Introduction, Providers & Pricing
Friday, 6th March 2026 | 10:00 AM - 12:00 PM IST
Definition, characteristics, and how cloud works
Public, Private, Hybrid, Multi-Cloud strategies
IaaS, PaaS, SaaS - Deep dive with real-world examples
AWS, Azure, GCP - Services, market share, use cases
Evolution, benefits, economics, and real-world success stories
CapEx vs OpEx, pricing strategies, cost optimization
The Foundation of Modern IT Infrastructure
NIST Definition: Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources that can be rapidly provisioned and released with minimal management effort or service provider interaction.
Provision resources instantly without IT involvement. Spin up servers, databases, and services with a few clicks.
Example: Launch an EC2 instance in 2 minutes
Scale resources up or down automatically based on demand. Handle traffic spikes without over-provisioning.
Example: Auto-scale from 2 to 200 servers
Pay only for what you use. No upfront capital investment. Convert fixed costs to variable costs.
Example: $0.023/hour for compute
Access resources from anywhere, anytime. Deploy globally across multiple regions for low latency.
Example: 99.99% uptime SLA
Cloud architecture delivers computing services over the internet through a network of data centers
Public, Private, Hybrid, Multi-Cloud
Description: Services delivered over the internet by third-party providers and shared across multiple organizations.
Providers: AWS, Azure, GCP, Oracle Cloud, IBM Cloud
Best for: Web apps, development/testing, general workloads, startups
Considerations: Shared infrastructure, limited customization, data residency concerns
Description: Dedicated cloud infrastructure for a single organization, either on-premises or hosted by a third party.
Providers: VMware, OpenStack, Azure Stack, AWS Outposts
Best for: Regulated industries, sensitive data, compliance requirements, full control needs
Considerations: Higher cost, requires expertise, limited elasticity
Description: Combines public and private clouds with orchestration between them for data and application portability.
Tools: Azure Arc, AWS Outposts, Google Anthos
Best for: Gradual cloud migration, burst capacity, legacy system integration
Considerations: Complexity, networking challenges, requires expertise
Description: Using multiple cloud providers (e.g., AWS + Azure + GCP) to meet different requirements.
Why Multi-Cloud: Avoid vendor lock-in, best-of-breed services, geographic requirements, redundancy
Tools: Kubernetes, Terraform, CloudHealth, Spot.io
Considerations: Management complexity, skill requirements, cost visibility
Industry Stat: 90% of enterprises use a multi-cloud strategy. Average organization uses 2.6 public clouds and 2.7 private clouds.
IaaS, PaaS, SaaS - Understanding the Stack
The shared responsibility model defines what you manage vs. what the provider manages
Key Insight: "Generally in life, the less we are responsible for that's not key to our purpose, the better"
IaaS provides virtualized computing resources over the internet. You get access to fundamental infrastructure like virtual machines, storage, and networking.
Best for: Teams that need maximum control over their environment, have specific OS/runtime requirements, or are migrating existing applications.
| Provider | Compute | Storage | Database | Networking |
|---|---|---|---|---|
| AWS | EC2 | EBS, S3 | RDS, DynamoDB | VPC, CloudFront |
| Azure | Virtual Machines | Blob Storage | SQL Database | Virtual Network |
| GCP | Compute Engine | Cloud Storage | Cloud SQL | VPC |
Host websites and web applications with full server control
Create dev/test environments that mirror production
Process large datasets with high-performance compute
Store backups and enable disaster recovery
PaaS provides a platform allowing customers to develop, run, and manage applications without the complexity of building and maintaining infrastructure.
Best for: Developers who want to focus on coding, not infrastructure. Ideal for web apps, APIs, and microservices.
| Provider | App Platform | Container Service | Serverless |
|---|---|---|---|
| AWS | Elastic Beanstalk | ECS, EKS | Lambda |
| Azure | App Service | AKS | Functions |
| GCP | App Engine | GKE | Cloud Functions |
Build and deploy RESTful APIs quickly
Serverless backends for mobile apps
Full-stack web app hosting
Containerized microservice architectures
SaaS delivers software applications over the internet, on demand, and typically on a subscription basis. Everything is managed by the provider.
Best for: Organizations that need ready-to-use software without IT overhead. Common for email, CRM, collaboration, and productivity.
Customer Relationship Management (CRM)
Enterprise SoftwareProductivity suite - Office, Teams, OneDrive
ProductivityGmail, Drive, Docs, Meet, Calendar
CollaborationTeam communication and collaboration
CommunicationProject management and documentation
Project MgmtVideo conferencing and webinars
CommunicationWho manages what? The responsibility shifts based on service model.
Important: Identity, accounts, and devices are ALWAYS the customer's responsibility - even in SaaS!
