The cloud is a set of services and technologies that enable the delivery of computing services over the Internet in real-time, allowing end-users instant access to data and applications from any device with Internet access.
It's useful to categorize cloud offerings as software-as-a-service, platform-as-a-service, and infrastructure-as-a-service. These services and platforms can be used individually or in combination with existing in-house infrastructure. Many cloud services can be deployed in-house, referred to as "the private cloud."
Software-as-a-Service (SaaS, or "cloud applications") is a model of applications delivered over the Internet, usually through a Web browser. Many of these services are consumer-oriented, but increasingly enterprise applications are coming from the cloud. In fact, a recent IDC report predicted that 76 percent of businesses would use SaaS in 2009, and that SaaS spending would grow more than 40 percent.1 Examples of cloud applications include Google's Gmail and Google Docs, Salesforce.com CRM, and Zoho Office. In each of these scenarios, applications and accompanying data is stored remotely and delivered over the Internet through a Web browser. Customers generally pay on a per-user subscription model.
Infrastructure-as-a-Service (IaaS) is the delivery of computer architecture over the Internet. It involves the use of remote computers (operating systems, databases, middleware, applications) and storage, oftentimes enabling cloud applications. The idea is that large, specialized players can operate server farms more efficiently than the average enterprise, so IT departments can supplement or replace their current servers with outsourced resources. This allows IT departments to scale up or scale down service levels quickly with minimal costs.
While IaaS adoption has lagged SaaS, it is quickly gaining traction. According to a recent Forrester report, 25 percent of enterprises are using or plan to use IaaS services in 2009.2
As with cloud applications, most cloud services follow a pay-for-use model where customers pay for computing time by the hour or by amount of storage used. Introduced in the 1990's as "Grid Computing," the maturation of the Internet and its associated technologies has furthered its adoption. More recently, Amazon's EC2 and S3 have emerged as the dominant cloud providers, but IBM, Force.com, AT&T, Google and dozens of startups offer IaaS products as well.
Platform-as-a-Service (PaaS) is the model whereby application middleware and build tools are made available to developers by a remote service provider. The most popular PaaS offerings include Saleforce's Force.com and Google's App Engine. Developers use these tools to rapidly build Web-based applications. PaaS development tools are usually free, with the customers paying only for the resultant hosted application.
Together, SaaS, IaaS, and PaaS represent the cloud ecosystem. Ultimately, organizations will piece together their own labyrinth of services from each category to increase application performance and decrease costs.
The Private Cloud
The increasing availability and use of virtual machines and software appliances are enabling the development of the private cloud: in-house environments that manage workloads using cloud methods. IBM forecasts that these "private clouds" can lower IT labor costs by 50 percent, improve capital utilization by 75 percent, reduce provisioning time from weeks to days, and reduce end user support costs by 40 percent. With these cost savings come increasing reliability and availability.
1Economic Crisis Response: Worldwide Software as a Service Forecast Update, 11/2008
2Conventional Wisdom Is Wrong About Cloud IaaS, 5/29/2009