High performance businesses have begun to use new models for funding IT within an enterprise. Instead of the traditional cost center model, advancements in service-oriented architecture (SOA) and service-oriented infrastructures (SOI) enable IT organizations to move to an environment in which IT is consumed and paid for on a metered basis. This "pay-per-drink" model has three advantages: It elevates the status of the IT organization from a cost center to a value center, it helps redefine the IT organization as a service provider, and it provides the commercial framework in which to run IT as a "business within a business." It also raises the stature of the CIO within the organization.
MORE ON Chargeback
IT chargeback represents a significant change in the culture of an enterprise. Unsurprisingly, it takes a while for the new model to be understood and embraced. It needs to be thoroughly road-tested first. For this reason, many high- performance IT organizations are choosing to introduce chargeback internally, between the application layer and the infrastructure layer (the systems, network and storage that support the applications). This allows the model to be tested and proven, before IT "raises the bar" and exposes it to the lines of business. (Also read Chargeback Demonstrates IT Value In the Enterprise.)
The introduction of chargeback for infrastructure usage has a very interesting side effect. If an organization unitizes the cost of its key infrastructure components such as compute time, storage capacity and network data transfers, using comparable metrics, the opportunity exists to meaningfully market test those costs against third parties. To the IT organization it provides a powerful tool to measure its own cost-effectiveness.
On the flip side, to infrastructure outsourcing (IO) providers, it represents an opportunity to demonstrate value by comparing their price per unit with the potential client's internal cost per unit.
However, for infrastructure outsourcing companies with a technology consulting arm, the opportunities afforded by the new model are even greater. Instead of a traditional "your mess for less" play, high-performance infrastructure outsourcers can engage in "transformational outsourcing." In this new model the client's existing environment is first transformed to a standard model amenable to unitized pricing, and then billed back to the client using pay-per-use pricing. The costs of the transformation are offset by the subsequent pay-per-use fees.
Why Do We Need a New Model?
Conventional infrastructure outsourcing is about transferring IT systems management headcount, depreciation and facilities costs from the client to the outsourcer in exchange for periodic fees. For the better part of 25 years it's been viewed as a means of saving money, acquiring expertise in new technologies, and accelerating product development lifecycles. In many respects, conventional IO has been successful. Forrester Research reports that 38 percent of conventional outsourcing arrangements are considered a success by customers and Dunn and Bradstreet reports that 50 percent of conventional outsourcing arrangements are still active after five years.
Nevertheless, industry research indicates that substantial amounts of outsourced capacity is left underutilized.
At the core of the problem is the mismatch between the fee structure of conventional outsourcing and the variable nature of the demand for infrastructure itself. A pay-per-use unitized pricing model addresses that disconnect. In addition, without pay-per-use pricing and standard units of measure, it's hard to determine if outsourced infrastructure is priced at a level that genuinely reflects value to the client.
The Impact of Service-Oriented Computing
Over the last three years, the discipline of enterprise architecture has especially come to the fore. EA concepts have been the driving force behind a broad transformation of traditional IT enterprise architectures—a move toward service-oriented computing. Architectural innovations such as service-oriented infrastructures have enabled IT organizations to redefine the nature of their infrastructures to meet the rapidly increasing demand for service-oriented architecture concepts such as mashups, user-driven integration and agile componentized application software.
For an infrastructure outsourcer, this trend in enterprise architecture provides a key stepping stone to measurable, standardized, units of infrastructure value. For application developers and application owners, it provides the opportunity to rethink the costs of deploying a new application. Instead of drawing down a capital expenditure budget to procure dedicated computer systems and storage to host a new application, the organization can alternatively contract with a next-generation infrastructure outsourcer to procure the hosting environment as operational expenditure, and then pay for it only when needed. Such services can be deployed and provisioned quickly via a standard "executable" service catalog. Application owners simply pick the services they require to host their application.
This move from capital-based procurement of infrastructure assets to as-needed purchases of infrastructure services, helps put a "real" value on IT infrastructure. Just as mobile phone users buy service by the minute and don't need to own or have a part in the running or selection of the cell phone towers, so the application owners have the opportunity to pay for their infrastructure on an as-consumed basis, without having an ownership or operations stake in the equipment itself. The application owners hold the infrastructure outsourcer accountable via standardized service-level agreements (SLAs) called "service classes." (For more, read "An Introduction to SLAs".)
In addition, pay-as-you-use computing opens up the opportunity for innovative chargeback models, allowing the infrastructure outsourcers to stay agile and responsive to the changing needs of their customers. Innovation in services pricing knows no bounds. Just as the cellphone providers can manage their network loading with "free nights and weekends" packages, the high performance infrastructure outsourcing organization can use the power of its pricing models to drive the most effective utilization of its assets. Higher utilization means higher return on the capital invested in IT, which means a more efficient outsourcing organization and lower unit prices to its clients.
Transformation Is the Name of the Game
It's clearly to the infrastructure outsourcers benefit to help potential clients transform their IT infrastructures to be compatible with the new pay-per-use IO model. The pay-per-use model is not only a fundamentally more efficient commercial vehicle; it also provides the outsourcer with the opportunity to capitalize on economies of scale. By building large infrastructures that can be charged on a pay-per-use basis to multiple application owners, the infrastructure outsourcer is able to absorb the individual peaks and troughs of demand.
Shareable and meterable IT environments allow outsourcers to boost asset utilization from as little as 15 percent to more than 65 percent over the typical 24-hour duty cycle. This doesn't necessitate the sharing of compute and storage across different clients&MDASH;which would cause significant security concerns. And it's both viable and practical within the domain of a single client.
For example, imagine a single server located in the U.S. and run by the IT department of a global investment bank: Rather than standing idle in the early hours of the morning after completing that last batch of portfolio analyses, the server could now be used for tasks that are not specific to any application or function. It can be "borrowed" by the bank's traders in Singapore, for example, who are operating on an entirely different time zone.
Of course, sharing common infrastructure resources across multiple applications is not new concept. And an outsourcing relationship is not a necessary pre-requisite. In fact, many high performance IT organizations have adopted dual-use or multi-use models for their IT assets in the interests of operating efficiency. However, infrastructure outsourcers using the outsourcing 2.0 model have the advantage that they have already institutionalized these savings into their operating environments and pricing, and industrialized the processes to support them. This enables a client to reap the rewards immediately.
Infrastructure Outsourcing Is Not a Silver Bullet
The transformation of human perception and enterprise architectures takes time. But it is fast becoming a practical reality. Current technology allows IT infrastructure usage metering and cross-charging to be implemented quickly and efficiently; either internally or in collaboration with an outsourcing partner. High performance IT organizations need to understand which option provides the most cost effective solution for the enterprise.
Mark Denne is the author of Software By Numbers (Pearson 2003). He has been involved in several large transformational outsourcing engagements and is currently leading one of the largest software as a service (SaaS) transformations in the industry. His background includes prior roles as a partner with the enterprise architecture practice at Accenture, managing principal for the VERITAS/Symantec utility computing practice, and the manager of Sun Microsystems' Java Center.