Your program should include the following components.
Four areas must be considered when determining the amount of space that will be required for your data center: the size of the data center itself, the space in the building that supports the data center, the space for staffing needs, and the space outside the building that supports the data center.
Review with the team what functional areas will be required in the disaster recovery site. These areas might include:
• Mainframe/server equipment
• Printer room
• Tape library/storage
• NOC, operations center, help desk
• Situation room
• Support personnel
• Test room
It is also important to consider the space that will be required by infrastructure equipment. Equipment that is typically located inside your data center includes CRACs (computer room air conditioner), PDUs (power distribution unit), UPS (smaller systems), and clean agent tanks.
Infrastructure equipment located outside the data center but within the building includes electrical switchgear, UPS (large systems) and battery room, and TVSS (transient voltage surge suppression. CRAC condensers or chillers and generator(s) will be located outside of the building. Areas must be dedicated for that equipment as well.
It is essential to understand the electrical and mechanical load of the planned facility. This will drive the sizing of the electrical and mechanical infrastructure, the most costly part of your data center.
The first step is to determine the load profile of the data center. To do this most accurately, it is necessary to develop an inventory of computer equipment with its corresponding kVA and BTU/h specifications. This may be information you already have. If not, it is listed on the back of every piece of electrical equipment. Add to this the load requirements for people, lights, and non-critical power and you will have an accurate profile of the requirements for the data center.
Think about future IS requirements and how they might affect your load profile. Involve executive management in these discussions. Inquire whether there is anything planned that will significantly change load requirements such as an acquisition.
Level of Protection
Review by discipline (architectural, structural, electrical, mechanical, and fire protection) the level of protection currently in place in your data center and decide what level will be used in the new site. Do you require redundancy? Would this be in the form of an N+# configuration or does your application require 2N? Might your requirements change in the future? What are the trends in your industry? Are there any upcoming regulations in your industry that might affect the type of infrastructure that is utilized?
There are many options to consider. Many data centers are employing dual power path architectures because most of their equipment has dual power supplies. With the challenges of heat densities in racks, there are new options as to how to achieve sufficient cooling in areas that need it the most. There are newer fire protection technologies that provide an increased level of protection such as early warning air sampling systems.
Once you have established the above listed criteria, determining a budget is relatively straightforward. You don’t even need a site to determine a solid preliminary budget. Look for a team partner with in-house experts that has experience with data center construction and equipment procurement.
A budget with significant detail can be helpful in your decision-making process. Your budget should be segmented by discipline: architectural, mechanical, electrical, fire protection, security and site monitoring, and general conditions. Within each discipline, there should be a breakout of major components. For example the electrical budget should list major components such as generator, switchgear, and UPS. The fire protection budget should list clean agent separately from pre-action sprinkler. Consideration should be given to various technologies within each discipline such as direct expansion versus central plant for HVAC. This will enable your team to evaluate some possible different scenarios.
With the program established, sites can be looked at with an entirely different perspective. Use the program to develop an evaluation report card that is brought to each site. Your team can grade a facility while they are on site and turn in the report card to be tabulated immediately. This will shorten the evaluation process and ensure that the team members looking at the site are using the same criteria to evaluate it.
Now the team is ready to look at potential sites. In a future article, we will discuss the due diligence that should be undertaken when inspecting various sites and how to make the best decision about the location of your future data center.
Tad Davies is executive vice president of Bick Group where he has been involved with data center planning and construction for 16 years. Since 1964, Bick Group has provided planning, building, and maintenance services for data centers of all sizes.