attach to EC2 Instance ✅

🧹 7. Cleanup (Best Practice)

After testing:

Delete snapshots (Actions → Delete snapshot)
Delete volumes (Actions → Delete volume)

This keeps your AWS storage costs low.

🔒 8. Quick Analogy

Think of EBS encryption like having a safe inside a secure truck:

The data (safe contents) is encrypted at rest

The truck (EC2 connection) encrypts data while moving

Every copy or duplicate safe (snapshot or volume) is also automatically locked

And you hold the key (KMS) that controls access.

Amazon EFS (Elastic File System)

🧩 What It Is

EFS (Elastic File System) is a fully managed Network File System (NFS) that allows multiple EC2 instances to share the same data simultaneously, even across multiple Availability Zones.

It’s highly available, scalable, and managed by AWS.
It behaves like a shared drive for your Linux servers.
You pay only for what you use — no need to pre-provision storage.

⚙️ 1. Core Characteristics

Feature	Description
Type	Managed NFS (Network File System)
Access	Mountable on many EC2 instances, across multiple AZs
OS Compatibility	Linux only (not compatible with Windows)
Protocol	NFSv4.1
Encryption	Supported at rest and in transit (via KMS)
Scaling	Automatic — grows and shrinks as files are added or removed
Payment	Pay-per-use (per GB stored per month)
Durability & Availability	Multi-AZ replication for Standard class

🌐 2. Architecture Overview

               ┌──────────────┐
us-east-1a →   │  EC2 Instance│
               └──────┬───────┘
                      │
us-east-1b →   ┌──────┴───────┐
               │  Amazon EFS  │  ←  Shared file system (NFS)
us-east-1c →   └──────┬───────┘
                      │
               ┌──────┴───────┐
               │  EC2 Instance│
               └──────────────┘

All instances can read/write to the same files at once — ideal for shared applications.

💡 3. Use Cases

Web content storage (e.g., WordPress shared uploads)
Content management systems (CMS)
Big data and analytics
Application data sharing between multiple servers
Media processing workflows

💰 4. Cost & Performance

EFS is ~3× more expensive than EBS gp2, but offers scalability and multi-instance access.
You don’t provision capacity — AWS handles growth automatically.

⚙️ 5. Performance Modes

Mode	Description	Use Case
General Purpose (default)	Low latency, best for typical use	Web servers, CMS
Max I/O	Higher latency but scales to thousands of connections	Big data, media processing, analytics

⚡ 6. Throughput Modes

Mode	Description	Example Use Case
Bursting (default)	Throughput scales with file system size; includes bursts	Most workloads
Provisioned	Set a fixed throughput (independent of storage size)	Consistent, heavy workloads
Elastic	Automatically scales up/down based on workload	Unpredictable or spiky workloads

Example:

1 TB = 50 MB/s base + bursts to 100 MB/s
Elastic mode can scale up to 3 GB/s read and 1 GB/s write

🗂️ 7. Storage Classes & Lifecycle Management

EFS supports storage tiers to optimize cost automatically.

Storage Tier	Description	Use Case
EFS Standard	Default, for frequently accessed data	Production, shared apps
EFS Infrequent Access (EFS-IA)	Lower storage cost, retrieval fee applies	Files not accessed often
EFS Archive	For rarely accessed files (e.g., once or twice a year)	Long-term storage
EFS One Zone	Stored in a single AZ (lower availability)	Dev/test environments
EFS One Zone-IA	One-AZ + infrequent access	Low-cost development data

You can configure lifecycle policies:

e.g., move files from Standard → IA after 60 days of no access.

✅ This can reduce costs by up to 90 %.

