CAT

UNIT 1
✅ UNIT 1 – FINAL REVISION NOTES (Cloud Automation Tools & Applications)

⭐ 1. Automation & Configuration Tools – Key Points

What is Automation?

 Automating repetitive cloud tasks (provisioning, deployment, scaling).

 Reduces manual effort, errors, and deployment time.

 Ensures consistency + repeatability (important outcome word).

Types of Tools

Category Examples Purpose

Provisioning Terraform, CloudFormation Create cloud infra

Configuration Management Ansible, Chef, Puppet Install software, set configs

Orchestration Kubernetes Manage container lifecycle

Why Automation?

 Faster delivery

 Standardization

 Version-controlled infrastructure (IaC)

 Scalability

 Lower operational cost

⭐ 2. Terraform – Key Points

What is Terraform?

 IaC tool from HashiCorp

 Uses HCL (HashiCorp Configuration Language)

 Cloud-agnostic (AWS, Azure, GCP, VMware)

Terraform Core Files

 [Link] → resources

 [Link] → inputs
  [Link] → outputs

 [Link] → variable values

 .tfstate → system state (VERY IMPORTANT)

Terraform Workflow — Ye 100% Exam Point

Write → Init → Plan → Apply → Destroy

Common Commands

 terraform init – initializes provider plugins

 terraform plan – preview

 terraform apply – deploy

 terraform destroy – remove infra

 terraform fmt – format code

 terraform validate – syntax check

State File Importance

 Stores the real-world resource mapping

 Needed for updates, destroys

 Can be stored remotely (S3 + DynamoDB lock)

⭐ 3. Terraform vs CloudFormation – Exam Difference

Feature Terraform CloudFormation

Developer HashiCorp AWS

Cloud Support Multi-cloud AWS only

Language HCL JSON/YAML

Speed Faster Slower

Reusable Modules Yes Limited

Community Very Large Smaller

State File External (tfstate) Managed by AWS

One-line Exam Answer

Terraform = Multi-cloud & HCL

⭐ 4. Deploying & Destroying AWS with Terraform

1. Install Terraform

2. Configure AWS credentials

3. Write .tf file defining resources

4. Run:

terraform init

terraform plan

terraform apply

To Destroy

terraform destroy

Typical AWS Resources You Deploy

 EC2 instance

 VPC

 Subnets

 Security Groups

 S3 bucket

 IAM role

⭐ 5. Packer – Final Summary

What is Packer?

 Tool for building machine images (AMI, Docker Image, VM Image)

 Also from HashiCorp

 Automates creation of golden images (important term)

Why Use Packer?

 Consistent images

 Faster deployment (pre-baked software)

 Works with Terraform

 Multi-cloud support

How Packer Works — Exam Point

Packer executes:

Builders → Provisioners → Post-processors

 Use Packer to create AMI

 Use Terraform to deploy the AMI

UNIT 2
✅ UNIT 2 – Complete Revision Summary (Cloud Automation Tools & Applications)

TOPICS COVERED

1. Introduction to Virtualization Technologies

2. Load Balancing & Virtualization

3. Hypervisors

4. Porting Applications

5. VM Provisioning & Manageability

6. VM Migration (steps, types)

7. Virtual Clusters & Resource Management

8. Virtualization for Data-Center Automation

⭐ 1. Virtualization – Core Definition

Virtualization allows multiple OS instances (VMs) to run on a single physical machine by abstracting
hardware resources.

Benefits:

 Better resource utilization

 Isolation & security

 Portability

 Lower cost

 Scalability

 Easy backup, cloning, migration

⭐ 2. Load Balancing in Virtualized Environments

Load balancing ensures even distribution of workloads across VMs or physical servers.

Why important?

 Avoid overload

 Ensure QoS
  Reduce downtime

 Maintain high availability

Types:

 Static

 Dynamic

 Hardware-based

 Software-based

⭐ 3. Hypervisors

Two types:

Type 1 (Bare Metal)

 Installed directly on hardware

 Examples: VMware ESXi, Xen, Hyper-V

 Fast, secure, good for data centers

Type 2 (Hosted)

 Installed on OS

 Examples: VirtualBox, VMware Workstation

 Slower, used for testing

⭐ 4. Porting Applications

Porting = moving applications across different VM environments or physical machines.

