Published on March 9, 2010

Finding and Resolving Data Center Weaknesses

Data centers are hubs of both the physical and virtual worlds, and weaknesses and single points of failure can threaten smooth operation from either side. The best approach to maintaining data center uptime is to prepare for malfunctions, failures, and attacks before they occur, regardless of whether they are physical or virtual.

Single-point failures are a major concern for the physical side of the data center. A single point of failure is essentially what it sounds like: a single point whose failure brings down the entire system. In designing systems for mission-critical or life-critical operation, single points of failure must be avoided at virtually any cost. Consider the data center from a design perspective: if only one power line, for instance, brings power to the facility, this power line is a single point of failure. Any of myriad potential accidents could bring down the line, thereby disrupting power and bringing data center operations to a dead stop. What is the solution in this case? Multiple power lines, preferably entering the facility at different locations, can prevent a single incident from cutting off all power. If one continues looking at the matter of power supply, one also finds that an accident at the power generation station can cut off power, regardless of how many lines come into the data center facility. Since data center designers and operators have no control over such things, backup power supplies can also help as a stopgap measure.

To read further, please visit the Data Center Journal website:
http://datacenterjournal.com/content/view/3626/43/