🔐 8. Security & Encryption

Encryption at rest: Managed by AWS KMS (AES-256)
Encryption in transit: Using TLS between EC2 and EFS
Access control:
- Security groups (control network access)
- POSIX file permissions (UID/GID, chmod)

🧠 9. Quick Comparison with Other AWS Storage

Feature	EBS	EFS	S3
Type	Block storage	File storage	Object storage
Access	One instance at a time	Many instances	Any via API
Protocol	Block (attached disk)	NFS	REST/HTTPS
Scaling	Manual (resize)	Automatic	Automatic
Use Case	Databases, OS	Shared files, web apps	Backups, data lake
Pricing	Per GB provisioned	Per GB used	Per GB used

🧩 10. Summary Table

Category	Option	Description
Performance Mode	General Purpose / Max I/O	Balance latency vs. throughput
Throughput Mode	Bursting / Provisioned / Elastic	Control how EFS scales throughput
Storage Class	Standard / IA / Archive / One Zone	Optimize cost and redundancy
File System Type	NFSv4	Linux only
Encryption	At rest & in transit	Uses KMS AES-256
Scalability	Auto, petabyte-scale	Thousands of connections

🧩 11. Exam Tip Summary

Question Mentions	Correct Answer
“Shared file system across AZs”	EFS
“Linux only, NFS protocol”	EFS
“Auto-scaling storage”	EFS
“Need cheaper infrequent access tier”	EFS-IA
“Single AZ, dev environment”	EFS One Zone
“POSIX permissions”	EFS
“Compare cost with gp2”	EFS ≈ 3× more expensive

Hands-On: Amazon EFS (Elastic File System)

🧩 Goal

Set up an EFS file system, connect it to two EC2 instances across different Availability Zones, and verify that both can share files simultaneously.

Step 1 — Create an EFS File System

Open the AWS Console → EFS → Create file system.
You can simply click Create, but to explore options, choose Customize.

Key Options

Setting	Explanation
File System Type	- Regional (multi-AZ, high availability — recommended for production) - One Zone (single AZ, cheaper — for dev/test only)
Automatic Backups	Enable it (recommended)
Lifecycle Management	Move files automatically between tiers (Standard → IA → Archive)
Encryption	Leave enabled (uses KMS AES-256)

Lifecycle Example

Condition	Action
File not accessed for 30 days	Move to Infrequent Access
File not accessed for 90 days	Move to Archive
File accessed again	Move back to Standard

Step 2 — Configure Performance and Throughput

Throughput Modes

Mode	Description	Use Case
Bursting (default)	Scales with storage size	Most general workloads
Elastic	Automatically adjusts throughput based on workload	Recommended — no tuning required
Provisioned	You manually set desired throughput	Predictable, high-traffic workloads

Performance Modes

Mode	Description	Best For
General Purpose	Low latency	Web servers, CMS
Max I/O	Higher latency but higher parallel throughput	Big data, analytics

✅ Recommended setting:
Throughput Mode: Elastic
Performance Mode: General Purpose

Step 3 — Configure Network Access

Select VPC: Choose your default VPC.
Mount Targets: Automatically created across available AZs.
Create Security Group:

Name: sg-efs-demo
Description: EFS Demo SG
Initially no inbound rules (AWS will handle these later).

Refresh the EFS setup page, select your EFS security group, and click Next → Create.

✅ The file system will now show status Available.
You’ll see a small usage (e.g., 6 KB), but costs remain zero until you actually store data.

Step 4 — Launch EC2 Instances

You’ll create two instances in different Availability Zones (to test EFS cross-AZ access).

Instance A

EC2 → Launch instance
Name: Instance-A
AMI: Amazon Linux 2
Type: t2.micro
Key pair: None (use EC2 Instance Connect)
Subnet: eu-west-1a
Storage: Default 8 GB gp2
File System:

Click Add file system → EFS
Select your created EFS
Mount point: /mnt/efs/fs1
AWS will automatically create required security groups and user-data scripts for mounting.

Click Launch Instance ✅

Instance B

Repeat the steps above with:

Name: Instance-B
Subnet: eu-west-1b
Select the same EFS file system.

Step 5 — Verify Network Configuration

After both instances are Running:

In the EFS console → Network tab, notice:
- Auto-created SGs like efs-sg-1 and efs-sg-2
- Each allows NFS (port 2049) inbound access
- Each is linked to the corresponding EC2 instance SG.

