Some cloud terms, like cloudstorming, cloudware and external cloud, are declining in popularity. Other terms are up-and-coming, like vertical cloud.

This list gives all the latest lingo to keep you up-to-date on the most popular terms for all things cloud:

Bare metal IaaS

Physical machine rental rather than virtual machine rental. Like dedicated-server hosting with added cloud attributes of self-service, quick scale-out and usage billing.

CDN (Cloud Distribution Network)

a worldwide collection of servers that copy content for end users, to speed up delivery. Useful for sending resources to millions of clients in milliseconds.

Cloud broker

Any agent between a client and a cloud service provider. A fuzzy term that applies to people manually connecting customers to suppliers, to applications automating the use of many APIs and to financial commodity traders who trade in cloud services.

Cloud management platform

A set of tools for creating a cloud service. Bringing cloud benefits of on-demand resources, consumption metering and self-service to the computing world requires automation of the whole technology stack. OpenStack, CloudStack and vCloud are cloud management platforms.

Cloud management system

An administration system for cloud resources. A collection of tools that can be dedicated to one provider only, such as the AWS management console, or work across many public and private clouds, such as the Rightscale multi-cloud platform.

Cloud marketplace

A financial market where cloud brokers trade cloud services as commodities. A cloud marketplace can also be a store, such as a list of IaaS providers or a library of business applications ready to be deployed.

Cloud encryption key

A huge random number. The number acts as a password in a cryptography system used to protect objects in cloud storage. A cloud encryption key is kept with its owner, not stored in the cloud.

Cloud operating system

A simple OS for end users, designed to act as a front end to many cloud services such as document editors, e-mail and file storage. The latest idea in a series including network OS (a system that is useless without a network connection), web OS (everything through a web browser) and Internet OS (Internet services do the heavy lifting). The phrase “cloud OS” is also used by every OS manufacturer promoting their product for cloud use.

Cloud application

Software built from distributed components that can cope with global reach, Internet connections, component failure, service attacks and massive scaling.

Cloud Oriented Architecture

The design of cloud applications, using SOA (Service Oriented Architecture) concepts and cloud provider services. Uses distributed computing items such as middleware.

Cloud portability

The ease of moving business applications between cloud providers. Provider APIs are not usually compatible but some cloud management systems provide portability. Not to be confused with interoperability (making cloud services play together nicely).

Cloud provider

An organization that offers computing services to Internet customers, run on multi-tenant virtualized infrastructure. A provider manages the infrastructure required to run cloud services. Also known as a cloud service provider. Not the same as an ISP (Internet Service Provider), which connects its customers to the Internet.

Cloud service

A computing service offered by a cloud provider to its customers. A service may be a low-level infrastructure component or a complete business application. A cloud service is rented to tenants, massively scalable and configured using customer self-service interfaces.

Cloud storage

A data file service offered by most cloud providers. Files are managed using web services.


A version of IT outsourcing, where the outside supplier a customer chooses is a cloud provider.


The marketing of any IT service as a cloud service to new cloud converts with a shaky understanding of cloud attributes. Similar to greenwashing for environmental marketing, pinkwashing for breast cancer marketing and whitewashing for all other marketing.


a collection of machines that work together to deliver a customer service. Cloud clusters grow and shrink on-demand. A cloud service provides an API for scaling out a cluster, by adding more machines.

Consumption-based pricing model

The method used by cloud providers to charge for their services. Customers are billed for resources they consumed in very small increments, rather than paying a flat rate per month. A flat rate fee is a subscription-based pricing model, designed to guarantee customer loyalty. Also known as Pay As You Go and Utility Billing.

One of the central functions of public cloud computing, along with customer self-service, multi-tenancy and on-demand service.

Customer self-service

Cloud service selection using a web browser-based control panel, an API or command line tools. A customer can create an account, add, manage and delete services, and cancel on-demand.

Database as a Service (DBaaS)

A managed database service running on cloud infrastructure. A DBaaS customer controls their data and can scale on demand, but does not have to administer any database application.

Disaster Recovery as a Service (DRaaS)

A managed business continuity service, which reproduces a customer’s business critical applications on cloud infrastructure. A replacement for the traditional enterprise combination of one production data center and one backup data center.

Distributed computing

A technical design and development area which makes customer services work in the cloud. Distributed computing tackles many problems: sharing hardware-level storage, application-level client sessions and other resources of many machines; duplicating applications across a cluster; coping with component failure and managing many machines.

Elastic computing

the ability to scale resources to meet requirements, such as doubling the size of a customer service in the day and halving it at night. CPU can be added to a virtual machine, machines can be removed from a cluster and storage can grow with no practical limit (beyond requirements of the customer). Scaling to the required size – not too much and not too little – and scaling in time to meet requirements can be difficult.

HaaS (Hardware as a Service)

A label for renting virtualized resources in the pre-AWS era (now replaced by IaaS). It was a complement to Saesforce’s SaaS (Software as a Service) concept.

Hosted application

Business software running on a service provider’s infrastructure, rather than at the organization’s premises. A hosted application is accessed over the Internet and usually has a web interface.

