CLOUD COMPUTING

MODULE: 4
Resource Management: Cloud resource provisioning plan

(advance reservation, on demand plan, spot instances), various scheduling and load
balancing techniques to improve QoS parameters, Resource Optimization algorithms,
task migration and VM migration technique.

📘Cloud Resource Provisioning Plan

🔷 1. What is Resource Provisioning?
✅ Definition (English):
Resource provisioning in cloud computing is the process of allocating cloud resources (like
CPU, memory, storage, bandwidth) to applications and users when they need them, efficiently
and on demand.

✅ Definition (Hinglish):
Resource provisioning ka matlab hai cloud me resources (CPU, RAM, storage, etc.) ko
allocate karna jab user ko zarurat ho — taaki performance achhi ho aur cost bhi kam aaye.
🔧 2. What is a Resource Provisioning Plan?
✅ English:
A resource provisioning plan defines how, when, and what kind of cloud resources (virtual
machines, storage, networks) are provided to users or applications based on their needs.

✅ Hinglish:
Provisioning plan ek strategy hoti hai jisme decide hota hai ki kitne resources, kis time, aur
kaise allocate kiye jayenge, taaki system smoothly chale aur koi wastage na ho.
🧠 3. Why is Resource Provisioning Important?
●​ 💡 Ensures optimal resource utilization​
●​ 💰 Controls cost​
●​ ⚡ Improves performance and speed​
●​ 📈 Supports scalability​
●​ ✅ Ensures availability & reliability​

🧱 4. Types of Resource Provisioning

Type English Definition Hinglish Explanation

Static Provisioning Resources are allocated in Pehle se fixed resources allocate

advance, fixed amount kar diye jaate hain

Dynamic Resources are allocated on Jab zarurat ho tabhi allocate kiya

Provisioning demand jata hai

Manual Admin allocates resources Human intervention ke through

Provisioning manually resource diya jata hai

Automated Software auto-allocates based System khud decide karta hai kitna
Provisioning on rules or triggers aur kab dena hai
🌐 5. Real-Life Examples
●​ ✅ Static Provisioning Example:​
A company allocates 5 VMs permanently to a team, regardless of use.​

●​ ✅ Dynamic Provisioning Example:​

An online store adds servers automatically during Diwali sales due to high traffic.​

🧠 6. Key Components in a Provisioning Plan

Component Description

Resource Monitor CPU, memory, bandwidth usage

Monitoring

Scaling Strategy Horizontal (add VM) or Vertical (add CPU)

Auto-Scheduler Automatically turns on/off resources

User Priority Which user/app gets priority access

Budget Constraints Cost limits and billing alerts

🖼️ 7. Simple Diagram – Provisioning Flow

[ User Request ]
↓
[ Provisioning System ]
↓
[ Check Resources Available ]
↓
[ Allocate VM/Storage/CPU ]
↓
[ Monitor & Scale if needed ]
✅ 8. Advantages of Provisioning Plan
●​ ✔️ Efficient usage of cloud resources​
●​ ✔️ Helps in autoscaling during peak loads​
●​ ✔️ Cost optimization – no wastage​
 ●​ ✔️ Improves reliability and performance​
●​ ✔️ Supports multi-tenant and large-scale apps​

❌ 9. Disadvantages
●​ ❌ Complexity in dynamic provisioning setup​
●​ ❌ Wrong planning may lead to resource shortage​
●​ ❌ May need monitoring tools and automation systems​

🧠 10. Tricks to Remember

🎯 Memory Tip:​
"Provision means ‘Provide’, so Provisioning Plan = How to provide cloud resources
smartly."

📌 Focus on:​
What to give (VM, CPU) + When to give (timing) + How to give (manual/auto)

📋 11.Q: What is a resource provisioning plan in cloud

computing?​
A: A resource provisioning plan is a strategy used to allocate cloud resources such as
CPU, memory, and storage to users or applications based on demand, ensuring
performance, scalability, and cost-efficiency.

✅ Summary Chart
Term Meaning
 Resource Allocation of cloud resources
Provisioning

Static Provisioning Pre-allocated fixed resources

Dynamic Provisioning On-demand allocation

Goal Performance + Cost saving + Auto

scaling

📘 Cloud Resource Provisioning Models

(Advance Reservation, On-Demand, Spot Instances)

🔷 1. 🔒 Advance Reservation (Reserved Instances)

✅ Definition (English):
Advance Reservation means booking or reserving cloud resources (like VM, storage, etc.) in
advance for a fixed time — usually for 1 or 3 years — at a lower cost.