AWS automatically handled all the NFS networking for you!

Step 6 — Test EFS Mounting

Connect to both instances using EC2 Instance Connect.
On Instance A, verify the mount:

   ls /mnt/efs/fs1/

Should list an empty directory.

Create a file:

   sudo su
   echo "hello world" > /mnt/efs/fs1/hello.txt
   cat /mnt/efs/fs1/hello.txt

Output:

   hello world

On Instance B, run:

   ls /mnt/efs/fs1/
   cat /mnt/efs/fs1/hello.txt

Output:

   hello world

✅ Both instances see the same file in real-time — confirming shared storage.

Step 7 — Clean Up

Terminate both EC2 instances.
In EFS, select and delete your file system.
Delete the auto-created security groups (efs-sg-1, efs-sg-2, etc.) to avoid clutter.

✅ Summary

Feature	Description
Type	Managed, scalable Network File System (NFS)
Multi-AZ Access	Yes (Regional type)
Auto-Scaling	Grows/shrinks automatically
Encryption	At rest and in transit (via KMS)
Performance Modes	General Purpose / Max I/O
Throughput Modes	Bursting / Elastic / Provisioned
Storage Tiers	Standard / IA / Archive / One Zone
Protocol	NFSv4.1
OS	Linux only
Pricing	Pay-per-use, 3× cost of EBS gp2

💡 What You Learned

How to create an EFS file system
How to enable lifecycle management and encryption
How to auto-mount EFS to EC2 instances
How to verify shared storage access across AZs

EBS vs. EFS vs. Instance Store

🧩 1. Overview

Storage Type	Description	Key Use Case
EBS (Elastic Block Store)	Block-level storage attached to a single EC2 instance (like a hard drive)	Boot volumes, databases, app data
EFS (Elastic File System)	Network File System (NFS) that can be shared across many EC2 instances	Shared data, web apps, CMS (WordPress)
Instance Store	Physical disk directly attached to the host machine	Temporary cache, scratch space, ephemeral data

2. Detailed Comparison

Feature	EBS	EFS	Instance Store
Type	Block storage	File storage (NFS)	Local ephemeral storage
Attachment	One instance at a time (except io1/io2 Multi-Attach)	Many instances across multiple AZs	Only to one EC2 (physically bound)
Availability Zone Scope	Locked to one AZ	Accessible across multiple AZs (Regional)	Same host only
Persistence	Data persists after instance stop/terminate (unless delete-on-termination enabled)	Data persists even if EC2s stop	Data lost when instance stops or fails
Protocol	Block device	NFSv4.1	Direct disk access
Performance	High IOPS (SSD or HDD options)	High throughput, scalable	Extremely high (direct hardware)
Scalability	Fixed size (must provision storage)	Auto-scales to petabytes	Fixed per instance type
Pricing	Pay for provisioned capacity	Pay for storage used (per GB)	Included with instance cost
OS Compatibility	Linux & Windows	Linux only	Linux & Windows
Use Case Examples	Databases, boot disks, logs	Web servers, CMS, shared content	Caches, temporary data processing
Encryption	Supported via KMS	Supported via KMS	Not supported
Backup	Snapshots (EBS Snapshots)	AWS Backup integration	Manual (user responsibility)

3. EBS (Elastic Block Store) Key Notes

Attached to one EC2 in a single AZ
Can migrate to another AZ by:

Taking a snapshot
Restoring the snapshot in a different AZ
- IOPS scaling:

gp2: IOPS increases with size (3 IOPS/GB)
gp3, io1, io2: IOPS and throughput can be set independently
- Multi-Attach: available only for io1/io2 — up to 16 instances in same AZ
- Root volumes deleted by default upon instance termination (configurable)
- Backups use I/O — avoid during high-traffic times

✅ Best for: Boot volumes, databases, app storage needing high performance and durability

