INFRASTRUCTURE EXPLAINED

One foundation.
Many ways to deploy.

Start with the needs of the application, then choose the environment and operating model to support it.

Deployment models and questions to ask
Model	What it means	What to validate
Public	Provider-operated infrastructure, often shared across customers.	Regional service availability, shared responsibility, egress and commitments.
Private	Cloud capabilities dedicated to one organisation, on premises or hosted.	Capacity ownership, skills, lifecycle upgrades and resilience.
Hybrid	Coordinated use of private infrastructure and public cloud.	Identity integration, data movement, monitoring and dependency failure.
Multicloud	Use of more than one cloud provider; integration is a separate design decision.	Portability, duplicated controls, contractual boundaries and interconnect.
Sovereign / isolated	A set of location, access and operational-control requirements; isolation may be an additional constraint.	Jurisdiction, personnel, keys, audit evidence and disconnected support.

Sovereignty is not a certification implied by this table. Assess requirements with qualified advisers and provider evidence.

ARCHITECTURE IN LAYERS

Physical foundations. Operational control.

Business services, inference and analytics; performance targets begin here.

VMs, containers, scheduling and repeatable deployment.

Resource capacity, durable data and reliable connectivity.

Power, cooling, physical security and site capacity.

Identity, encryption, monitoring, recovery and cost accountability.

Conceptual layer model; operational responsibility changes by contract and deployment.

Document acceptable processing locations, backup locations and support access.

Define recovery time and recovery point objectives; test restoration instead of assuming replication is enough.

Allocate incident response, patching, identity, key ownership and cost responsibility.

Then validate end-to-end latency and model full lifecycle costs, including migration and data movement.