✅ Definition (Hinglish):
Advance reservation ka matlab hai ki aap resources pehle se book kar lete ho (jaise 1 saal ke
liye) — isme price kam hota hai, but flexibility kam hoti hai.

✅ Real-life Example:
A company reserves 10 virtual machines for a year for their permanent website — saves cost
but can't cancel easily.

🔄 2. ⚡ On-Demand Plan
✅ Definition (English):
In On-Demand, you request cloud resources whenever you need them, and you are charged
per hour or per minute with no long-term commitment.

✅ Definition (Hinglish):
On-Demand plan me aap jab chaho tab resources le sakte ho, koi advance booking nahi —
aapko use ke hisaab se payment karna hota hai.

✅ Real-life Example:
A startup uses 2 VMs during a product launch event for 3 days only — pays only for 3 days.

🕒 3. 🧩 Spot Instances
✅ Definition (English):
Spot instances are unused cloud resources offered at very low prices, but they can be taken
back anytime if demand increases. Best for temporary or non-critical jobs.

✅ Definition (Hinglish):
Spot instances wo resources hote hain jo bache hue hote hain — saste milte hain but kabhi
bhi cancel ho sakte hain. Ye bas temporary tasks ke liye best hote hain.

✅ Real-life Example:
A company uses spot instances to run data analysis at midnight when demand is low —
saves 70% cost.

🧠 4. Difference Table: (Important for Exam)

Feature Advance On-Demand Plan Spot Instances
Reservation

🔒 Commitment 1–3 years No commitment No guarantee

💰 Cost Low (discounted) High (pay-per-use) Very low

⚙️ Flexibility Low High Very low

⏳ Availability Guaranteed Always available Only when idle

⚠️ Risk None None Can be terminated

anytime

🧪 Use Case Long-term apps Short-term needs Non-critical batch jobs

🖼️ 5. Diagram: Provisioning Models Overview

+-----------------+ +------------------+ +------------------+
| Advance Reserve | | On-Demand | | Spot Instance |
+-----------------+ +------------------+ +------------------+
| | |
Book in advance Use anytime Use only if idle
(low price) (pay-per-use) (super cheap but risky)

RESOURCE PROVISIONING PROCESS IN CLOUD COMPUTING :

✅ 6. Advantages & Disadvantages

Model Advantages Disadvantages

Advance ✔️ Cost-effective for long-term use ❌ Inflexible, upfront commitment

Reserve
 On-Demand ✔️ Flexible, simple to use ❌ Expensive for long-term use
Spot Instance ✔️ Very cheap ❌ Can be interrupted anytime

🎯 7. Tips to Remember
Trick: “AOS” — Always Order Smartly

●​ A = Advance Reserve (Always booked)​

●​ O = On-Demand (Order anytime)​

●​ S = Spot Instance (Smart but risky)​

🔁 Think like booking travel:

●​ 🚂 Advance = Book train in advance for discount​

●​ 🚖 On-Demand = Call Uber when needed​

●​ ✈️ Spot = Take last-minute flight deal (cheap but can cancel)​

📝 8. Exam Answers
✅ Q: What is the difference between On-Demand and Spot Instances? (2/3
marks)

Ans: On-Demand instances provide resources whenever needed and are reliable,
but cost more.​
Spot Instances are low-cost, use spare resources, and can be terminated anytime
— suitable for temporary or non-critical jobs.

📘 Topic: Scheduling & Load Balancing Techniques to

Improve QoS Parameters
✅ 1. What is QoS?
✅ Definition (English):
QoS (Quality of Service) refers to how well the cloud service performs in terms of:

●​ Response time ⏱️​

●​ Reliability ✅​
●​ Availability 🌐​

●​ Throughput 📶​

✅ Definition (Hinglish):
QoS ka matlab hota hai cloud service ki performance ki quality — jaise response kitna fast
hai, service kitni reliable hai, aur system kitna available hai.
✅ 2. Why Scheduling & Load Balancing is Needed for
QoS?
Scheduling and load balancing help:

●​ ⚖️ Properly distribute workload​

●​ 🚀 Increase speed and response time​
●​ 💡 Reduce overload and downtime​
●​ 🔄 Improve resource utilization​

🔄 3. Scheduling Techniques in Cloud Computing

🔹 (i) Round Robin Scheduling
English:​
Each job is assigned to the next server in order, like taking turns.

