Polymeric Technology Meets Data Centers
Data centers are mission critical facilities that run operations with profound significance. They are prevalent in industries including healthcare, telecommunications and information technology, BFSI, government and public sector, energy, and many more.(1) Businesses and infrastructures rely on these data centers to store sensitive data and support mission critical applications.
The continuous operation of these facilities in these industries is critical and unplanned downtimes are costly. Therefore, there is a need to protect these data centers. Polymeric technology can offer solutions to these facilities and allow them to continue their critical operation across various industries.
TYPES OF DATA CENTERS
Data centers represent an expansive and profitable market. There are currently three main types of data centers: on-premises, colocation, and cloud.
Types of collocation data centers
Colocation data centers can be further categorized as carrier-neutral or non-carrier-neutral data centers:
Downtime in Data Centers
The cost of downtime for data centers has been thoroughly researched by Ponemon Institute, sponsored by Emerson Network Power, since 2010. A research paper is published every 3 years with the latest one published for 2016. The study for 2016 analyzed 63 data centers ranging in industries from communications to healthcare to transportation. Out of the 63 data centers, 10 were colocation data centers. The average square footage of the data centers analyzed were 14,090 sq. ft. with a maximum square footage of 55,000 sq. ft. (8) The colocation data centers are most likely on the higher end in terms of square footage.
According to the study for 2016, the average cost per minute and average total cost of an unplanned outage was around $9,000/minute and $740,000, respectively. The total cost of an unplanned outage was as high as $2,400,000. Colocation data centers are most likely on the higher end of the costs due to the larger sizes. The majority of costs come from lost revenues and business disruptions. It is important to note that the cost of outages has steadily increased since 2010. (8) This trend may only continue as businesses become more and more dependent on data centers.
Maintenance in Data Centers
Preventative vs. reactive maintenance
The importance of preventative maintenance in data centers cannot be overstated. The contrary to preventive maintenance is reactive maintenance. Under reactive maintenance, repairs are only carried out after the equipment has broken down. About 55% of the maintenance in the average facility is conducted under reactive maintenance and the percentage is notably higher in data centers.
Maintenance in data centers includes HVAC system, Uninterruptible Power Supply (UPS), generators and building. The cost of maintenance in data centers is roughly 2% of the annual budgets. However, there are many unintended consequences of reactive maintenance that is often overlooked. Repairing equipment after it is broken means that there can be loss or corruption of data, shorter life time expectancy of equipment, increased energy costs, inefficient use of resources, and safety issues. This all leads to higher costs in the long-term. There is also the added complication of being unable to control maintenance budgets.
By implementing preventative maintenance, data centers can reduce emergency repairs and unplanned downtime. Problems can be fixed before they become serious. In addition, preventative maintenance leads to system reliability. (10)
HVAC in Data Centers
Temperature and humidity
It is important to monitor the temperature and humidity in data centers. Electronics generate heat during operation that needs to be dissipated. If the heat is not dissipated and the temperature is too high, electronics can experience reduced efficiency and even failure. In data centers, the high number of electronics operating leads to a high heat generation that needs to be constantly cooled. Normally, higher end equipment automatically shut down when the temperature is too high. A shutdown results in downtime. In addition, the system will usually experience computation errors and application errors before it shuts down.
The cooling fans in the system may be overworked during this time as well, reducing life time expectancy of the equipment. With equipment that does not have this automatic shutdown, the consequences of high temperatures are more severe. Low humidity can cause buildup of static electricity on the equipment and high humidity can cause corrosion to occur on the equipment. In order to prevent any reduced efficiency and downtime in these data centers, it is crucial to control the temperature and humidity.
There are three key areas to measure temperature at a data center: the bottom front of the rack (where the coldest cold air should be), the top front of the rack (where the hottest cold air should be), and the top back of the rack (where the hottest hot air should be). These three areas can be checked to ensure the temperature is not too high. The recommended cold air (front of the rack) temperature should be 64-80°F. The recommended hot air (back of the rack) temperature should have less than 35°F difference from cold air, typically less than 105°F. The humidity should be 40-60% relative humidity. (11)
Types of HVAC systems in data centers
Due to the importance of temperature and humidity control, data centers place emphasis on HVAC systems. There are two main types of HVAC systems for data centers: CRAC and CRAH.
Heat Exchangers Failure
Types of HEX failure
Heat exchangers can experience failure for many reasons. There are four main types of heat exchanger failure: mechanically, chemically induced corrosion, mechanically and chemically induced corrosion, and fouling. (13)