IT organizations are pushing to reduce their operating costs and go green. In pursuit of these objectives they...
are consolidating data centers, addressing power consumption, implementing virtualization and assessing cloud computing. As usual, IT is in a rush to buy a solution, click "install" and move on to the next crisis. The problem is that the supposed solutions are built on a fragile house of cards, because many IT organizations and data centers do not have an accurate understanding of the services they provide and the discrete server and application configuration items (CIs) that make up those services. This problematically violates one of the most basic tenets of process improvement -- that the people, processes and technologies of the current state are accurately understood.
Simply put, if the current state is unknown, then how does an organization begin to set, let alone attain, desired objectives? In what order are things addressed? Which solutions make the most sense? These questions and many others cannot be adequately addressed. As a "fix," some groups take their visionary people in a room and ask, "What do we need to do?" or, more likely, "What should we buy?" Even then, the answer will be based on what these people know. A visionary and a walking configuration database are two very different things. A visionary may know where the organization needs to go but have little to no idea of current state. A person who knows quite a bit about which systems are in place and how they connect may be great at details but not have a vision. You can get very different outcomes in terms of solution designs depending on who is in the room.
Certainly if we want to improve power consumption and cooling, an understanding of what services are being provided today is critical information. In many data centers, the services and their configurations are often propagated in a manner akin to folk songs, and those with significant tribal knowledge are the shamans who pass along historical knowledge by word of mouth regarding what is where, why it exists and how the pieces connect.
Inventory the data center, understand configuration items
To understand a data center's current state means that a technical inventory is needed. It should include the relationships of CIs to one another as well as an understanding of processes and procedures relevant to those CIs. This data can then be used to populate a Configuration Management System (CMS) as well as to help identify process improvement opportunities. To be clear, the configuration data doesn't need to be incredibly detailed, but the vital data does need to be understood and accommodated. Indeed, one mandate is to start as simple as possible and evolve the data managed over time.
At this point, change management enters as a critical control process that is used not only to manage the risks associated with changes but also to govern the updating of the CMS to ensure that the data represented in it is accurate and timely. Without change management, the inventory investment will be wasted and the configurations will drift from what were inventoried to new and uncontrolled states.
Armed with this information about CIs, existing processes and procedures, IT organizations can make informed decisions about what to do next. Investigation must be done as to whether services can actually exist on virtual or shared servers, or if they require dedicated servers. Capacity modeling techniques can be applied and plans made about optimizing existing data centers, new approaches, and so on. For example, the Uptime Institute found that up to 30% of systems in a data center are ghosts -- meaning they consume power and space but deliver no business value because they're no longer needed and nobody told IT. There would definitely be space, power and cooling capacity benefits if those systems were identified and addressed as part of the inventory process.
In closing, improving a system, whether it's all of IT or a specific data center, requires that the current state be understood in terms of people, process and technology. For data centers that want to leverage green IT to reduce operating costs, accurate information about existing IT services must be available for effective and efficient management planning. Organizations that try to skip an understanding of current state risk making decisions on incomplete or erroneous data, and find that their new "solutions" aren't really solutions at all and result in higher levels of capital investment and operating expenses with lower levels of support. Instead, it's best to take the time and invest the money to understand CIs in the data center from the perspectives of people, process and technology.
ABOUT THE AUTHOR: George Spafford (http://www.spaffordconsulting.com/) is a Principal Consultant with Pepperweed Consulting and an experienced practitioner in business and IT operations. He is a prolific author and speaker, and has consulted and conducted training on regulatory compliance, IT governance and process improvement.
What did you think of this feature? Write to SearchDataCenter.com's Matt Stansberry about your data center concerns at email@example.com.