Cloud Service Models
- Cloud computing delivers IT resources over the internet using three primary service models that define what the provider manages versus what the customer controls
- Each model shifts responsibility boundaries - understanding these boundaries is crucial for network engineers designing hybrid infrastructures
- Service models follow a stack approach: IaaS provides the foundation, PaaS adds development tools, SaaS delivers complete applications
Infrastructure as a Service (IaaS)
- Provider manages: Physical hardware, virtualization layer, networking infrastructure, storage systems
- Customer manages: Operating systems, middleware, runtime environments, applications, data
- Most flexible model - gives network engineers full control over virtual network configuration
- Examples: Amazon EC2, Microsoft Azure VMs, Google Compute Engine
- Use cases: When organizations need custom network topologies, specific OS configurations, or legacy application support
- Network considerations: Customer configures VPCs (Virtual Private Clouds), subnets, routing tables, security groups, and VPN connections
Platform as a Service (PaaS)
- Provider manages: Everything in IaaS plus operating systems, middleware, runtime environments
- Customer manages: Applications and data only
- Development-focused model - abstracts infrastructure complexity while providing development frameworks
- Examples: Google App Engine, Microsoft Azure App Service, Heroku
- Use cases: Rapid application development, microservices architectures, when teams lack infrastructure expertise
- Network considerations: Limited network customization - provider handles load balancing, auto-scaling, and basic security
Software as a Service (SaaS)
- Provider manages: Complete technology stack from hardware to application
- Customer manages: User access, data input, and configuration settings within the application
- Consumption model - users access fully-functional applications via web browsers or APIs
- Examples: Office 365, Salesforce, Google Workspace, Cisco Webex
- Use cases: Standard business applications, when minimal IT overhead is desired
- Network considerations: Focus shifts to bandwidth planning, Quality of Service (QoS), and secure internet connectivity
Service Model Comparison
| Model | Customer Control | Provider Responsibility | Network Flexibility | Use Case |
|---|---|---|---|---|
| IaaS | High | Hardware + Virtualization | Full (VPCs, routing, firewalls) | Custom environments, legacy apps |
| PaaS | Medium | Infrastructure + Platform | Limited (managed load balancing) | Application development |
| SaaS | Low | Everything except data | Minimal (connectivity only) | Standard business applications |
Vocabulary
- Multi-tenancy: Multiple customers share the same infrastructure while maintaining data isolation
- Elasticity: Automatic scaling of resources based on demand (differs from scalability which is manual)
- API (Application Programming Interface): Programmatic interface for managing cloud resources
- VPC (Virtual Private Cloud): Isolated virtual network environment within public cloud infrastructure
- Hybrid Cloud: Combination of on-premises infrastructure with public cloud services
Notes
- Service models are not mutually exclusive - organizations typically use multiple models simultaneously (e.g., IaaS for custom apps, SaaS for email)
- Network latency becomes critical in cloud architectures - consider proximity of cloud regions to users when designing solutions
- Security responsibility follows the service model: IaaS requires more customer security configuration, SaaS shifts most security to provider
- Cloud networking often uses SDN (Software-Defined Networking) principles - traditional networking concepts apply but implementation differs
- For CCNA context: Focus on how traditional networking concepts (VLANs, routing, NAT) translate to cloud virtual networking
- Cost implications: IaaS offers most control but requires most management overhead; SaaS has predictable costs but less customization
- Always verify which networking features are available - cloud providers may limit certain protocols or configurations (e.g., broadcast traffic, custom routing protocols)