Session - 27 (18-01-2025)

Main topics covered (Auto Scaling, Load Balancing, EC2/IPs, Bootstrapping)

Recap: Auto Scaling + Load Balancer

Auto Scaling definition: Automatically scale out/in EC2 instances based on load.

Key capacity parameters:

• Minimum (lower boundary)
• Maximum (upper boundary)
• Desired capacity (initial number of instances; example: desired = 4, min = 2, max = 6/8)

Scaling types:

• Manual scaling: manually edit desired capacity
• Scheduled scaling: scale based on time/day
• Dynamic scaling: automatic scale out/in based on load

Auto Scaling Group (ASG) + Launch Template:

• ASG launches instances using a Launch Template
• Instructor emphasis: without a Launch Template, ASG can’t know how to launch instances
• A launch template includes the steps/configuration needed to launch EC2

Integration mentioned:

• Notifications via SNS
• Traffic distribution via Elastic Load Balancer (ELB)

ELB concept:

• ELB is a managed service (not a server)
• Distributes traffic across multiple EC2 instances across Availability Zones

IP handling for ELB:

• ELB IPs are dynamic
• Recommendation: use the DNS name instead of the IP

---

Load Balancer types (AWS ELB)

AWS ELB includes 4 types:

• Classic
• Application
• Network
• Gateway

Classic:

• “previous generation”
• Works with HTTP/TCP

Application Load Balancer (ALB):

• latest generation / commonly used
• Supports HTTP/HTTPS
• Key feature: routing rules
• Forwards traffic to Target Groups

Network Load Balancer (NLB):

• For very high performance
• Supports TCP/UDP/TLS
• Static IP support: 1 IP per Availability Zone

Gateway Load Balancer:

• For security appliances
• Uses port 6081 (as stated)

---

Target Groups (ALB core concept)

• Target group must be created before the load balancer
• A target group contains targets (typically EC2 instances)
• A single load balancer can have multiple target groups (e.g., for different apps)
• Routing rules determine how traffic is forwarded to targets
• Instructor note: load balancer listens for HTTP/HTTPS only

---

New topic: EC2 IP address types (Public, Private, Elastic IP)

The 3 IP types taught

1. Private IP

   • Mandatory by default on every EC2 instance
   • Example range: starts with 172.31.x.x in AWS default VPC
   • Used for communication inside the VPC (EC2-to-EC2 within VPC)

2. Public IP

   • Optional
   • Used to access EC2 from the internet
   • Dynamic: changes on stop/start
   • Recommendation: generally don’t expose EC2 to the internet unless necessary

3. Elastic IP (EIP)

   • Static
   • Prevents IP change even after stop/start
   • Instructor note: AWS provides “five EIPs free” but charges when unused/idle (policy explained with a “chocolates” analogy)

---

Access scenarios explained

• Internet user → needs Public IP/EIP
• Company user via VPN → can use Private IP
• EC2 instances within AWS/VPC communicate using Private IP

---

Credentials recap

• Linux access: default user ec2-user
• Password/access via PEM key file

---

EC2 metadata vs user data + automation (bootstrap script)

Instance metadata

• Defined as “data about instance”
• Includes: who launched it, AMI, key pair, security groups, attached volumes, etc.
• Access via:
  • EC2 console Details tab
  • IMDS URL: http://169.254.169.254/latest/meta-data
• Uses curl to fetch metadata content (called out as a certification-type detail)

User data

• Defined as scripts/config provided at instance launch
• In the EC2 7-step launch flow, user data is in the user data box
• Treated as a bootstrap script that runs at boot time
• Example automation: install Tomcat automatically without manually logging in
• Script types by OS:
  • Linux: shell script
  • Windows: PowerShell

Practical use case:

• Launch multiple instances (e.g., 10) with the same setup using user data
• Alternative: create an AMI from a configured instance, then launch from AMI

---

EC2 launch “7-step” structure (checklist emphasized)

Instructor consolidated a 7-step EC2 theory framework:

1. Name and Tags
2. Choose AMI (example: Red Hat for Linux)
3. Choose instance type (example: T2 micro; other types are billable)
4. Key pair (for login; uses PEM)
5. Security group (inbound rules emphasized)
6. Storage / Volumes
   • root volume OS
   • volume types listed: GP2, GP3, IO1, IO2, ST1, SC1, Standard
7. VPC/subnet selection and then Launch
   • example: default VPC with 3 subnets for 3 AZs

Also noted:

• Default volume sizes:
  • Windows 30 GB
  • Linux 8/10 GB
• Volume size can be increased/decreased was mentioned as “No” (per transcript)

---

Advanced add-on: Global Accelerator (GA), unicast vs anycast, and entry points

Latency concept + CloudFront

• Far distance (e.g., US → Mumbai) causes high latency
• Suggested solution: CloudFront using edge locations

Global Accelerator (GA) introduced

• Provides two static IPs
• Uses anycast behavior (traffic goes to the nearest endpoint)
• Sits in front of load balancer / works with CloudFront (as described)
• Instructor scenario: customer asks for website IP → provide GA IPs
• Mentioned cost: $18/month

Unicast vs Anycast (as taught)

• Unicast: each server has a unique IP
• Anycast: many servers can share the same IP; the nearest responds

Entry points for AWS traffic mentioned

• Load balancer and CloudFront as entry points
• WAF/Shield options: WAF and DDoS protection can be enabled (checkbox-style integration mentioned)

---

Course/plan announcements (reviews, practice, schedule)

• EC2 theory completed; next starts EC2 practicals (from Monday).
• Practicals approach:
  • instructor demonstrates
  • students should observe first (no parallel clicking while being taught)
• Students encouraged to revise EC2 topics beforehand:
  • pricing models, instance types, volume types
  • snapshots, AMIs
  • security groups/inbound-outbound rules, etc.
• Communication: use Telegram (linked via Google Classroom)
• Certification focus later: Solution Architect Associate level (not “practitioner”)
• Suggested study flow: revise EC2 topics → then do practicals
• Course has a structured timeline:
  • first month: basics/EC2
  • next month: more services

---

Main speakers / sources

• Primary speaker: The course instructor (referred to as “Sir”; name not explicitly given in subtitles)
• Student/asker mentions (not main speakers):
  • Avaneesh Pandey, Zakir, Naveen, Roshan, Pranjal, Abby, Tarun, Vishal