Hybrid cloud

A combination of clouds, such as an organization’s private cloud running on-premise, used for confidential processing, and an off-premise public cloud, used for batch processing. A hybrid cloud is administered by a cloud management system that can provide interoperability. No standards exist yet for cloud interoperability.

Infrastructure as a Service (IaaS)

Data center resources offered as cloud services. IaaS removes the overhead of dealing with premises, hardware, networking and some infrastructure software -the cloud provider administers these systems so the customer does not have to. A cloud customer uses IaaS to build business services – IaaS provides components, not finished products. NIST call IaaS, Paas and SaaS “delivery methods”.


Integration software used in distributed computing. Getting enterprise applications to play well together is easier using an ESB, ETL system or application server. Outside the SOA and cloud world, middleware is any software which does the job of connecting components.


The sharing of infrastructure between many customers. Multi-tenancy is central to public cloud services, as multi-user is to Linux and multi-site is to web hosting. A private cloud removes security worries that other tenants are hostile strangers.


A type of database system that can be massively scaled out for cloud use. The structure required by a traditional RDBMS caused problems for distributed computing when updating records, joining tables and answering millions of queries.

Object storage

A simple service that stores documents that can be accessed with keys. Cloud object storage holds vast amounts of data and the web service front-end is easy to integrate into cloud applications. Usually cheaper and slower than the file storage used by virtual machines.

SDN (Software Defined Networking)

A data network that is automatically provisioned when a customer orders a cloud application. A recent addition to the automated provision of machines, storage and applications.

SDS (Software Defined Storage)

An attempt to apply the SDN concept to data storage. A data store that is automatically provisioned when a customer orders a cloud application.

On-demand service

A service that allows a customer to create, change and destroy resources at any time, using a self-service interface, with little or no delay.


A project backed by many cloud providers to create a cloud management platform. The collection of applications produced by the OpenStack project is also called OpenStack.

PaaS (Platform as a Service)

A type of cloud service aimed at giving developers everything they need to code applications. Like IaaS, PaaS removes the administrative overhead of dealing with lower technology stack layers and also removes some development overhead. Collections of software tools and libraries are installed, configured, integrated and bundled into self-service options.

Physical machine

A traditional computer. Can be rented by a bare-metal IaaS provider. A cluster of physical machines can be used to run a cloud management system.

Private cloud – cloud services for the use of one organization only. The organization may run its own cloud management platform on-premise or may rent off-premise dedicated services from a cloud provider. NIST call public cloud and private cloud “deployment models”.

Public cloud – Distributed computing services available over the Internet. Comes with scalability, a consumption-based pricing model, customer self-service, multi-tenancy and on-demand service.

Resource pool – The collection of behind-the-scenes computing resources a cloud provider can rent out. Resources from infrastructure pools of CPU, storage, DNS names and IP addresses are constantly allocated to customers and returned to their pools. Similar to a car pool and nothing like eight-ball pool.


Business software running in a public cloud, available for immediate use. Examples are, Google Docs and Basecamp. A SaaS provider is an ASP (Application Service Provider).

Scale out

A way of coping with more customer demand by adding more machines to a cluster. Machines may be added in other data centers to be closer to where the demand is. An application being scaled out must be able to run on a distributed computing platform. Also known as horizontal scaling. Scaling out and scaling up are methods of elastic computing. A web service is scaled out.

Scale up

Adding CPU, memory or storage resources to one machine. Virtual machines can be scaled up on-demand, as opposed to traditional on-premise physical machines. An RDBMS is scaled up.

SDN (Software Defined Networking)

A new approach to maintaining network machines. Network control is managed remotely. Networks can be added, changed and removed on-demand, along with other cloud infrastructure. Making SDN work requires more layers of complexity.

Cloud service migration

Moving a business application between different cloud providers. The challenge of moving between providers requires cloud portability work. Also covers moving a business application from an organization’s computer room to cloud services – the challenge of moving to cloud in the first place requires distributed computing work.

SLA (Service Level Agreement)

A contract which specifies the minimum a customer will get for his money. The basic measurement is service availability – the EC2 SLA from AWS offers some money back, under certain conditions, if the service fails to stay up at least 99.95% of the time.

Unstructured data

The data found in a NoSQL database and cloud storage. An organization can capture all unstructured data it produces and put it in cheap cloud storage for later analysis, rather than analyzing and reducing data first then storing the result. Unstructured data is simpler than the structured data found in an RDBMS.

Utility computing

Any rental of computing power, including cloud computing. The term was born when mainframe computing power was rented out to customers using a very expensive consumption-based pricing model, similar to the way public utilities rented out electricity, gas and telephone services.

Vertical cloud

Specialized cloud services tailored to a particular industry segment, such as finance, education or health care. Current cloud services are horizontal – generalized resources that can be used by any kind of organization. Similar to SaaS, but not limited to ready-made business services.

Virtual machine

A computing node in a cloud, offered by every IaaS cloud provider. A virtual machine is a software-defined computer that can be created, destroyed and (by some providers) scaled up on-demand.

VPC (Virtual Private Cloud)

A lightweight alternative to a private cloud. A cloud provider supplies a secure network to a customer, who can then place public cloud resources within the secure network.


A quantity of data fed to these number crunching overgrown calculators. An organization can evaluate performance from different providers by setting up identical applications and workloads on each virtual machine.