What Is Cloud Computing? IaaS, PaaS & SaaS

Cloud computing is the delivery of computing services - servers, storage, databases, networking, software, analytics, and intelligence - over the internet with pay-as-you-go pricing. Instead of buying, owning, and maintaining physical data centers, organizations rent access from cloud providers like AWS, Azure, and GCP. The five essential characteristics per NIST SP 800-145 are on-demand self-service, broad network access, resource pooling, rapid elasticity, and measured service. The three primary service models are IaaS (raw infrastructure), PaaS (managed platform), and SaaS (ready-to-use applications). The four deployment models are public, private, hybrid, and multi-cloud. US salary bands for cloud roles in 2026 range from $75-100K for cloud support to $260K-$400K+ for principal cloud architect.

Cloud Service Models Compared (IaaS vs PaaS vs SaaS)

Cloud Deployment Models

Definition and Essential Characteristics

The formal NIST definition of cloud computing (NIST SP 800-145, 2011) identifies five essential characteristics, three service models, and four deployment models. These are the canonical reference used by governments, enterprises, and cloud providers globally.

Five Essential Characteristics

• On-demand self-service: Users provision computing resources (server time, storage, networking) automatically without requiring human interaction with the provider.
• Broad network access: Resources are available over the network and accessed through standard mechanisms used by heterogeneous client platforms (mobile, laptop, desktop, IoT).
• Resource pooling: Provider resources are pooled to serve multiple consumers using a multi-tenant model with physical and virtual resources dynamically reassigned according to consumer demand.
• Rapid elasticity: Resources can be scaled out or in, often automatically, to match demand - from a single instance to thousands in minutes.
• Measured service: Cloud systems automatically control and optimize resource use by applying metering at a level appropriate to the service. Resource usage is monitored, reported, and billed transparently.

Three Service Models

IaaS (Infrastructure as a Service)

Raw computing infrastructure delivered as a service. The customer rents virtual machines, storage, networks, and load balancers and is responsible for the operating system, middleware, runtime, applications, and data. Use cases: lift-and-shift migrations from on-premises, custom application architectures requiring specific OS or runtime configurations, big data and HPC workloads, dev/test environments. Examples: AWS EC2, EBS, VPC; Azure Virtual Machines, Virtual Network, Blob Storage; Google Compute Engine, Persistent Disk, VPC.

PaaS (Platform as a Service)

A managed platform for developing, deploying, and running applications without managing the underlying infrastructure. The provider handles the OS, runtime, middleware, and often the database; the customer focuses on application code and data. Use cases: web application deployment, API services, microservices, mobile backends, data pipelines. Examples: Heroku, AWS Elastic Beanstalk, Google App Engine, Azure App Service, AWS App Runner, Google Cloud Run, Vercel, Netlify (for frontend). Modern PaaS increasingly overlaps with CaaS (Containers as a Service) and FaaS (serverless).

SaaS (Software as a Service)

Ready-to-use applications delivered over the internet. The provider manages everything from infrastructure to application code; the customer manages only their data, user access, and configuration. Use cases: email, collaboration, CRM, ERP, HR, finance, marketing automation, design tools, customer support, security operations. Examples: Microsoft 365, Google Workspace, Salesforce, HubSpot, Slack, Dropbox, Box, Zoom, Adobe Creative Cloud, ServiceNow, Workday, Atlassian Jira and Confluence, GitHub, GitLab, Figma, Notion, Monday.com, Asana, Zendesk, Datadog, Splunk Cloud, ServiceNow.

Four Deployment Models

Public Cloud

Infrastructure owned and operated by a third-party provider and shared across many customers. The largest category by spend (AWS, Azure, GCP, Alibaba Cloud). Benefits: no upfront capital cost, global reach, broadest service catalog, fastest time to deploy. Trade-offs: less direct control, data residency concerns, ongoing operating cost at scale can exceed on-premises equivalent.

Private Cloud

Infrastructure dedicated to a single organization, either hosted on-premises (OpenStack, VMware vSphere/VCF, Nutanix) or in a hosted private environment (AWS Outposts, Azure Stack, Google Distributed Cloud, IBM Cloud Private). Benefits: full control, data residency, predictable costs, can use existing infrastructure investments. Trade-offs: highest capex, slower to scale, requires in-house operations expertise. Used primarily by regulated industries (financial services, defense, government) and large enterprises with heavy legacy investment.