AWS, Azure, GCP - Market Leaders
AWS Strengths: Most mature platform, largest ecosystem, extensive documentation, global presence (30+ regions), marketplace with 10,000+ solutions.
Azure Strengths: Best for Microsoft enterprises, hybrid cloud (Azure Arc), Office 365 integration, Active Directory, strong compliance certifications.
GCP Strengths: Best-in-class data analytics (BigQuery), AI/ML leadership, Kubernetes originator, premium network, competitive pricing with sustained use discounts.
Real-World Cloud Success Stories
Pizza restaurants face extreme demand variations - 50% higher than Friday nights during Super Bowl, but only once per year!
Build for peak capacity (14 servers). Pay for maximum 24/7/365. Waste capacity 364 days/year.
Scale on demand (3→14→3 servers). Pay only when scaling up. Match capacity to actual demand.
Factories used to run their own generators to power their operations. High capital cost, maintenance burden, specialized expertise required.
Utility companies provide power cheaper and more reliably. Pay for what you use. Focus on your core business, not power generation.
The Shift: "Compute is going the exact same way as power utilities. With scales of economy, it's very unlikely you can run at the same efficiency as these Mega Cloud operators."
Pizza restaurants, retail holidays, tax season - scale up during known busy periods
News sites, viral products, unexpected traffic - auto-scale based on demand
No upfront investment. Scale with success. Fail fast without sunk costs.
Create/delete environments frequently. Pay only when running.
Keep DR ready but not running. Only pay during failover tests or actual events.
AI/ML workloads, big data processing - use specialized hardware temporarily
Streams 1+ billion hours/week on AWS. Uses auto-scaling to handle peak loads, CDN for global delivery.
AWS | Auto-scaling | CDNUses Azure for vaccine research, HPC for drug discovery, compliant data storage.
Azure | HPC | ComplianceAll-in on AWS. Improved security posture, faster innovation, reduced data center costs.
AWS | Security | MigrationGCP handles 500M+ downloads. Real-time location data, massive database writes/second.
GCP | Real-time | ScaleBuilt on AWS from day one. Scaled to 150M+ users without owning servers.
AWS | Startup | ScaleUses GCP for music recommendations, BigQuery for analytics, ML for personalization.
GCP | ML | AnalyticsKey Insight: "At the peak of inflated expectations, all you care about is FEATURES. At the plateau of productivity, you focus on INTEGRATION with your existing environment."
"This applies to Cloud, AI, marriage, kids... everything!"
Only protects against planned scenarios
Protects against any single point of failure
Design Principle: "Don't rely on single instances - design for multiple instances distributed across availability zones from the start."
Reality Check: Hybrid will be the norm for most organizations. Some workloads stay on-prem due to latency requirements or legacy dependencies.
CapEx, OpEx & Cloud Cost Management
How it works: Pay per second/hour with no long-term commitment. Highest flexibility, highest cost.
Best for: Development, testing, uncertain workloads, short-term projects
Example: t3.micro EC2 instance = $0.0104/hour (~$7.50/month)
How it works: Commit to 1 or 3 years for significant discounts. All upfront, partial upfront, or no upfront payment options.
Savings: Up to 75% compared to On-Demand
Best for: Stable production workloads with predictable usage
How it works: Bid on unused capacity. Can be terminated with short notice (2-minute warning).
Savings: Up to 90% compared to On-Demand
Best for: Batch jobs, data processing, CI/CD, fault-tolerant workloads
How it works: Pay only when code executes. No charges for idle time. Automatic scaling.
Example: AWS Lambda = $0.20 per 1M requests + $0.0000166667 per GB-second
Best for: Event-driven apps, APIs, data processing, low/variable traffic
Use monitoring data to select the optimal instance types and sizes for your workloads.
Potential savings: 20-40%
Automatically adjust capacity based on demand. Scale down during off-peak hours.
Potential savings: 30-50%
Use Reserved Instances or Savings Plans for predictable workloads.
Potential savings: 40-75%
Leverage spot for fault-tolerant, flexible workloads.
Potential savings: 70-90%
Estimate costs before you build
Monitor and optimize spending
Let's estimate the monthly cost of a typical 3-tier web application:
Pro Tip: Set up budgets immediately after creating a cloud account. Unexpected costs are the #1 surprise for new cloud users!
On-demand, scalable, pay-as-you-go computing services delivered over the internet
IaaS (infrastructure), PaaS (platform), SaaS (software) - choose based on control vs convenience
AWS (33%), Azure (23%), GCP (11%) - each with unique strengths and service offerings
On-Demand, Reserved, Spot, Serverless - optimize costs with the right mix
Next Steps: Try the free tiers of AWS, Azure, and GCP. Use the pricing calculators. Set up budget alerts before deploying anything!
Recommended Learning Paths:
Questions & Discussion
Session 2: Thursday, 12th March 2026
Operational Reliability, Monitoring, Security & Migration