Hinglish:​
Har job ko ek ke baad ek server pe assign kiya jata hai — ek turn system.

Example:​
If 3 servers (A, B, C) are there, then jobs go like A → B → C → A…

✅ Advantage: Simple and fair​

❌ Disadvantage: Ignores job size/load

🔹 (ii) First Come First Serve (FCFS)

English:​
Jobs are scheduled in the order they arrive.

Hinglish:​
Jo request pehle aayi, usko pehle process kiya jata hai.

✅ Advantage: Easy to implement​

❌ Disadvantage: Long jobs delay short ones

🔹 (iii) Priority-Based Scheduling

English:​
Jobs are given priority levels — higher priority jobs execute first.

Hinglish:​
High-priority tasks ko pehle chalaya jata hai — emergency system jaisa.

✅ Advantage: Critical jobs handled faster​

❌ Disadvantage: Low priority jobs may starve

🔹 (iv) Deadline-Based Scheduling

English:​
Jobs with the closest deadlines are scheduled first.
Hinglish:​
Jo task jaldi complete karna hai, use pehle assign karte hain.

✅ Advantage: Meets deadlines effectively​

❌ Disadvantage: Complex scheduling logic

⚖️ 4. Load Balancing Techniques

🔸 (i) Static Load Balancing
English:​
Tasks are assigned to servers at the beginning and don’t change.

Hinglish:​
Kaam starting me hi fix kar diya jata hai, badme change nahi hota.

✅ Advantage: Less overhead​

❌ Disadvantage: Not suitable for dynamic workloads

🔸 (ii) Dynamic Load Balancing

English:​
Tasks are assigned at runtime based on current load.

Hinglish:​
Real-time me dekh ke kaunsa server free hai, us par kaam diya jata hai.

✅ Advantage: Better performance and resource use​

❌ Disadvantage: Needs monitoring tools

🔸 (iii) Honeybee Foraging Algorithm

Inspired by bees searching for food — tasks are allocated based on availability and workload
just like bees share food info.

✅ Used in cloud for energy efficiency

🧠 5. Key QoS Parameters Improved by Scheduling &
Load Balancing
QoS Parameter How it improves

Response Time Tasks are scheduled efficiently

Reliability Load is spread, no single-point failure

Throughput More jobs are completed per second

Availability Load balanced = less crash chances

Resource Utilization Avoids idle or overused resources

📊 6. Diagram – Basic Load Balancing

 [Incoming Tasks]
↓
[Load Balancer System]
↙ ↓ ↘
[VM1] [VM2] [VM3]
(Low) (Idle) (Busy)
Assign to best available VM

✅ 7. Summary Table
Technique Type Purpose Use When

Round Robin Scheduling Fair task distribution Equal job sizes

Priority Scheduling Scheduling Important jobs first Critical tasks

Dynamic Load Balancing Load Balancing Real-time load shift Changing workload

Static Load Balancing Load Balancing Fixed distribution Known environment

📋 8. Exam Writing Tips

✍️ 2 Marks Sample Answer:
Q: What is Round Robin Scheduling?​
A: It is a scheduling method where each job is assigned to the next server in a circular order,
like turns. It is simple and fair but doesn't consider job size.

✍️ 5 Marks Answer Tip:

Q: Explain any two scheduling and two load balancing techniques used to improve QoS.​
Include:

●​ Definitions​

●​ Real-life examples​

●​ Advantages + 1 disadvantage​

●​ Simple diagram (optional)​

●​ QoS parameters affected​

🎯 9. Trick to Remember
🔑 Trick Word: “RFPD – SDHS”
●​ R – Round Robin​

●​ F – FCFS​

●​ P – Priority​

●​ D – Deadline​

●​ S – Static LB​

●​ D – Dynamic LB​

●​ H – Honeybee Algorithm​

●​ S – Scalability boost​
📘 Resource Optimization Algorithms
✅ 1. What is Resource Optimization?
🔹 Definition (English):
Resource Optimization in cloud computing means intelligently managing cloud resources
(CPU, memory, storage, bandwidth, etc.) so that performance is high, cost is low, and no
resources are wasted.

🔹 Definition (Hinglish):
Resource Optimization ka matlab hai ki cloud ke resources ko smartly use karna — na zyada
lagana, na kam — taaki cost bhi kam ho aur performance bhi achha mile.

✅ 2. Why Resource Optimization is Important?