Hybrid Cloud

Combines public and private cloud with orchestration and data movement between them. The typical pattern for regulated industries: sensitive workloads on private cloud or on-premises, web-facing and elastic workloads on public cloud. Modern hybrid patterns are increasingly abstracted by Kubernetes (Anthos, Azure Arc, AWS EKS Anywhere) and by managed database services that span on-premises and cloud (Azure SQL Managed Instance, AWS RDS on Outposts, Google AlloyDB Omni).

Multi-Cloud

Using multiple public cloud providers simultaneously for redundancy, vendor diversification, or best-of-breed service selection. Common patterns: AWS for primary compute, GCP for data analytics and ML, Azure for Microsoft ecosystem integration. Multi-cloud introduces complexity in networking, identity, observability, and FinOps. Major multi-cloud management platforms include Terraform/Pulumi (infrastructure as code), Kubernetes (workload portability), and tools like Spot by NetApp, Kion, and Morpheus (cloud governance).

Core Cloud Services

Every major cloud provider offers the same broad categories of services. Understanding the categories helps navigate any provider.

Compute

Virtual machines (EC2, Azure VMs, Compute Engine), managed containers (ECS, EKS, AKS, GKE), serverless functions (Lambda, Azure Functions, Cloud Functions), and edge compute (Lambda@Edge, Cloudflare Workers). For a deeper dive into cloud engineering practices, see our What Is Cyber Security guide for the cloud security lens.

Storage

Object storage (S3, Azure Blob, Cloud Storage) for unstructured data and backups; block storage (EBS, Azure Disks, Persistent Disk) for virtual machine disks; file storage (EFS, Azure Files, Filestore) for shared file systems; archival storage (S3 Glacier, Azure Archive, Cloud Storage Archive) for long-term retention at $0.001-0.004/GB/month.

Databases

Managed relational databases (RDS, Azure SQL, Cloud SQL, Aurora, AlloyDB) for OLTP; data warehouses (Redshift, Synapse, BigQuery, Snowflake, Databricks) for analytics; NoSQL (DynamoDB, Cosmos DB, Firestore, MongoDB Atlas) for document, key-value, and graph workloads; in-memory caches (ElastiCache, Azure Cache for Redis, Memorystore); ledger databases (QLDB) for audit trails.

Networking

Virtual private clouds (VPC, VNet, VPC), load balancers (ALB, NLB, Application Gateway, Cloud Load Balancing), content delivery networks (CloudFront, Azure CDN, Cloud CDN, Cloudflare), DNS services (Route 53, Azure DNS, Cloud DNS), direct connections (Direct Connect, ExpressRoute, Cloud Interconnect), and software-defined WAN (AWS Transit Gateway, Azure Virtual WAN, Google Cloud Router).

Security and Identity

Identity and access management (IAM, Microsoft Entra ID, Cloud IAM), key management (KMS, Azure Key Vault, Cloud KMS), secrets management (Secrets Manager, Azure Key Vault Secrets, Secret Manager), web application firewall (WAF, Azure WAF, Cloud Armor), DDoS protection (Shield, Azure DDoS Protection, Cloud Armor), and security posture management (Security Hub, Defender for Cloud, Security Command Center). The shared responsibility model means cloud security requires both provider controls and customer configuration. For cloud security training, the TutorsBot Cloud Security training covers the architecture, controls, and governance that map to these roles.

AI and Machine Learning

Managed ML platforms (SageMaker, Azure ML, Vertex AI) for training, deploying, and managing ML models. Generative AI services (Bedrock, Azure OpenAI, Vertex AI with Gemini) for LLM applications. Foundation models from Anthropic, Meta, Mistral, Cohere, and OpenAI available through provider marketplaces. Specialized hardware (AWS Trainium, Google TPUs, NVIDIA H100 GPUs) for AI training at scale. For an introduction to the ML concepts behind these services, see our What Is Machine Learning guide.

Cloud Computing Costs and FinOps

Cloud cost management has become its own discipline, formalized by the FinOps Foundation. Major cost categories: compute (typically 50-65% of spend), storage (10-20%), data transfer/egress (5-15%, often the most surprising line item), managed services (10-20%), and support/licensing (5-10%).