Virtualization simplifies porting by:

 Removing hardware dependency

 Using VM templates

 Providing identical environments

⭐ 5. VM Provisioning

Provisioning = creating and configuring new VMs.

Traditional Provisioning

Took hours/days:

 Buy server
  Install OS

 Install software

 Configure hardware

Virtual Provisioning

Takes minutes:

 Use templates

 Self-service interface

 Instant OS / app deployment

⭐ 6. VM Migration

Migration = moving a running VM from one host to another.

Why?

 Load balancing

 Hardware maintenance

 Server consolidation

 Energy saving

⭐ Live Migration Steps (Very Important)

1. Start migration

2. Pre-copy memory pages

3. Suspend VM & copy last “dirty pages”

4. Transfer CPU + device states

5. Resume VM on destination

6. Redirect network → Remove source copy

Downtime is only a few milliseconds.

⭐ 7. Virtual Clusters

Virtual cluster = group of VMs acting like a cluster, independent of physical machine boundaries.

Features:

 VMs can be replicated

 Size grows/shrinks dynamically

 VM failures don’t affect host

Key issues:

 VM placement

 Memory migration

 VM image storage

 Resource scheduling

✅ 1. Hardware Virtualization

The hardware is abstracted so multiple operating systems can run on a single physical system.

How?

Using a hypervisor.

Types:

 Full Virtualization – guest OS is not modified (e.g., VMware)

 Para Virtualization – guest OS is modified to work with hypervisor (e.g., Xen)

 Hardware-assisted Virtualization – uses CPU extensions like Intel VT-x, AMD-V

Example:

VMware ESXi, Hyper-V, KVM

✅ 2. Operating System (OS) Virtualization

Multiple isolated OS environments (containers) run on the same kernel.

Key idea:

No hypervisor → same OS kernel shared.

Examples:

 Docker

 LXC

 Kubernetes containers

Use case:

Microservices, DevOps, cloud-native apps

✅ 3. Server Virtualization

One physical server is divided into multiple virtual servers.

Benefits:
  Better resource utilization

 Easy provisioning

 Isolation

Examples:

VMware vSphere, Microsoft Hyper-V

✅ 4. Storage Virtualization

Combines multiple physical storage devices into a single logical storage pool.

Types:

 Block-level

 File-level

Examples:

RAID, SAN, NAS, VMware vSAN

✅ 5. Network Virtualization

Abstracts networking hardware to create virtual networks.

Components:

 Virtual switches

 Virtual routers

 VLANs

 SDN (Software Defined Networking)

Examples:

Cisco ACI, VMware NSX

✅ 6. Application Virtualization

Applications run in isolated containers without being installed on the OS.

Examples:

 Citrix XenApp

 VMware ThinApp

 Microsoft App-V

Benefit:
Avoids installation conflicts → portable apps

✅ 7. Desktop Virtualization

User desktops run inside VMs on a central server.

Types:

 VDI (Virtual Desktop Infrastructure)

 Remote Desktop Services (RDS)

Examples:

 VMware Horizon

 Citrix Virtual Desktop

UNIT 3
✅ UNIT 3 – IMPORTANT POINTS

1. Cloud vs Data Center

 Cloud offers on-demand scalability, flexibility, and rapid provisioning.

 Data Centers require manual hardware purchase, installation, maintenance → time-

 Cloud eliminates over-provisioning and under-provisioning problems.

2. Why Scalability Matters

 Scalability ≠ growth | It means meeting changing demand.

 Demand fluctuates seasonally, weekly, hourly.

 Under-provisioning → Performance issues, errors, downtime, revenue loss.

 Over-provisioning → Wasted resources, higher cost.

3. Types of Scaling

A. Horizontal Scaling (Scale Out / Scale In)

 Add or remove instances (more servers).

 Used for load balancing, fault tolerance, redundancy.

