Cloud Fundamentals

Key Characteristics of Cloud Computing

The five NIST characteristics of cloud computing, with examples, common extra traits and caveats on what a cloud does not guarantee by default.

On this page
  1. The five essential characteristics
  2. Other traits people list
  3. How the characteristics connect
  4. What the characteristics do not guarantee
  5. Common misconceptions
  6. Frequently asked questions

The National Institute of Standards and Technology (NIST) defines five essential characteristics of cloud computing in Special Publication 800-145, published in 2011. They are on-demand self-service, broad network access, resource pooling, rapid elasticity and measured service. Many other traits are often listed alongside them, but the five are the standard definition.

The five essential characteristics

On-demand self-service

You can provision computing resources, such as servers, storage or network capacity, when you need them, without asking a person at the provider. You usually do this through a web console, a command-line tool, an SDK or an API. Launching a virtual server on Amazon EC2 is a common example.

Broad network access

Cloud services are available over the network, using standard devices such as laptops, phones and tablets. The network does not have to be the public internet. Private connections, such as a dedicated link from an office, also count.

Resource pooling

The provider serves many customers from a shared pool of physical and virtual resources. Resources are assigned and reassigned dynamically based on demand. Customers generally do not control the exact physical location of their resources, though they can often choose a country or region. This is why the cloud is described as multi-tenant.

Rapid elasticity

Capacity can be provisioned and released quickly, and in some cases automatically. To the customer, capacity often appears unlimited. Elasticity is about matching resources to demand as it rises and falls. For a full explanation, see what is elasticity in cloud computing?

Measured service

The cloud system meters the resources you use, such as storage, processing, bandwidth and active accounts. Usage is monitored and reported, which gives both provider and customer a basis for billing and control. Pay-as-you-go pricing follows from this. Private clouds also meter usage, for example to charge internal departments, even when there is no per-use bill to an outside provider.

Other traits people list

Many guides add further traits. These are useful, but they are not part of NIST’s five:

  • Multi-tenancy: Many customers share the same infrastructure, with their data kept separate. NIST treats this as part of resource pooling.
  • Scalability: The ability to handle growth over time. It is related to elasticity but is not the same thing.
  • High availability and resiliency: The ability to keep running through failures. This comes from design choices, such as spreading workloads across availability zones.
  • Automation: Provisioning, scaling and recovery done by software rather than by hand.
  • Security and compliance: Controls and certifications offered by providers, although responsibility is shared with the customer.
  • Pay-per-use pricing: A common result of measured service.
  • Economies of scale: Large providers spread infrastructure costs across many customers.

Lists differ between sources, and some include five, seven or ten or more items. When a list does not match NIST’s five, it is usually because the author added their own traits.

How the characteristics connect

The five characteristics together explain why cloud computing differs from owning servers. Self-service and rapid elasticity let you change capacity quickly. Resource pooling lets providers offer that capacity at lower cost. Measured service turns usage into a bill you can track. For the business outcomes that follow, see the benefits of cloud computing.

What the characteristics do not guarantee

Having the characteristics does not make a system reliable, secure or cheap by default:

  • High availability: Uptime depends on your design. A single instance in a single zone can still fail. Providers publish service level agreements (SLAs), often in the range of 99.9% to 99.99% depending on the service and setup. Check the SLA for what you use. Zones and regions are covered in cloud regions and availability zones explained.
  • Security: Providers secure the infrastructure, but you configure access, data protection and resources. See the shared responsibility model.
  • Cost: Pay-per-use can cost less for variable workloads. A steady, always-on workload can cost more than running it yourself, so measure the total cost before you decide.

Common misconceptions

  • NIST lists scalability, security or multi-tenancy as essential characteristics. It does not. Those appear in other lists.
  • Elasticity and scalability are the same. Elasticity is automatic matching to demand. Scalability is the ability to grow.
  • Broad network access means the public internet only. Private connections also count.
  • Measured service always means pay-per-use. Private clouds may meter usage for internal charging rather than for a bill.
  • High availability comes with the cloud. It has to be designed in.

Frequently asked questions

What are the 5 characteristics of cloud computing?

They are on-demand self-service, broad network access, resource pooling, rapid elasticity and measured service, as defined by NIST in SP 800-145.

Why do different sources list different characteristics?

Some guides add traits such as multi-tenancy, security and automation, which are useful but not part of NIST’s five. Others list them to match a particular product or course.

Which characteristic saves the most money?

It depends on the workload. Measured service and rapid elasticity together can reduce spending on idle capacity, but only if you configure scaling limits and monitor usage. For a broader view of cloud advantages, see the 12 benefits of cloud computing.

The complete guide to cloud fundamentals

How Cloud Computing Works (With Diagram)
Read the guide

Keep exploring

Guides from other topics

All topics →