Cost Optimization Strategies

• Right-sizing: Match instance types and sizes to workload requirements. Tools: AWS Compute Optimizer, Azure Advisor, Google Active Assist.
• Reserved instances and savings plans: 1-3 year commitments for 30-72% discount on stable workloads.
• Spot instances: Up to 90% discount for fault-tolerant workloads (batch processing, CI/CD, dev/test).
• Storage tiering: Move cold data to lower-cost storage classes (S3 Intelligent-Tiering, Azure Blob lifecycle policies, Cloud Storage Class Transitions).
• Autoscaling: Scale capacity to match demand; turn off non-production resources outside business hours.
• Egress reduction: Co-locate compute and data in the same region, use CDNs for content delivery, consider Cloudflare or other providers for cache-heavy workloads.
• Tagging and chargeback: Tag every resource with team, environment, and cost center to enable accurate chargeback and accountability.

For a structured cloud cost management career path, the FinOps Foundation offers FinOps Certified Practitioner and FinOps Certified Professional certifications. Cloud cost optimization roles typically pay $130-200K in the US in 2026.

Career Paths in Cloud Computing

Cloud Engineer

The most common entry-level role. Responsibilities: deploy and manage cloud infrastructure, write infrastructure as code (Terraform, CloudFormation, Pulumi), configure networking and security, automate operations with scripts and CI/CD. Salary bands 2026: $115-160K, senior $160-210K. Common certifications: AWS Solutions Architect Associate, Azure Administrator Associate AZ-104, GCP Associate Cloud Engineer. For a structured entry path, the TutorsBot DevOps and Cloud Engineering training covers cloud platforms, infrastructure as code, and deployment practices that map directly to these roles.

Cloud Architect

Designs cloud architectures for complex workloads. Responsibilities: lead architecture reviews, evaluate new services, design multi-region and disaster recovery architectures, establish cloud governance, mentor engineering teams. Salary bands 2026: $180-260K, principal $260-400K+. Common certifications: AWS Solutions Architect Professional, Azure Solutions Architect Expert AZ-305, GCP Professional Cloud Architect.

DevOps Engineer / Site Reliability Engineer (SRE)

Bridges development and operations. Responsibilities: CI/CD pipeline design and operation, observability (Prometheus, Grafana, Datadog, Honeycomb), incident response, on-call rotation, infrastructure as code, container orchestration (Kubernetes). Salary bands 2026: $140-200K, senior $200-280K. Common certifications: AWS DevOps Professional, CKA (Certified Kubernetes Administrator), CKAD, Terraform Associate, GitHub Actions Certified.

Cloud Security Engineer

Specializes in securing cloud workloads. Responsibilities: implement IAM and conditional access, configure security services (GuardDuty, Defender for Cloud, Security Command Center), run cloud security posture management, respond to cloud incidents, audit configurations. Salary bands 2026: $130-180K, senior $180-260K+. Common certifications: AWS Security Specialty, Azure Security Engineer Associate, CCSP, CCSK. The TutorsBot Cloud Security training covers the architecture and governance that map to these roles.

Data Engineer

Builds data pipelines and analytics platforms on cloud. Responsibilities: design ETL/ELT pipelines (AWS Glue, Azure Data Factory, Google Cloud Dataflow), manage data warehouses (Redshift, BigQuery, Snowflake, Databricks), implement streaming architectures (Kafka, Kinesis, Pub/Sub, Event Hubs), ensure data quality and lineage. Salary bands 2026: $130-180K, senior $180-250K. Common certifications: AWS Data Analytics Specialty, GCP Professional Data Engineer, Databricks Data Engineer Associate.

ML Engineer

Deploys and operates ML systems at scale. Responsibilities: build training and inference pipelines (SageMaker, Vertex AI, Azure ML), implement MLOps practices (model versioning, monitoring, A/B testing), optimize model serving for cost and latency, manage vector databases for RAG systems. Salary bands 2026: $140-200K, senior $200-300K. Common certifications: AWS ML Specialty, GCP Professional ML Engineer, Azure AI Engineer Associate.

