According to the EPI Data Centre Training Framework, lithium-ion batteries are becoming more popular in data centres due to their higher energy density, longer lifespan, and lower maintenance costs compared to lead-acid batteries1. However, lithium-ion batteries also have some drawbacks, such as higher initial cost, stricter safety requirements, and potential thermal runaway risks1. Therefore, a battery management system (BMS) is essential to monitor and control the voltage, current, temperature, and state of charge of each battery cell or module, and to prevent overcharging, over-discharging, or overheating23. A BMS can also communicate with the UPS system and provide information on the battery status, performance, and health2.

References: 1: EPI Data Centre Training Framework, Module 5: Power, Section 5.3.2: Battery Technologies, Page 5-19 2: Benefits of Lithium-ion batteries for data centers at the edge,Data Center Knowledge, Page 1 3: Lithium-Ion Batteries in Data Centers, Data Center Systems, Inc, Page 1

The expected time to recover or repair from a system from a failure is defined as MTTR (mean time to recovery or mean time to repair). MTTR is the average time it takes to restore a system or product to its normal operation after a failure. MTTR includes the time spent on detecting, diagnosing, and fixing the problem. MTTR is an important metric for measuring the reliability and availability of a system or product, as well as the efficiency and effectiveness of the incident management process. A lower MTTR indicates a faster and more reliable recovery or repair process.

References:

•EPI Data Centre Professional (CDCP®) Preparation Guide, page 17

•What is MTTR? | IBM

•Explained: All Meanings of MTTR and Other Incident Metrics

Class D fires involve combustible metals or combustible metal alloys such as magnesium, sodium and potassium. These metals can react violently with water, air, or other chemicals, and require special extinguishing agents1

References: 1: EPI Data Centre Professional (CDCP®) Reference Materials, page 16.

IP protection grades are a way of showing the effectiveness of electrical enclosures in blocking foreign bodies such as dust, moisture, liquids, and accidental contact. IP stands for Ingress Protection or International Protection, and it is defined by the international standard IEC 60529. IP ratings consist of the letters IP followed by two digits and an optional letter. The first digit indicates the level of protection the enclosure provides against access to hazardous parts and the ingress of solid foreign objects. The second digit indicates the level of protection the enclosure provides against the ingress of water or fluids. The higher the number, the better the level of protection. For example, IP65 means the enclosure is dust-tight and can withstand water jets from any direction. IP68 means the enclosure is dust-tight and can be submerged in water under specified conditions.

References: EPI Data Centre Training Framework, CDCP Preparation Guide, IP code - Wikipedia, [IP Ratings Explained | Ingress Protection Rating | IP Codes | Updated 2022]

Cost of downtime is the total amount of money lost due to a data centre outage or disruption. It can be classified into two categories: direct and indirect. Direct costs are the immediate and measurable expenses incurred during or after an outage, such as lost revenue, lost productivity, recovery costs, compensation costs, penalties, etc. Indirect costs are the long-term and intangible impacts of an outage, such as reputation damage, customer dissatisfaction, loss of market share, legal liabilities, etc. Both direct and indirect costs can vary depending on the type, duration, and severity of the outage, as well as the industry, size, and location of the data centre.

References:

•EPI Data Centre Training Framework1

•EPI Data Centre Competence Framework2

•Understanding the Cost of Data Center Downtime3

•Uptime Institute’s 2022 Outage Analysis Finds Downtime Costs and Consequences Worsening

•[INFOGRAPHIC] The Cost of Downtime: 21 Stats You Need to Know

Fiber-optic cables use light as the source to transmit data. Light pulses are modulated to carry information through an optical fiber. The light is confined in the core of the fiber by total internal reflection at the core-cladding interface. The light travels along the fiber with minimal loss or interference, making it suitable for long-distance and high-bandwidth applications.

References: EPI Data Centre Training Framework, Principle of Data transmission through fiber optic cables, Fiber-optic communication

Availability and reliability are two important aspects of data centre performance that measure how often the system is operational and how dependable it is. According to the EPI Data Centre Framework, availability is the percentage of time that a system or component is in an operable state, while reliability is the probability that a system or component will perform its required function under given conditions for a specified period of time. Both availability and reliability can be affected by various factors, such as design, maintenance, human error, power supply, security, etc. One of the factors that can have a significant impact on both availability and reliability is cooling. Cooling is essential for maintaining the optimal temperature and humidity levels for the IT equipment and preventing overheating, which can cause failures, downtime, and reduced lifespan. Inadequate cooling can result from insufficient capacity, poor airflow management, faulty components, or environmental conditions. Inadequate cooling can reduce the availability and reliability of the data centre by increasing the risk of thermal stress, hot spots, performance degradation, and equipment damage. Therefore, cooling is a critical factor that can affect availability and reliability in a data centre.