 Cons: May require load balancers, might cost more if many instances.

B. Vertical Scaling (Scale Up / Scale Down)

 Change the instance size (bigger CPU/RAM).

 Pros: Simple, effective right-sizing saves cost.

 Cons: Usually causes downtime, expensive if not right-sized.

4. Cloud Scalability Methods

A. Manual Scaling

 Engineer changes resources manually.

 Risk: Human error, forgetting to scale down.

B. Scheduled Scaling

 Pre-defined scaling times (e.g., evenings, weekends).

 Useful when demand pattern is predictable.

C. Automatic Scaling (Auto Scaling)

 Based on metrics: CPU, memory, network utilization.

 Most efficient & cost-effective.

 Ensures availability + performance.

5. AWS / Azure Scaling Strategies

 AWS: EC2 Auto Scaling, Dynamic Scaling, Predictive Scaling.

 Azure: VM Scale Sets (VMSS).

 Understand right sizing → selecting correct instance type to avoid costs.

 Reserved Instances (RI) may limit flexibility but reduce cost.

6. Configuration Management

Unmanaged Configuration

 Manual configuration of systems.

 Error-prone, non-repeatable, inconsistent.

Managed Configuration

 Uses tools like Ansible, Puppet, Chef, SaltStack.

o Idempotency (same result every time)

o Automation

o Version control

o Consistency in environments

o Infrastructure as Code (IaC)

7. Why Configuration Management is Important

 Avoids configuration drift.

 Reduces manual effort.

 Improves reliability & repeatability.

 Supports DevOps workflows.

 Enables fast provisioning + automated patching.

UNIT 4
⭐ Types of Load Balancers in Google Cloud

Google offers two categories:

🔶 1. Application Load Balancers (Layer 7 – HTTP/HTTPS)

 Proxy-based

 URL routing, Host-based routing

 SSL offloading

 Works with serverless, Compute Engine VMs, GKE, hybrid backends

External Application LB

 Built on GFE or Envoy

 Supports:

o Global (multi-region)

o Regional

o Classic

Internal Application LB

 Built on Andromeda + Envoy

 Can be:

o Regional

o Cross-region

🔶 2. Network Load Balancers (Layer 4 – TCP/UDP/IP)

Two types:

 Proxy Network LB (Reverse Proxy)

 Passthrough Network LB (Direct Server Return)

🟣 Proxy Network Load Balancer (Layer 4 Reverse Proxy)

Features

 Terminates TCP traffic at LB

 LB forwards traffic to backend

 Supports TCP only

 Supports advanced routing & cross-region configuration

External Proxy NLB

 Built on GFE or Envoy

 Modes:

o Global

o Regional

o Classic

Internal Proxy NLB

 Only internal VPC clients can connect

 Modes:

o Regional

o Cross-region

🟢 Passthrough Network Load Balancer (Layer 4 – DSR Mode)

 Does not proxy traffic

 Traffic passes directly to VM

 Maintains client’s source IP

 Supports:

o TCP

o UDP

o ESP, GRE, ICMP, ICMPv6

External Passthrough NLB

 Built on Maglev

 Internet-facing

 Can use:

o Backend services (recommended)

o Target pools (legacy)

Internal Passthrough NLB

 Built on Andromeda

 Only internal VPC access

 Premium tier only

🔥 Key Differences – Proxy NLB vs Passthrough NLB

Feature Proxy NLB Passthrough NLB

Traffic handling Terminates TCP Direct pass-through

Layer L4 Proxy L4 Passthrough

Source IP preserved? No Yes

Backend response Goes back via LB Direct to client (DSR)

Protocols TCP only TCP, UDP, ESP, GRE, ICMP

Use case Advanced L4 control High-performance, low latency

2. Choosing a Load Balancer

Choosing a load balancer depends on:

Load Balancer Traffic Type

Application Load Balancer HTTP/HTTPS (Layer 7)

Proxy Network LB TCP (with optional TLS offload)

Passthrough Network LB TCP, UDP, ICMP, ESP, GRE (Layer 4)