✔️ Avoid resource wastage​
✔️ Improve QoS (response time, availability)​
✔️ Reduce energy and operational cost​
✔️ Handle sudden traffic/load

✅ 3. Real-life Example
Imagine a cloud-based e-commerce app like Amazon:

●​ During a sale: high traffic → allocate more resources​

●​ After sale: reduce resources​

This dynamic allocation is done using resource optimization algorithms.

✅ 4. Common Resource Optimization Algorithms

Algorithm Name Type Purpose

🔄 Genetic Algorithm Bio-inspired Find best resource combination using

evolution strategy

🧠 Ant Colony Optimization Bio-inspired Find optimal path/resource like ants do

⚖️ Load Balancing Algorithm Heuristic Balance workload among servers

🎯 Particle Swarm Optimization Bio-inspired Swarm intelligence to find best resource

use

📏 Bin Packing Algorithm Greedy Fit tasks into minimum number of VMs

🔢 Integer Linear Programming Mathematical Optimize under constraints (CPU, time)

(ILP)

🔍 Reinforcement Learning AI-based Learn & optimize over time

🔸 (i) Genetic Algorithm (GA)

🧬 Works like natural evolution:
●​ Population = Set of solutions​

●​ Best ones “survive” → combine & mutate → new generation​

Used to: Minimize cost and maximize performance.

✅ Advantage: Good for complex environments​

❌ Disadvantage: Slow convergence

🔸 (ii) Bin Packing Algorithm

🎯 Objective: Pack tasks into minimum number of VMs without overflow.
Used to: Save money & reduce idle VMs.

✅ Advantage: Efficient use of servers​

❌ Disadvantage: Not suitable for dynamic environments

🔸 (iii) Ant Colony Optimization (ACO)

🪳 Inspired by how ants find food — use “pheromones” to guide path.
Used to: Find best path or configuration to assign cloud tasks.

✅ Advantage: Decentralized and adaptive​

❌ Disadvantage: Takes time to stabilize

🔸 (iv) Particle Swarm Optimization (PSO)

🦅 Inspired by birds flocking — agents “fly” toward best solution.
Used to: Search optimal cloud resource usage.

✅ Advantage: Fast and scalable​

❌ Disadvantage: May get stuck in local optima
🔸 (v) ILP – Integer Linear Programming
🔢 Solves optimization equations under constraints (time, bandwidth, CPU, etc.)
✅ Advantage: Mathematically optimal solution​
❌ Disadvantage: Complex, slow for large systems

📈 5. Diagram – Resource Optimization Flow

[User Requests]
↓
[Monitor Resource Usage]
↓
[Optimization Algorithm]
↓
[Choose Best Resource]
↓
[Execute Task Efficiently]
✅ 6. Advantages & Disadvantages
Advantage Disadvantage

🚀 Better performance & speed 🧮 Some are computationally heavy

💰 Reduced cost 📊 Needs accurate monitoring tools
⚖️ Balanced load 🧠 Complex algorithm implementation
🔋 Less energy consumption 🐌 Slow convergence (in some cases)

🎯 7. QoS Parameters Improved

●​ ✅ Response Time ⏱️ – Faster task execution​
●​ ✅ Availability 🌐 – Prevents resource overloading​
●​ ✅ Throughput 📊 – More tasks in less time​
●​ ✅ Resource Utilization 📈 – No idle VMs or overuse​
●​ ✅ Scalability 🧩 – Easily handles sudden spikes​

🧠 8. Memory Trick
Trick Word: "GAB-API" —

●​ G = Genetic Algorithm​

●​ A = Ant Colony​

●​ B = Bin Packing​

●​ A = AI-Based (Reinforcement)​
 ●​ P = PSO​

●​ I = ILP​

✍️ 9. Exam Writing Tips

📌
Q: What is the purpose of resource optimization in cloud computing?​
A: Resource optimization ensures efficient use of cloud resources like CPU and memory to
improve QoS and reduce cost using smart algorithms like Genetic Algorithm, Bin Packing, etc.

📌 5 Marks Answer Format:

●​ Write definition in 1–2 lines​

●​ List 3 types with 1-line explanation each​

●​ Add 1 real-life example​

●​ Table/diagram optional​

●​ Close with 2–3 QoS improvements​

📘 Topic: Task Migration and VM Migration

Techniques

✅ 1. Definition
🔹 Task Migration :
Task migration means moving a specific task/process (small unit of work) from one virtual
machine (VM) or physical server to another to improve performance, reduce load, or save
energy.

