Cloud Computing – Clarification of Key Terms

As an end-user, the range of terms in cloud computing can be confusing. Here is a brief introduction to explain all the key terms in one place.

Cloud Computing covers a multitude of differing offerings. The Cloud Industry Forum defines it as covering any remotely hosted IT service of any  type, including multi-tenanted systems accessed via the internet. This is typically using a web browser on the user's computer or mobile device.

Cloud services are provided by a "provider", or a set of providers working together. Typically a software author will work with one or more hosting companies that "host" and run the system for them.

Cloud services are similar to traditional outsourcing, both legally and in practice. Use of cloud services, including  the relationship with the provider(s), needs to be approached in a similar manner.

You’ll also see cloud computing referred to as “On-Demand”, to distinguish it from the traditional way of running computing “On-Premise”. On-Demand is a marketing term that implies more flexibility than is really available, but refers to the relative speed of adding additional capacity, and the speed of selecting (if not implementing) cloud services compared to on-premise.

There are two ways of sub-dividing the market, which combine into potentially 12 forms of cloud computing. The “Public SaaS” combination is of most relevance to most businesses, and tends to be the focus of the cloud computing articles in this blog:

(1) By Type of Service

“SaaS (Software as Service)”: Packaged software for a specific application ("app"), such as Customer Relationship Management, payroll or accounting. This was formerly known as "Application Service Provision" (ASP)

“PaaS (Platform as a Service)”: Programming environment where apps can be developed and deployed

“IaaS (Infrastructure as a Service)”: Raw computing such as processing and storage, and possibly network. The user typically controls the software systems and apps loaded and used (subject to licensing), much as they would otherwise do with on-premise equipment.

(2) By Type of Deployment

Each of the three categories above can be deployed in any of four ways:

“Public Cloud”: Multi-tenanted systems where various end-user organisations share the same hardware set and usually the same database. Usually adequate for commercial purposes, if the right functionality and security tools are available.

“Private Cloud”: A system dedicated to one end-user organisation, and customer/suppliers etc they wish to share with. This can be externally with a third party, or internally as a web-enabled on-premise solution. With SaaS, this is typically using the same app as is available in the public cloud. The private service, if available, is typically more expensive than the public one, as private ones require additional administration by the provider. The benefits of a private system are:

• To mitigate risks of disclosing confidential data (if segregation, encryption and other security techniques are not deemed adequate in the shared public system)
• Better control of administration, such as timing of upgrades and any recovery of the database (as control in these areas is rarely offered in a public system)

“Hybrid Cloud”: A set of applications run across a mixture of private and public cloud services

“Community Cloud”: Apps shared between various organisations or individuals in a community.

As mentioned above, most of the articles on cloud computing in this blog focus on Public SaaS, as this is of most relevance to most businesses. The significant benefits, and how to manage the risks, are covered in those other articles.