4. EFS (Elastic File System) Key Notes

Managed NFS (Network File System) service
Can be mounted by hundreds of EC2 instances
Works across multiple Availability Zones (Regional)
Automatically scales to petabytes — no need to pre-size
Pay-per-use pricing model
Higher cost than EBS (≈ 3×), but shared and scalable
Supports storage tiers:
- Standard (frequent access)
- Infrequent Access (IA) and Archive (for cost savings)
Lifecycle policies move files between tiers automatically

✅ Best for: Shared Linux workloads — e.g. WordPress, content sharing, CMS, home directories, analytics

5. Instance Store Key Notes

Local physical disks attached to the host EC2 server
Offers very high IOPS and throughput
Data lost when instance stops, terminates, or hardware fails
No persistence or snapshot support
Used for temporary data, caches, or buffers
Included in instance pricing (no separate cost)

✅ Best for: Temporary or fast scratch data — not for permanent storage

6. Practical Example

Scenario	Recommended Storage
Web server with persistent logs	EBS (gp3)
WordPress cluster shared across AZs	EFS
Database needing consistent IOPS	EBS io1/io2
Temporary processing (e.g., caching)	Instance Store
Big data or media streaming	EFS (Max I/O)
Boot volume for EC2	EBS

7. Key Takeaways for Exams

Question Mentions	Correct Answer
“Attach to only one EC2 instance”	EBS
“Mount on multiple EC2s across AZs”	EFS
“Temporary storage, lost on stop/terminate”	Instance Store
“Auto-scaling storage, NFS, Linux only”	EFS
“Block-level storage for databases”	EBS
“Need Multi-AZ shared storage”	EFS
“Physically attached storage”	Instance Store

💡 Analogy

EBS = your personal SSD drive — fast, reliable, one user at a time.

EFS = a shared network drive — multiple users access together.

Instance Store = temporary workspace on a computer — gone when powered off.

🧹 AWS Clean-Up After the EBS & EFS Labs

1. Delete the EFS File System

Go to EFS Console → File systems.
Select your file system → Actions → Delete file system.
Copy the File System ID (for example, fs-0abcd1234efgh5678) and paste it into the confirmation box.
Click Delete ✅

Wait a few seconds — status will change to Deleting and then Deleted.

2. Terminate EC2 Instances

Go to EC2 Console → Instances.
Select all running or stopped instances from this lab.
Instance state → Terminate instance.
Confirm termination.

When done, state changes from Shutting down → Terminated.

3. Delete Unused Volumes

Navigate to EC2 Console → Volumes (under Elastic Block Store).
Look for any volumes with State = available (these are detached).
Select them → Actions → Delete volume → Confirm.

This removes any leftover EBS storage that could otherwise incur charges.

4. Delete Snapshots

In EC2 Console → Snapshots, select any snapshots you created during practice.
Actions → Delete snapshot → Confirm deletion.

Snapshots persist in S3 behind the scenes, so deleting them stops storage billing.

5. Delete Extra Security Groups

Go to EC2 Console → Security Groups.
Select and delete unnecessary groups created during this section (for example, efs-sg-1, efs-sg-2, or sg-efs-demo).
Keep the default security group for your VPC — do not delete it.
If deletion fails:

Ensure all EC2 instances using that SG are terminated.
Retry deletion after a few minutes.

6. (Optional) Check Load Balancers or Other Resources

If you created or tested any Elastic Load Balancer or Target Groups, delete them as well:

EC2 Console → Load Balancers → Actions → Delete
Confirm deletion

7. Verify Everything Is Gone

Instances tab → No running/stopped instances.
Volumes tab → Empty.
Snapshots tab → Empty.
EFS console → No file systems.
Security groups tab → Only default groups remain.
Billing console (optional) → Monitor “Zero cost” for EC2, EBS, EFS.

✅ Result:
Your AWS environment is now completely clean — no file systems, no instances, no volumes, and no leftover snapshots.
You’re ready to start fresh in the next section without accumulating charges.