🔹 Task Migration (Hinglish):

Task migration ka matlab hai kisi ek specific task ya process ko ek machine se dusri
machine me shift karna, taaki performance better ho ya load kam ho jaye.
🔹 VM Migration :
VM Migration means moving the entire virtual machine (OS + apps + data) from one physical
host to another without interrupting the service.

🔹 VM Migration (Hinglish):
VM migration ka matlab hai poori virtual machine ko (jisme OS aur apps hain) ek server se
dusre server par shift karna, bina system down kiye.

✅ 2. Why Migration is Needed?

✔️ Load Balancing ⚖️​
✔️ Energy Saving 🔋​
✔️ Hardware Maintenance 🛠️​
✔️ Fault Tolerance 🧯​
✔️ Improved QoS (speed, uptime)

✅ 3. Real-life Example
Suppose a cloud data center is hosting multiple VMs. One server is overheating due to heavy
usage. So:

●​ 👉 Either heavy tasks are migrated to another VM (Task Migration)​

●​ 👉 Or the entire VM is migrated to another physical server (VM Migration)​

✅ 4. Types of VM Migration Techniques

🔸 (i) Live Migration
●​ VM is moved without downtime​

●​ User doesn’t feel interruption​

●​ Uses memory page copying and synchronization​

Hinglish:​
Live migration me VM shift hoti hai bina system ko band kiye.

✅ Best for: Real-time apps​

❌ Slow in case of large VMs

🔸 (ii) Cold Migration

●​ VM is turned off → then moved → then turned on​

●​ Downtime is there​
Hinglish:​
Cold migration me VM ko pehle shutdown karte hain, fir migrate karte hain.

✅ Simple method​
❌ Not suitable for live services

🔸 (iii) Hot Migration

●​ Similar to live migration, but used during load balancing​

●​ Can occur even when VM is active​

✅ Less interruption​
❌ Requires strong network

✅ 5. Types of Task Migration Techniques

🔹 (i) Static Task Migration
●​ Tasks are pre-assigned to servers​

●​ No runtime movement​

✅ Less overhead​
❌ Not flexible

🔹 (ii) Dynamic Task Migration

●​ Tasks are moved at runtime, based on load or system health​

✅ Improves performance​
❌ Needs monitoring tools
✅ 6. Diagram – VM Migration vs Task Migration
[Task Migration] [VM Migration]
┌─────────────┐ ┌──────────────┐
│ Task A │ │ VM1 │
│ on VM1 │ │ (OS+Apps) │
└────┬────────┘ └─────┬────────┘
│ │
▼ ▼
Migrate Task Migrate entire VM
▼ ▼
┌─────────────┐ ┌──────────────┐
│ Task A │ │ VM1 │
│ on VM2 │ │ on Host B │
└─────────────┘ └──────────────┘

✅ 7. Advantages & Disadvantages

▶️ Task Migration
Advantages Disadvantages

Faster movement Only small tasks can be

moved

Helps in real-time load balance Complex scheduling needed

 Less network bandwidth May fail if task is I/O intensive
needed

▶️ VM Migration
Advantages Disadvantages

Full environment moves High data transfer time

Useful for maintenance May cause performance

delay

Improves fault tolerance Needs advanced

infrastructure

✅ 8. Key Differences (Table)

Feature Task Migration VM Migration

Unit of Movement Only a process/task Full VM (OS + data + apps)

Overhead Low High

Speed Faster Slower (especially live migration)

Use Case Load balancing, quick fixes Maintenance, hardware failure

Downtime Minimal or none Depends (Live = none, Cold = yes)

✅ 9. Impact on QoS
✅ Improves availability​
✅ Reduces response time​
✅ Prevents overload or failure​
✅ Maintains uptime during hardware upgrades

🧠 10. Memory Trick to Remember

"TV = Task (Small), VM (Big)"
 ●​ T = Task = Tiny move​

●​ VM = Whole Environment moves​

✍️ 11. Exam Writing Tips

📌 2 Marks Sample Question:
Q: What is VM Migration?​
A: VM Migration is the process of moving a virtual machine from one physical host to another,
either live (without downtime) or cold (with downtime), to improve performance, balance load, or
for maintenance.

📌 5 Marks Answer Format:

1.​ Write definitions of both techniques​

2.​ Mention 2 types of each​

3.​ Use a comparison table​

4.​ Optional: Draw a simple diagram​

5.​ Add real-life example + benefit​