Today: Financial Models and Utility Computing
Most IT managers and staff are used to working with tangible service assets they own; hardware, software, tools, and other assorted infrastructure, applications, and platforms. Based on our available resources and organizational capabilities (knowledge, skills, processes, etc.), we would design, plan, develop (or buy), and deploy services in support of our clients. This required the IT organization to use the business's capital to invest in IT assets, which may or may not be optimally utilized (consider hardware utilization, applications and associated licenses, data center infratructure and environmentals, etc.). Capital investments require a substantial commitment of time and upfront money to create a capability, long before a payback period begins (when the new service is delivering value in operations). If the business's needs evolve, substantial changes in service warranty are needed (capacity, service continuity, security, and availability) and will require rearchitecture, re-provisioning, and ultimately waste time and resources.
Most businesses seek a different model (and not just for IT!). Businesses generally prefer operating costs (pay-as-you-go) to capital costs, because we can stop buying it if we don't need it, or request more capacity or a different level of continuity or security (availability is a bit messier, as we'll discuss later this week) for an understood price per service. Utility Computing is really nothing more than delivering on the promise of ITIL Service Level Management, where for an understood and agreed price we will deliver an agreed level of service at a defined level of quality. The biggest difference then is WHERE we will deliver it...
In the Cloud Computing context, as with the ITIL Service Model, we seek to deliver value to customers while avoiding the ownership of certain costs and risks. Depending on the level of Cloud Computing your organization may consider, this includes potentially choosing not to own
- Applications - we'll use providers with SaaS (software-as-a-service) models to host and deliver key software applications, especially generic ones like productivity applications (a la Google docs) and ones where access across a number of devices and locations is a strong benefit (CRM applications like salesforce.com)
- (Development) Platforms - we'll count on our cloud providers to provide our developers with rich toolsets for building and deploying cloud applications; for example, Microsoft Azure extends .NET developers tools to build in the cloud.
- Infrastructure - we'll take advantage of the benefits of server virtualization, grid computing use of processing capabiliites, and enormous server farms managed by providers such as Amazon, Google, IBM, and many others to scale our storage, data management, and processing needs as they evolve, and make agility and real Capacity management far more accessible and realistic.
As our ITIL readers know, effective provisioning of whole services to customers involve the correct combination of these three things aligned with the business's needs and managed to deliver a reliable underpinning of their key business processes. Clearly cloud computing has enormous implications for the entire service lifecycle. To deliver "utility" computing, we must do more than just create pay-as-you-go models; we must actually deliver the right level of utility!
Next time, we'll talk about Cloud Architecture