Skip to content

Cloud Characteristics

For an IT environment to qualify as a cloud, it must exhibit a specific set of characteristics that enable the remote provisioning of scalable and measured resources. The greater the degree to which these characteristics are supported and utilized, the greater the value proposition of the platform.

CharacteristicCore idea
On-Demand UsageSelf-service provisioning, fully automated after configuration
Ubiquitous AccessAccessible from any device, anywhere, via standard protocols
Multitenancy & Resource PoolingShared infrastructure that keeps tenants isolated
ElasticityAutomatic, transparent scaling in response to runtime demand
Measured UsagePay only for what you use; built-in monitoring and reporting
ResiliencyRedundant implementations that failover automatically

Also called on-demand self-service usage — a cloud consumer can unilaterally provision cloud-based IT resources without requiring any human involvement from the cloud provider.

  • Self-service: the consumer configures and activates resources through a portal, CLI, or API
  • Fully automated: once configured, usage scaling and management requires no further manual interaction from either party
  • Practical result: eliminating the provisioning bottleneck — new capacity appears in minutes, not days

A cloud service must be widely accessible — not restricted to specific devices, locations, or network conditions.

To achieve this, cloud architectures must support:

  • A variety of client devices — desktops, laptops, tablets, mobile phones, IoT devices
  • Multiple transport protocols — HTTP/S, WebSocket, gRPC, and others
  • Diverse interfaces — web UIs, REST APIs, SDKs, CLI tools
  • Multiple security technologies — TLS, OAuth 2.0, API keys, mTLS

These two characteristics are closely linked and together underpin the economics of cloud computing.

Multitenancy allows a single instance of a software program or IT resource to serve multiple different consumers (tenants) simultaneously, while keeping each tenant completely isolated from and unaware of the others.

Resource pooling builds on multitenancy and virtualization to aggregate large-scale physical and virtual IT resources into a shared pool that serves many consumers at once.

PropertyDetail
Dynamic assignmentIT resources are assigned and reassigned based on live consumer demand
MechanismTypically implemented through statistical multiplexing
Consumer viewConsumers perceive dedicated resources; actual allocation is shared

Elasticity is the automated ability of a cloud to transparently scale IT resources in response to runtime conditions — or based on thresholds predetermined by the consumer or provider.

  • Automatic — no manual intervention required; the cloud monitors and responds to demand signals
  • Transparent — consumers don’t see the underlying provisioning; they just see consistent capacity
  • Proportional — scaling is tied directly to measured usage, which enables reduced upfront investments and pay-as-you-go cost models
  • Range — the breadth of elasticity available depends directly on how large the provider’s IT resource pool is; larger pools offer wider scaling ranges

Measured usage is the cloud platform’s ability to accurately track resource utilization and charge consumers only for what they actually use.

This characteristic has two dimensions:

DimensionWhat it means
BillingConsumers pay only for the specific resources consumed, or for the exact timeframe they were accessed
Monitoring & ReportingThe platform continuously records usage data, which is valuable even in private clouds that don’t charge per use

Measured usage is closely linked to on-demand usage — you can only charge for what you can measure, and you can only scale automatically if you know current utilization.


Resilient computing involves preconfiguring redundant implementations of IT resources across physically separate locations so that failure is handled automatically — without requiring manual intervention.

ScenarioHow resiliency handles it
A resource instance failsProcessing automatically transfers to a redundant implementation
An availability zone goes downFailover occurs across locations within the same cloud
A full cloud environment failsFailover can occur across multiple different clouds

The practical outcomes for consumers:

  • Increased reliability — applications tolerate individual component failures
  • Increased availability — systems remain accessible even during infrastructure events

These characteristics are not independent — they reinforce each other:

  • On-demand + elasticity → capacity is always right-sized without manual effort
  • Measured usage + on-demand → costs track actual consumption, not reserved capacity
  • Multitenancy + resource pooling → the economics work at scale
  • Ubiquitous access + resiliency → the service is available anywhere, even during failures

An IT environment that lacks any of these characteristics is not fully cloud — it may be hosted infrastructure, managed hosting, or a private data center, but it doesn’t deliver the full value proposition of cloud computing.