This post launches the special report series exploring the importance of data center power quality, and its impact on modern facilities.
This special report series examines the equipment, operational characteristics and conditions that can influence various points in the data center power chain, which ultimately can the impact the power quality and operation of the IT equipment.
Understanding Key Elements of Data Center Availability
Generally speaking, the term data center “availability” is focused on maintaining the security of a physical facility and the power and cooling infrastructure systems that support the IT equipment (ITE). The data center industry has several different organizations which have developed guidelines and terminologies which are commonly used as an availability classification reference.
It should be noted that in addition to the Uptime Institute, the Telecommunications Industry Association “TIA” also has a similar rating structure (1-4) contained in their TIA-942-B document. It is a design specification referenced by many in the industry.
Also to be noted that the redundancy requirements apply to power and cooling systems, as well as other critical infrastructure elements of the data center.
Counting the 9s
In addition to the four redundancy levels noted above, data center owners, operators and customers are concerned about remaining operational 24×7. From the facility perspective, this is primarily focused on ensuring there is no loss of power and cooling, and is referenced annually in the form of a percent (i.e. 99.999%), and informally, expressed as the number of 9s: the proverbial “Five 9s”.
In addition to the four redundancy levels, data center owners, operators and customers are concerned about remaining operational 24×7.
However, while all this is the industry norm, it does not necessarily address the issue of quality of the power delivered to the ITE. In fact, it is possible to have some undetected power quality issues (such as random transients), which may impact the ITE or network resulting is data errors or loss, yet not be considered as “downtime” since power was never lost.
The power path can range from a very basic single path design or contain as many redundant components as required to deliver desired level of fault tolerance which is considered by many as the primary measure of evaluating electrical availability to the IT equipment and cooling systems, as well as other miscellaneous loads, such as lighting, and other security and support functions.
Power Management vs Power Quality Monitoring
The impact of some of these type of issues, (interruptions and sag/undervoltage), are relatively apparent and can be detected by basic power metering and recorded by logging software. Others, such as Transients, which can be random and extremely short, as well as waveform distortions, are not normally detected by typical power meters.
Moreover, older and smaller data centers in particular, usually lack basic power measurement or monitoring systems. In those cases, the only way they become aware of a potential power problem is when IT support personnel report the issue. In some cases, even when redundant A-B power paths exist, if one side of redundant circuits fail, IT administrators may not be aware of the exposure since the IT equipment that have dual power supplies will continue to operate. Even many newer, larger data centers which may have more comprehensive power measurement at the branch circuit level, do not have a complete picture of some of the issues of assessing power quality to IT equipment.
In the coming weeks, this special report series will further explore topics like IT power supplies, PDU transformers, PQ monitoring, power quality at the edge and more.
Download the full report, “Understanding the Importance of Power Quality in the Data Center,” to explore the importance of power quality in the data center.