Failing to create a plan for a data center project can be a recipe for disaster. Mistakes made in the planning process can be exacerbated once the project is underway, causing delays, cost overruns and a potentially compromised system – never mind the ensuing frustration of project participants and the end-user.
The time put into detailing project development steps can save time and money in the long run. Involving people in the planning process who have knowledge about each step is essential to avoiding unexpected hazards down the road. Plotting steps and associated tasks in sequential order as in a data flow model can take the project from concept to completion in an organized manner.
There are two principal parts of a data center project plan: process planning and system planning. Process planning involves assigning responsibility for each task or step, determining project management, making decisions regarding outsourcing, establishing a budget and refining the schedule.
System planning translates the original concept into a detailed design plan. Here are five essential components of system planning:
1. Determine IT parameters.
This step starts with a specific business need that requires a change to the organization’s IT system. The step looks at three factors: criticality, capacity and growth plan. Criticality refers to the importance of the data center to the business, rated on a scale of 1 to 4. Capacity is the maximum IT power load during a data center’s lifetime. This number will be the “maximum final load” parameter in the IT load profile. The growth plan addresses the expected IT load over the data center lifetime – including initial IT load, minimum final load, maximum final load and ramp-up time (in years).
Leaders of the end-user organization are involved in this first step. The discussion focuses on a specific business need – i.e., “We need a backup data center” – as opposed to details about implementation – i.e., “The data center should be 500 watts per sq. ft.” The latter is too detailed and can bog down the process. Discussions about implementation by people who are not technically oriented can steer a plan off course.
2. Develop system concept.
Developing the system concept involves using the components of step one – criticality, capacity and growth plan – to devise a general concept of the physical infrastructure system. It is based on choosing a reference design that reflects the criticality and capacity, but that also can be scaled to support the growth plan.
This step involves choosing a reference design – a prototype of a collection of attributes including criticality features, power density, scalability features and instrumentation level – as well as selecting the physical room in which the data center will be housed.
3. Determine user requirements.
Each project will have components particular to the end-user – specific preferences the user would like – as well as constraints. Preferences may include things the end user wants but could change or alter if new information becomes available. Constraints are pre-existing conditions that are out of the customer’s hands and are either difficult or impossible to change, such as a room’s ceiling height or a preexisting law to which the project must conform.
4. Generate specification.
The specification is the “rulebook” that must be followed in creating the detailed system design. This includes standard specifications, which do not vary from project to project (such as regulatory compliance and best practices) as well as user specifications (preferences and constraints). The specifications outlined here must be fulfilled by the detailed system created in the final step.
5. Generate detailed design.
A detailed design for the installed system is the final step in the planning process. This design includes detailed lists of components, exact floor plan of racks (including power and cooling equipment), installation instructions, the project schedule and “as built” characteristics of the design, including efficiency, density and expandability.
System planning is essential in taking an initial concept to final detailed design. For more information on developing a system plan, see APC by Schneider Electric’s white paper, Data Center Projects: System Planning.
6 years ago
Executing a data center project requires detailed planning. You never know what might go wrong. Your blog on the parameters to be considered while planning the entire process is very systematic – exactly the kind of information I was looking for!