In my previous blog post, I had given you a quick look into what is a datacenter and how it is classified. In this blog post, I will be posting about the some of the key concepts and terminologies used in a datacenter.
PUE - Power Usage Effectiveness
This is one of the key terms that is widely used in a datacenter. The term PUE stands for Power Usage Effectiveness. It is the measure of how effectively datacenter a uses its power. It is calculated as the total power utilized by the IT equipment inside the datacenter. An ideal value of PUE is 1.0.
DCiE - Datacenter Infrastructure Efficiency
As the name suggests, it is the performance metric which is used to calculate the energy efficiency of the datacenter. It is the inverse of PUE.
DCIM - Datacenter Infrastructure Management
The integration of IT and facility management (such as lighting, cooling, etc) is called datacenter infrastructure management (DCIM). This integration is done to centralize monitoring, management and capacity planning of datacenters. This helps IT managers to identify the interdependence between the IT system and the facility and eliminate any source which can cause risk to datacenter availability. A DCiE of 40% means that only 40% of the total power used goes to IT equipment and the remaining 60% goes to cooling and ventilation. The larger the DCiE value, the better it is for the datacenter.
CFD - Computational Fluid Dynamics
This is a modeling system which is applied in datacenter to see how air flows within the datacenter. By studying the existing air-flow pattern, it becomes easier design an optimal air-conditioning system to ensure better performance of datacenter.
HVAC - Heating, Ventilation and Air-conditioning
The term HVAC refers to heating, ventilation and air-conditioning system built for the datacenters. These are critical components of the datacenter. HVAC itself is a major discipline where concepts of heating, ventilation and air-conditioning are studied in-depth. This technology is applied to data centers as it houses a huge number of computers generating a lot of heat and humidity. As the size of datacenter increases, so does the requirement and redundancy factor of these equipments.
Rack Unit
Commonly referred as RU, it is unit of measurement to describe the dimensions of the server unit which is to be housed inside the datacenter. One RU equals to 1.75 inches. The two common types of rack used are the 19-inch and 23-inch rack units. These are arranged in rows enabling easier access to the servers.
REC - Renewable Energy Credits/Certificates
REC for short are trade-able commodities which are available in the United States. This commodity is proof that 1 megawatt-hour (MWh) of electricity is generated from renewable energy resources such as solar, wind, geothermal, etc. These RECs can be bought and traded by the datacenters to offset any increase in power consumption in their centers.
IDS - Intrusion Detection System
An ID is a device or a piece of software which monitors the datacenter’s network or systems to ensure that there are no malicious intrusions. Whenever there is a violation of the network such as an unauthorized access to the system or the network, a signal is immediately sent to the datacenter administrator/management.
CRAC - Computer-room Air-conditioning
These are units placed inside the datacenter to monitor temperature, humidity, air flow, etc. This is a critical component of the datacenter as it monitors and reports if there is any raise or lowering of temperature/humidity. This would enable the datacenter management to ensure optimal performance.
kWh - kilo Watt hour
Power to the datacenter is sold in kWh units. The standard definition of a 1 kWh is the amount of energy delivered in one hour at power level of 1000 watts.
PDU - Power Distribution Unit
These are units within the datacenter which is responsible for delivering power to the datacenter.
UPS - Uninterruptible Power Systems
These are generally the backup units (generators, batteries) installed within the datacenter. These are used only when the main power source is disrupted due to any reason. The size of the UPS varies depending upon the size and power required to run the datacenter without any downtime.
ETR - Estimated Time of Repair
It is the total time taken to recover from a planned or unplanned outage in a datacenter. Usually this would be a finite value. Only in case of natural disasters such as flood or fire, this value would be infinite. Before formalizing datacenter availability, SLAs are drafted to include this value into the agreement.
SLA - Service Level Agreement
This is a service contract between the ISP (internet service provider) and companies. This contract is used to define the company’s availability objectives and requirements. This agreement makes it the provider’s responsibility to ensure guaranteed uptime or availability. It includes the technical definition of mean time between failures (MTBF) and mean time to repair/recovery (MTTR) in addition to the above mentioned factors.
If I have left out some key terms that you think are important for someone who handles datacenter, do drop a line in our comment section.