Centralization of IT is definitely a good start in cost savings; however, it is not, by any means, all that it is cracked up to be without a vision for the future with an architecture to match. What kind of savings were realized by decommissioning systems, applications, and common infrastructure components? These will be minimal in context to the true sharing of resources to provide standardized, commoditized, and meaningful service levels (platinum, gold, silver, etc.). A utility model is a blast from the past but still viable. Cloud computing, at its very root, is a derivative of utility computing.
The average utilization of a physical server is less than 30%. The average utilization of storage is less than 50%. The average utilization of network is less than 70%. All of this is a symptom of the fact that there have been seperate silos of visibility, control, and architecture. While separation of expertise depth is absolutely required, Enterprise Capacity Planning coupled with alignment between IT and the business will drive cost reduction. Ironically, it will also drive agility, rigor, efficiency, complexity, and results.
If we focus solely on cost reduction, we negatively effect how we serve the business (ultimately the end customer) in the terms of agility, rigor, efficiency, and complexity. The direct and indirect dependencies among these significant execution components in conjunction with cost reduction correlates to the results experienced by the overall business. You might think me a contrarian; however, I am not. Cost reduction is absolutely necessary.
My concern is really over making certain that service management is delivered upon flawlessly while consolidating, standardizing, and commoditizing the environment. For instance, implementing virtualization as an element of cost reduction without understanding the increase in complexity, decrease in efficiency, and decrease in rigor in operations is an evaporating oasis. The costs associated in the operations could offset if not overwhelm the costs savings associated with a smaller physical server footprint. The linkages to backup, disaster recovery, administration, storage, and process need to be considered.
Enterprise Capacity Planning is a program which epitomizes the adage, “the whole is greater than the sum of the parts”. Too often we look at capacity management myopically. We focus on a particular portion of the environment (i.e. storage, network, server, etc.). Enterprise capacity planning, on the other hand, encompasses all of the elements of the environment, including HVAC, power, floor space, and the like. Therefore, we must also understand our loads, pipeline for the future, and end of life components.
As a part of the overall scheme to manage IT, we must first establish services that are consumed by internal and external customers. Unlike the historical consumption of resources (i.e. hardware, software, and storage), we must manage a shared set of resources efficiently, cost effectively, rigorously, and simply with agility and justice for all. A consumer whom is over taxing these shared resources must balance that consumption with compensation. Likewise, a consumer whom is under-utilizing these shared resources must be right-sized (“balanced”) for proper consumption and compensated proportionately.
If we are to do this, we must understand the key metrics which are important to the services we provide. This dictates that we examine “ourselves” on a regular basis to adjust and adapt to the changes in demand, market, and industry. The business of IT is no longer solely inward focused; therefore, the connection with the business element must be firmed up. Assumptions and assertions are no longer acceptable. This provides the basis for which we can start to look at cost reduction.
The fundamental “business” of IT must be understood for the organization which it supports. More recently IT supports external customers such as cloud and SaaS consumers, not just the systems on which customers rely internal to the organization. To meet this challenge and manage capacity and IT, we must build a firm base with constant governance and continuous improvement. We must stop thinking in terms of silos (albeit important in the overall scheme) and consider the whole when planning for capacity and managing IT long term. While network and server capacity management is fairly mature, storage capacity management is fairly infantile, not for a lack of effort. Therefore, we must take advantage of current efficiencies and shore up our inefficiencies as delivered in service form. Vagueness and assumption must be driven from the IT and Capacity Management landscape in return for specifics, standards, and policies which support the execution of a strategy based upon a vision for the organization.
Tactically, I advocate looking at decommissioning, virtualization, consolidation, standardization, and commoditization with an eye on the direct and indirect relationships to agility, complexity, efficiency, and rigor . In addition, I would look very closely at:
* Consuming as much as 70% of capital IT budgets, it will
provide the strongest return for cost reduction
* De-duplication, Single Instancing, Archiving, and Thin
Provisioning are a few techniques / technologies
that will assist (not sustainable without a change in
business as usual – processes and people)
* Orphaned storage (allocated but unused), under-utilized
/ wasted space (file-system & database), and stale
data (untouched / unaccessed) will provide the most
immediate impact if addressed in your
today without additional purchases
- Process Automation
* Consuming most of the operational IT budgets, it will
provide the best return on this side of the equation
* Provisioning, Authorization (rigor), Control,
Standardization, and Monitoring / Incident Management
are a few areas where you will find significant consumption
of man hours that could be automated
- Listen / Incentivize / Empower
* Inefficiencies are witnessed at all levels – show your
commitment to hearing and acting upon all input