Week 10 · lesson
Lesson 3: Cloud Models, Services, and Characteristics
Core path: 42 minutes
"The cloud" is not one architecture.
A technician should separate where the cloud is deployed, what service layer is delivered, and which operating characteristics the service uses.
Deployment models
Public cloud
Infrastructure is operated by a cloud provider and offered to many customers through provider services.
Private cloud
Cloud infrastructure is dedicated to one organization, whether hosted internally or by a provider.
Hybrid cloud
An organization uses both private and public cloud resources as part of one operating model.
Community cloud
Infrastructure is shared by organizations with common requirements, policies, or mission concerns.
Deployment model answers:
Who is the infrastructure intended to serve, and how is it shared?
Service models
IaaS
Infrastructure as a Service provides resources such as virtual compute, storage, and networking. The customer retains more responsibility for the guest OS, applications, configuration, and data.
PaaS
Platform as a Service provides a managed application platform/runtime. The customer focuses more on the application and data while the provider manages more of the underlying platform.
SaaS
Software as a Service provides the finished application/service for the user to consume. The customer still manages responsibilities such as identities, data, sharing, settings, and appropriate use.
Service model answers:
Which technical layers does the provider operate for me?
Shared vs dedicated resources
Cloud platforms may use pooled/shared infrastructure across customers or provide resources dedicated to one customer/workload.
Do not confuse shared infrastructure with "everyone can see everyone else's data." Isolation and authorization are still required.
Multitenancy
Multitenancy means one service/infrastructure platform supports multiple separate customers or tenants while maintaining logical separation between them.
Metered utilization
Cloud usage can be measured for billing, limits, or capacity planning.
Technicians should recognize that usage can include more than compute time. Network transfer may matter too, including ingress and egress depending on the provider/service model.
Elasticity
Elasticity is the ability to scale resources up or down as demand changes.
Scaling capability is not the same thing as infinite capacity or zero cost.
Availability
Cloud designs can use redundancy and distributed infrastructure to improve service availability.
High availability reduces some failure risks. It does not mean the service can never go down.
File synchronization
Cloud services commonly synchronize files/state across devices or locations.
Synchronization is not automatically backup. A deletion or unwanted change can sometimes synchronize too.
Guided comparison
Requirement A:
Run a custom server OS and application while controlling the guest OS.
A likely service model is IaaS.
Requirement B:
Use a hosted productivity application without administering servers.
A likely service model is SaaS.
Requirement C:
Keep sensitive internal workloads in organization-dedicated infrastructure while using provider capacity for another workload.
That may point toward a hybrid deployment model.
Student action
For eight fictional requirements classify:
deployment model: public / private / hybrid / community
service model: IaaS / PaaS / SaaS
shared or dedicated resource concern:
metered-use concern:
availability requirement:
file-sync or multitenancy concern:Not every scenario needs every characteristic. Defend only the ones that actually matter.
Evidence checkpoint
Keep these three questions separate:
- Deployment model: who shares/owns the cloud environment?
- Service model: which technical layers are delivered?
- Characteristics: how are resources consumed, scaled, shared, synchronized, and kept available?
Read it. Prove it.