The late Harvard Business School professor Theodore Levitt is quoted as saying “People don’t want quarter-inch drills. They want quarter-inch holes”. Put in terms of IT, “A service is a means of delivering value to customers by facilitating outcomes customers want to achieve without the ownership of specific costs and risks”. This is the definition used by ITIL.
Achieving IT value without owning the underlying assets is the essence of cloud computing, and NIST defines the basic service models of cloud as Software as a Service (SaaS), Platform as a Service (PaaS), and Infrastructure as a Service (Iaas). The key characteristics of these models are detailed in the following diagram[i]:
These service models form the three tiers of cloud computing.
- At the base layer, Infrastructure as a Service (IaaS) represents the foundation of all cloud computing. System Managers can leverage IaaS to deploy services without owning the facility, network, or underlying IT infrastructure.
- Next, Platform as a Service (PaaS) builds upon laaS to deliver value to the customer further up the IT stack, allowing Developers and Deployers to implement applications without the need to maintain the operating system and some basic middleware components.
- Finally, at the top of the IT stack, Businesses Users can simply use Software as a Service (SaaS) applications already deployed by others to satisfy their business objectives.
To put this into layman’s terms, the analogy of renting versus buying a car is often used to explain the premise behind cloud computing. When we need to get somewhere, the goal is to get from point A to point B, not necessarily to own and/or operate the vehicle. Similarly, the business value of IT should focus on outcomes produced, not the underlying infrastructure. Cloud computing allows customers to achieve their desired IT outcomes without the burden of ownership.
One key characteristic of commercial cloud computing is that it allows customers to shift their IT dollars from capital expenses (CAPEX) to operational expenses (OPEX). Historically, on premise IT infrastructure must be refreshed every three to five years to mitigate the risk of critical failures and to take advantage of the rapid evolution of IT, as measured in performance, size, density, or power/cooling improvements. Funding for this continual reinvestment, along with any capacity increase necessary to accommodate predicted growth, must compete with other organizational needs and investment priorities that may be difficult to forecast over that same future period. In the public sector, this can be quite a challenge when agencies are asked to “do more with less” and to respond to evolving threats and new mandates while planning budgets two or more years in advance.
Another key consideration is that infrastructure investments may require scalability for expected peak demand, resulting in underutilized assets during periods where demand is lower. If demand changes unexpectedly in the short term—either up or down, the on premise infrastructure cannot quickly scale to match demand because procurement, installation, and testing can often take months to accomplish. Thus, the scale of on premise IT infrastructure becomes fixed over the short term and represents a design constraint on the applications it supports. Peak demand cannot be satisfied if it exceeds this fixed capacity, and excess capacity exists in periods where demand is less than peak.
The diagram below compares the capacity of on premise IT funded as CAPEX at periodic intervals versus predicted and actual demand:
Where capacity exceeds demand, or visa-versa, the business incurs an opportunity cost. In the case of overcapacity, the funds used on the excess capacity have been wasted and cannot be recovered. In the case of under capacity, the business suffers lost revenue, loss of goodwill, or both because demand cannot be satisfied. If customers cannot access the application(s) or experience degraded performance, the business’s reputation will be harmed.
Cloud computing represents a fundamentally different investment paradigm compared to IT hosted on premise. There is no need to fund capital investments in IT capacity, incurring no CAPEX, and capacity can be provisioned (and de-provisioned) quickly per the actual demand curve. Customers incur operational costs (OPEX) only for the capacity they use, which is typically easier to predict across shorter time intervals. CAPEX must typically be planned across longer time periods and may carry other constraints. In the cloud, IT need not be provisioned day one to accommodate peak demand, only average or minimum demand, because scalability is an inherent quality of the service model. Other fixed costs, such as rent and salaries of staff who operate on premise IT, don’t apply to cloud and represent further potential cost savings.
When analyzing “build versus buy” options for IT investments, managers today have no choice but to incorporate cloud into the decision process. This adds an additional level of complexity given the differences between on premise and cloud computing, and one must also recognize the transformation opportunities available when using the cloud. Though difficult to quantify initially, the benefits of these possibilities should prove valuable in the long term.
How is your company investing in the cloud? We’d love to hear from you!
Do Good. Have Fun. Add Value.
[i] Based partly on: Guidelines for Secure Use of Cloud Computing by Federal Departments and Agencies, Federal CIO Council, September.