Cloud Delivery Models

Cloud delivery models are specific, prepackaged combinations of IT resources that a cloud provider offers to consumers. The three foundational models differ primarily in two dimensions: how much administrative control the consumer has, and how much infrastructure responsibility the provider carries. Everything above IaaS builds on it; everything above PaaS builds on that.

Model	Consumer control	Provider responsibility
IaaS	Full — raw infrastructure, OS up	Physical hardware + network layer
PaaS	Moderate — platform and app layer	Infrastructure + middleware + runtime
SaaS	Minimal — UI and usage configuration only	Everything

Infrastructure as a Service (IaaS)

A self-contained IT environment comprised of infrastructure-centric “raw” IT resources — virtualized hardware, networks, connectivity, and operating systems.

The central resource is the virtual server, leased by specifying processor capacity, memory, and local storage
Resources are not preconfigured — the consumer takes on full responsibility for software setup, configuration, and monitoring
Grants full administrative control over the virtualized infrastructure

Provider responsibilities: Provision and manage physical processing, storage, and networking hardware; monitor consumer usage for billing.

Consumer responsibilities: Everything above the hypervisor — OS installation, middleware, application deployment, network rules, security policies, scaling configuration.

Working with IaaS: Consumer View

Accessing virtual servers

Server type	Access method
Windows	Remote Desktop client (GUI-based)
Linux / Mac	SSH client (text-based shell)

Managing cloud storage

Access type	Protocol	How it works
File-based	NFS, CIFS	Files organized like standard OS folders
Block-based	SAN, iSCSI, Fibre Channel	Formats scattered data into cohesive files for network optimization
Object-based	Web API (e.g., S3)	Accessed via a web interface; not integrated into the OS directly

Consumer control capabilities

Manage virtual IT resource lifecycles — power on, restart, shut down
Configure load balancers and automated scaling triggers
Define network access rules, firewalls, and logical network perimeters
Manage IAM credentials and virtual server image storage (backups, imports, exports)
Select hardware specs (CPU, RAM, storage), high-availability options, and geographic regions
Track costs and monitor SLA metrics

Building IaaS: Provider View

Virtual server images: Providers pre-assemble environments with predefined configurations; snapshots capture current state for vertical scaling and data replication
Data centers: Multiple geographically diverse facilities increase resiliency, enable load balancing, reduce latency, and satisfy regulatory data residency requirements
Scalability: A Virtual Infrastructure Manager (VIM) automates dynamic vertical scaling; load balancers manage horizontal scaling. Scaling can be manual (consumer-driven via portal) or automatic (via an automated scaling listener)
Monitoring: Tracks virtual server lifecycles (uptime/billing), storage allocation, network traffic (QoS and billing), failure conditions (SLA compliance), and event triggers (regulatory audit)
Security: Encryption, digital signatures, IAM/SSO mechanisms, cloud-based security groups isolated via hypervisors, and hardened virtual server images

Platform as a Service (PaaS)

A predefined “ready-to-use” environment of already deployed and configured IT resources, tools, and frameworks specifically designed to support the application delivery lifecycle.

Built on top of an IaaS layer — the consumer is abstracted from bare infrastructure
Grants moderate administrative control — focused on the application layer, not the infrastructure layer
Common uses: extending on-premises environments into the cloud, fully replacing on-premises infrastructure, or building cloud services to offer to other consumers

Provider responsibilities: Preconfigure the platform, provision all underlying infrastructure and middleware, monitor usage.

Consumer responsibilities: Developing, testing, deploying, and managing cloud-based applications.

Working with PaaS: Consumer View

Development: IDEs outfitted with libraries, frameworks, APIs, and tools that emulate the cloud deployment environment. Developers can write, test, and run code locally before deploying to ready-made environments. Both SQL and NoSQL database structures are supported for development data storage
Control capabilities
- Select software platforms, development frameworks, and instance types (front-end/back-end)
- Control application lifecycle — deploy, start, stop, version
- Configure scalability — active instance thresholds, usage quotas, load balancers
- Manage IAM credentials and security settings (accessible network ports)
- Schedule resource availability — activate on request, shut down after inactivity to save costs

Building PaaS: Provider View

Development tools: Custom or plugin-enabled IDEs that simulate the cloud runtime locally, including executable servers and simulated security restrictions
Resource management: Custom resource management systems for consumers to control virtual server images and configure multitenancy
Scalability: Relies on automated scaling listeners and load balancers evaluating network traffic against workloads
Reliability: Maintained through failover systems and automated service relocation — outages are shielded from consumers
Monitoring: Tracks ready-made environment instance usage (time-based), data persistence (object count + storage size), network usage, failure conditions, and event triggers
Security: Primarily inherits security mechanisms established in the underlying IaaS environment

Software as a Service (SaaS)

A software program positioned as a shared cloud service and offered as a product or generic utility to a wide range of consumers.

Grants very limited administrative control — restricted to usage configuration and front-end UI access
SaaS architectures are predominantly built on multitenant environments

Provider responsibilities: Implement, manage, maintain, provision, and monitor the cloud service.

Consumer responsibilities: Usage-level configuration only — no infrastructure or platform concerns.

Working with SaaS: Consumer View

Integration: SaaS services typically expose refined APIs, making them easy to incorporate into larger distributed solutions
Data awareness: Free third-party SaaS products often include background programs that collect usage data — understand the privacy tradeoffs before adopting
Control capabilities
- Manage user accounts, profiles, and access authorizations
- Configure select security, availability, and reliability settings
- Set manual and automated scalability limits
- Monitor SLAs and track usage costs

Building SaaS: Provider View

Architecture: SaaS implementations require highly specialized architectures driven by distinct business logic. Implementation mediums include mobile applications, REST services, and web services
Key patterns used: Service Load Balancing, Dynamic Failure Detection and Recovery, Cloud Balancing (for extreme concurrent usage), Dynamic Scalability (vertical and horizontal auto-scaling)
Monitoring: Tracks Tenant Subscription Periods (time-based billing), Application Usage (per user or security group), and Tenant Application Functional Module (function-based billing for service tiers — e.g., free vs. paid)
Security: Inherits deployment environment security, then layers on additional mechanisms required by specific business processing logic

Combining Delivery Models

The three models form a natural provisioning hierarchy and can be stacked:

Combination	What it enables
IaaS + PaaS	A PaaS environment running on top of IaaS. The PaaS provider may lease the underlying IaaS from a different provider for economic, capacity, or legal data residency reasons
IaaS + PaaS + SaaS	An organization uses PaaS (on IaaS) to develop and deploy a SaaS service they then offer to external consumers — becoming a cloud provider themselves

Delivery Submodels

Many “as a Service” specializations exist that map back to the three foundational models:

Submodel	Base model	What it delivers
Storage as a Service	IaaS	File, object, or long-term archive storage devices
Database as a Service	PaaS	Managed database access within a ready-made environment
Desktop as a Service	IaaS	Virtual desktop environments hosted in the cloud
Security as a Service	SaaS	Security features (identity, threat detection, compliance) as a shared service
Cloud-Native Delivery	SaaS	Applications built and deployed in lightweight containers that scale dynamically
Communication as a Service	SaaS	Hosted telephony, messaging, and conferencing
Integration as a Service	PaaS	Middleware and API integration platforms
Testing as a Service	SaaS	On-demand test infrastructure and automation frameworks
Process as a Service	SaaS	Business process automation delivered as a managed service