Judging infrastructure

As CIOs are pushed to speed up responsiveness to the business while reducing costs, it is increasingly important that they know how to judge the contribution of their infrastructure organization. CIOs should insist that the infrastructure organization follow methods whose primary focus is to reduce cost of change, while also measuring the key areas of complexity and utilization.

META Trend: By year-end 2003, the results of increased focus on operations excellence, end-to-end infrastructure management, and relationship management will hasten the movement of many IT organizations (ITOs) toward a trust-filled relationship with business partners, decreasing the percentage of ITOs viewed as a cost center from 70 percent to 50 percent. Further relationship movement – from trust- to respect-filled – will demand that CIOs focus on process, especially those that exist primarily at the borders of business and technology (i.e., program management, strategic planning, enterprise architecture development). Only CIOs that follow a value management process – categorizing, capturing, and communicating value – will reach a relationship maturity of collaboration by 2007 (8 percent).

Although 50 percent of CIOs have a technology background, our research indicates that most are missing the point in judging their infrastructure staff’s effectiveness. A focus on reducing cost or maximizing use in initial designs – common objectives within infrastructure organizations – will result in CIOs inadvertently supporting infrastructures with significantly higher total lifetime costs. High-performing CIOs have recognized the need for end-to-end infrastructure patterns that drive reuse and adaptability. Indeed, throughout 2003/04, high-performing CIOs (less than 20 percent of Global 2000 CIOs) will extend their governance models to ensure that infrastructure designs support reusability as the primary goal, as opposed to scalability or reduced initial cost. Indeed, by 2006/07, high-performing IT organizations will have merged the principles-based philosophy of enterprise architecture design and reusable infrastructure patterns and services into a common set of processes and products whose objective is to reduce the total cost of change.

CIOs wishing to ensure that infrastructure is designed and leveraged properly should evaluate three areas:

– Commitment to reducing total costs of change

– Decrease in complexity

– Proper management of utilization

The first area – commitment to reducing cost of change -is an approach, while the other two – decreasing complexity and properly managing use – are indicators of effective infrastructure design and implementation. Management of all three areas produces effective asset management.

Reducing the Total Cost of Change

The primary design goal for information systems must be to enable rapid change in business processes and in the applications and technical infrastructure that enable them. Indeed, the fastest increasing percentage of total infrastructure cost is the cost of change.

Focusing on decreasing initial costs by concentrating on scalability or extendability will result in higher than average total costs over an asset’s life cycle. CIOs should insist that infrastructure staff use the concepts of concurrent engineering and integrated product development. This proven practice, as detailed by the National Institute of Standards & Technology, results in 30 percent to 70 percent less development time, 65 percent to 90 percent fewer engineering changes, and 20 percent to 90 percent less time to market.

Reducing Complexity

Reduction of complexity is simply measured by the number of different products within a specific domain. Although architects model domains and infrastructure developers implement patterns, both activities are oriented toward reducing complexity. In high-performing IT organizations, reduction of complexity is the result of fully inculcated architectural principles and standards. For most ITOs, however, true principles-based enterprise architecture remains elusive. We recommend that CIOs insist on a quarterly inventory of the number of product types within each major technology domain (i.e., middleware, management tools, database, network) and measure performance of the infrastructure group in reducing the different types therein.


Effective CIOs judge infrastructure use according to technology domain rather than at the elemental level. For networks, 35 percent to 40 percent capacity is considered appropriate (leaving enough “headroom” for peak loads). Likewise, storage should remain fairly steady at 75 percent capacity. CPU capacity should change, of course, based on the platform. Measures of effective CPU use are 5 percent for servers, 25 percent for servers in a distributed environment, and 70 percent to 85 percent for mainframe CPUs. In general, managing to these high-level metrics attains two goals:

– Provides an indication of staff performance, relating to overall usage metrics

– Promotes the start of necessary training of new infrastructure groups in multi-level reporting, relating to high-level metrics (for example, data is repackaged and communicated based on receiver value rather than on sender desire)

The extent to which infrastructure staff insists on centralized asset management is another indicator to CIOs of being on the right track. Indeed, our research indicates that costs can easily be reduced by 25 percent ot 30 percent through initial centralization of asset management and a continual 5 percent reduction year over year. Combining a philosophy of reducing cost of change with complexity reduction and proper use management will result in further, optimal asset management.

Business Impact: A CIO’s ability to properly judge infrastructure adaptability will help infrastructure staff focus on reducing the total business life-cycle cost of technology-based business solutions.

Bottom Line: CIOs must judge the effectiveness of their infrastructure staff regarding dedication to reducing the cost of change, ability to centralize asset management, and willingness to manage and communicate utilization on a domain rather than an elemental level.