New Age -- A holistic approach to planning data centers can avoid capacity problems
July 22, 2002
Appeared in California Real Estate Journal
The data center overbuilding of recent years has companies more concerned than ever with designing and building cost-effective, efficient, flexible and reliable centers that deliver the most value for the investment.
To achieve these goals, an organization first looks at its business plan, information technology plan and projected growth, then plans increases in its data center capacity to support that growth. The challenge for a company is to keep its infrastructure aligned with its growth. If a company overbuilds, it ties up capital in unneeded infrastructure, such as overbuilt capacity. If it underbuilds, a capacity shortage could limit its growth, necessitating costly upgrades to increase capacity.
Keeping Infrastructure Aligned
A company can address this challenge through:
* Senior management involvement. Senior executives should actively participate in the process of planning data centers. The chief operating officer or other senior executives, such as the chief information officer, could head a team of facilities managers, information technology professionals, consultants and others in planning and designing centers.
The chief operating officer can provide the big picture, which helps the team in making planning decisions. For example, the chief operating officer may know about confidential business plans that show the organization's growth in e-commerce could triple its computing needs over the next few years. The other members of the team need to know this in planning future capacity.
* A flexible approach. A company can plan a flexible data center that not only meets its current needs but also accommodates its future growth.
For example, a company could build a facility that is partly used as a data center and partly as office space. As the company grows, the office space can be converted into a data center. However, knowing that the conversion is likely to occur, a company would design the data center infrastructure in a scalable format. This will assure that capacity can be added in logical chunks without the large cost or operational interruptions that would be required to rebuild the infrastructure.
* Targeted space planning. One reason for past overbuilding is that companies tended to assume that more space would be needed for everything. Instead, a selective approach to space planning is needed.
For example, a financial services company may need more servers to support more transactions, but it may not need more space for them because servers are becoming smaller and more powerful. However, it may require more space for power equipment, generators and coolers to support these servers. An understanding of the company's information technology equipment and strategy - and how that strategy is evolving - will greatly improve space planning.
* Selective overcapacity. Building more capacity for capacity's sake doesn't make sense. It does make sense to initially build excess capacity in key infrastructure components, such as main switchgear, certain electrical feeders or the mechanical piping system. This will enable the organization to expand its capacity in the future by adding scalable components to the system such as additional uninterrupted power supply modules without the substantial cost of replacing equipment and systems. Over the long term, this will reduce the data center's lifecycle costs.
It isn't enough for a company to align a data center's capacity, infrastructure, equipment and systems with the organization's current and future growth requirements. It must also ensure that the center is reliable.
Matching the center's reliability level to its vision and business use is key. Is the center a high-transactional facility where a minute of downtime could mean millions of dollars in lost revenue? Or is it a facility supporting supply-chain processes that can be offline for a short time without a direct revenue loss? The first facility will require a higher level of reliability and a correspondingly higher level of investment in infrastructure, systems and equipment.
Organizations must develop a targeted approach to reliability to ensure that their significant investment in data centers is consistent with their business needs. Building a highly reliable data center will result in significant life cycle cost and consume capital and operational resources that may well better support the business elsewhere.
When a data center project is initiated, hard decisions must be made regarding reliability, based on a detailed cost/benefit analysis.
Reliability must be built into a data center through redundant servers, state-of-the-art storage systems and faculty infrastructure systems, and it must also be proven reliable before operation begins. This requires a company to assure proper commissioning and test backup systems when construction is completed.
Granted, this is a time- and labor-intensive process. For example, testing a 15,000-square-foot data center can take two days and add about 5 percent to the total cost of the system. But the expense is small compared with the enormous losses that can result if the system should fail at some point. Only integrated systems testing under condition that simulate real loads will ensure that the facility systems are performing as designed and will work when called upon.
To continue to maintain the intended reliability through the life of the data center, the operational system must be part of an appropriate maintenance program. At high reliability levels, this would include a predictive maintenance model and regular testing under load to ensure that the system will perform in an emergency - that when it comes time for a backup generator to work, it does work.
In the final analysis, a data center isn't only about cost control or system reliability or building the right capacity for future growth. It is about all of these things. Planning and building a center requires a holistic approach that provides a company with the most efficient, reliable, cost-effective center for its investment.
Daniel McNary, PE, is a senior vice president of Syska Hennessy Group, a national consulting, engineering, technology and construction firm, and leads the design/build and construction division of OnlinEnvironments, the firm's national provider of critical facilities services.