Clarifying a Cloudy Picture

How software-as-a-service relates to cloud computing.

As I have said before, ask six people to define cloud computing and you’ll likely receive at least half a dozen answers. This would be further complicated if you then ask them to explain how software-as-a-service (SaaS) relates to cloud computing.

Organizations that use software-as-a-service are accessing application software through a subscription-based pricing model; this is also commonly referred to as on-demand software. The application is hosted on hardware external to the deploying organization (as opposed to being installed on-premise/in-house). This results in reduced hardware and software acquisition and startup costs as well as eliminating or greatly reducing ongoing maintenance expenses. Although the classic example of an SaaS offering is the CRM and sales force automation software from, numerous other vendors have also adopted this delivery channel.

SaaS is not necessarily a new concept as evidenced by the number of application service provider (ASP) vendors that sprang up in the mid-1990s to offer their clients Internet access to application software that the ASPs remotely hosted at their data centers.

Companies can use the cloud to remotely access hardware computing resources including raw computer power and storage and associated virtualization and provisioning functionality. Although many users and vendors consider this to be the essence of cloud computing, it is merely just one part of it; I like to consider this part as “hardware-as-a-service” (HaaS), a term I consider to be more precise than the sometimes used “platform-as-a-service” (PaaS) because some vendors use it to include their remotely hosted on-demand software.

Among the vendors in this market are Amazon and Google. Although “the cloud” is almost always considered to be synonymous with services available via the Internet, it dates back to at least the early 1970s when computing power was available, via dial-up phone lines, from time-sharing vendors. In addition to these hardware services, many time-sharing companies offered application software services to their users as well.

Cloud computing, in its broadest terms, consists of both hardware and software components. Using the cloud for services (such as remote meetings and other collaborative efforts) may seem like another category, but these services are enabled by a combination of hardware and software services. Both hardware and software can be utilized on an on-demand basis as cloud services.

Many organizations can use hardware-as-a-service for augmenting their own computing resources, often in times of peak computing demand or for special one-time projects; most companies with software-as-a-service application offerings host their application software off premises. From the perspective of the user organization it is therefore hosted in the cloud although the hosting site might be the vendor’s own data center or a third-party site that the SaaS vendor uses as its own HaaS vendor. I believe it would be helpful for users to describe their cloud computing requirements, and vendors to market their cloud computing capabilities, by categorizing them as either software-as-a-service, hardware-as-a-service, or both.

Cloud computing appeals to organizations of all sizes. It allows small and midsize organizations to utilize resources that they could not otherwise afford to acquire and maintain on their own. Even large organizations with massive data centers are looking at cloud computing as a way to quickly augment their available resources to accommodate new user demands, be they be operational or analytical. After all, almost all BI practitioners have seen needed projects delayed or simply not started because it required additional hardware or software resources and additional administrative efforts that were not immediately available. Funding for these incremental resources may require multiple layers of approval even in good economic times; obtaining them from the cloud on an on-demand basis is usually much easier to justify.

Furthermore, on-demand cloud services can generally be implemented much quicker than obtaining and installing hardware and software in-house. From the perspective of BI users, the cloud can be the vehicle that enables them to undertake BI analysis that their organizations might otherwise not have been able to undertake, at least in a relevant timeframe.

The cloud is rapidly expanding as more organizations utilize cloud services to augment their own in-house capabilities or, in some cases, to ultimately replace them. Many business intelligence vendors already recognize this and are proactively moving ahead to get their heads into the cloud. Among them are SAP BusinessObjects which, in 2007, first launched, a cloud-based, on-demand offering of Crystal Reports, and Business Intelligence OnDemand, a data warehouse offering for analyzing data from via BusinessObjects Web Intelligence.

Database vendor Kognitio offers cloud-based data warehousing which it cleverly brands as “Data Warehousing as a Service,” and data warehousing appliance pioneer Netezza recently launched a cloud-based data warehousing service in conjunction with partner AppNexus, a HaaS cloud computing vendor. Additionally, both IBM and Microsoft are undertaking aggressive cloud computing initiatives.

Cloud computing, whether SaaS, HaaS, or both, is definitely more than a passing fad. User organizations and vendors must recognize it for what it is: a growing component of an organization’s overall future computing architecture.