How to Start Learning Cloud Computing

Begin with the foundations: Linux command line, networking fundamentals (TCP/IP, DNS, HTTP, load balancing, firewalls), and basic scripting (Python or Bash). Then learn one cloud provider deeply - AWS is the broadest and has the most learning resources, but Azure is dominant in enterprise and GCP is strongest in data and ML. Use the provider's free tier and free training (AWS Skill Builder, Microsoft Learn, Google Cloud Skills Boost). Earn an associate-level certification to validate the foundational knowledge, then build hands-on projects (static website hosting, three-tier web application, data pipeline, CI/CD pipeline). For specialized tracks (security, data, ML, DevOps), earn the relevant specialty or professional-level certification. For a structured, mentor-guided entry path, the TutorsBot DevOps and Cloud Engineering training consolidates the curriculum end-to-end.

For related foundational topics, see our What Is Cyber Security and What Is Machine Learning explainers.

Frequently Asked Questions

Is cloud computing the same as virtualization?

Virtualization is the technology that creates virtual versions of physical resources (a virtual machine, a virtual network). Cloud computing is a broader model built on virtualization but also including networking, automation, self-service, multi-tenancy, and pay-as-you-go pricing. A private cloud typically runs on virtualization (VMware, OpenStack, Hyper-V); a public cloud runs on massive-scale virtualization plus automation, orchestration, and self-service portals.

Is cloud computing more expensive than on-premises?

It depends on the workload and how you optimize. For new workloads, cloud is almost always cheaper because of zero upfront capital expenditure and the ability to right-size continuously. For workloads that are predictable, steady-state, and run for many years, on-premises can be cheaper if you fully utilize the hardware over its depreciation life. The break-even is typically 3-5 years for steady workloads. Hybrid is often the optimal answer: cloud for elastic and variable workloads, on-premises for steady-state core workloads.

What is serverless computing?

Serverless computing (more accurately, "serverless" or FaaS - Functions as a Service) is a cloud execution model where the cloud provider dynamically allocates compute resources on demand and the customer is billed only for the actual compute time used. The developer writes functions that respond to events (HTTP requests, queue messages, file uploads, scheduled triggers) and the provider manages all the infrastructure. AWS Lambda, Azure Functions, Google Cloud Functions, and Cloudflare Workers are the leading serverless platforms. Serverless is ideal for event-driven workloads, APIs with variable traffic, scheduled tasks, and glue code between services.

What is Kubernetes?

Kubernetes (K8s) is the open-source container orchestration platform originally developed by Google and now maintained by the Cloud Native Computing Foundation (CNCF). Kubernetes automates the deployment, scaling, management, and operations of containerized applications across clusters of machines. Major managed Kubernetes services include Amazon EKS, Azure AKS, and Google GKE. The Certified Kubernetes Administrator (CKA) and Certified Kubernetes Application Developer (CKAD) are the standard certifications. Kubernetes has become the de facto standard for container orchestration and is widely used in cloud-native application architectures.

What is the future of cloud computing?

The major trends in 2026 and beyond: generative AI and large language models driving massive new cloud spend (the major providers now compete heavily on AI infrastructure); edge computing pushing compute closer to where data is generated (Cloudflare, AWS Wavelength, Azure Edge Zones, Google Distributed Cloud Edge); quantum computing as a cloud service (AWS Bracket, Azure Quantum, Google Quantum AI); sustainability and carbon-aware computing (cloud providers committing to net-zero and offering carbon-aware workload scheduling); confidential computing protecting data in use (Intel SGX, AMD SEV, NVIDIA H100 confidential computing); and platform engineering as a discipline (internal developer platforms that abstract cloud complexity for application developers).

Resources and Next Steps

The authoritative sources listed (NIST, AWS, Microsoft Azure, Google Cloud, FinOps Foundation) are the canonical references for definitions, service catalogs, and cost management. For students and career-changers pivoting into cloud engineering, DevOps, or cloud architecture roles, the TutorsBot DevOps and Cloud Engineering training covers cloud platforms, infrastructure as code, and deployment practices end-to-end. For cloud security specialization, the TutorsBot Cloud Security training covers architecture, controls, and governance. For related foundational topics, see our What Is Cyber Security and What Is Machine Learning explainers.