References:

•EPI Data Centre Framework1

•EPI Data Centre Training Framework2

•EPI Data Centre Competence Framework3

•Data Center Availability, Reliability Hinge On Numerous Factors4

•A Look at Data Center Reliability and Availability5

Water sprinkler systems are designed to protect the structure of a building from fire by suppressing or extinguishing the flames with water. Water sprinkler systems are typically installed in the ceiling or walls of a building and are activated by heat or smoke detectors. Water sprinkler systems can reduce the risk of fire spreading and causing structural damage to the building.

References:

•EPI Data Centre Professional (CDCP®) Preparation Guide, page 28

•Fire Protection Systems for Data Centers | EPI

Waveform distortion is a type of AC power quality anomaly that occurs when the shape of the voltage or current waveform deviates from the ideal sinusoidal shape. Waveform distortion can be caused by nonlinear loads, such as rectifiers, inverters, variable frequency drives, and electronic devices, that draw current in pulses or harmonics. Waveform distortion can result in overheating, reduced efficiency, malfunctioning, or damage of equipment.

References: EPI Data Centre Training Framework, CDCP Preparation Guide, 5 anomalies in AC power that can damage your home devices, 9 Most Common Power Quality Problems

Data centre design and infrastructure standards can vary from country to country, depending on the local regulations, codes, and practices. Therefore, it is important to check the local standards before designing, building, or operating a data centre in a specific location. Compliance to only international standards may not be sufficient or adequate to meet the local requirements, which could result in legal, financial, or operational risks. For example, some countries may have stricter fire safety, environmental, or energy efficiency standards than the international ones. Some countries may also have different electrical standards, such as voltage, frequency, or plug types. By checking the local standards, you can ensure that your data centre is compliant, safe, and efficient in the local context.

References:

1: Data Center Design: Which Standards to Follow?2, page 1, section 1 3: The Most Important Data Center Design and Infrastructure Codes and Standards4, page 1, section 1 5: Explaining the new family of ISO Data Centre Standards6, page 1, section 1 7: Standards for data centre certification. Is a changing coming?8, page 1, section 1

Data centres are essential for supporting the IT operations and applications of various businesses across different industries. Data centre downtime can have a negative impact on the business results, such as loss of revenue, customer satisfaction, productivity, reputation, and competitive advantage. According to a web search, the average cost of data centre downtime in 2020 was $8,851 per minute, and the average duration of a data centre outage was 95 minutes1. This means that a typical data centre outage could cost a business over $840,000 in direct and indirect losses1. Therefore, data centre downtime can have a significant impact on the business results, regardless of the industry or sector.

References: Unplanned Downtime Costs More Than You Think, Downtime, Outages and Failures - Understanding Their True Costs, The Unplanned Downtime Nightmare — And How Operators Can Avoid

Shielded twisted pair cables (STP) are Ethernet cables that feature additional protection against electromagnetic interference from external sources, such as radio waves, microwaves, or other network cables. This is achieved by wrapping each pair of wires with a conductive shield, usually made of foil or braided wire, and then enclosing the entire cable with another shield layer. However, this shielding is not effective against low frequency electromagnetic fields (EMF) from power cables, which can induce currents and voltages in the network cables and cause signal distortion or data loss. Low frequency EMF can only be reduced by increasing the distance between the power and network cables, or by using a tre-foil cable arrangement, which is a special configuration of three power cables twisted together to cancel out the magnetic fields they generate.

References: STP Cable: Your Shield Against Network Disturbances; What is Shielded Twisted Pair Cable? - Advantages, Disadvantages; Shielded vs. Unshielded Cables: What’s the Difference? - Cable Matters.

The casters and feet under the rack are used to support the weight of the rack and its equipment, and to allow the rack to be moved if needed. However, the casters and feet should also be designed to avoid putting too much pressure on the floor tile, especially if the data centre uses a raised floor system. A too heavy point load on the floor tile can cause the tile to crack, deform, or collapse, which can damage the rack, the equipment, and the underlying infrastructure. To prevent this, the casters and feet should be larger, so that they can distribute the weight over a larger area and reduce the point load. The casters and feet should also be compatible with the floor type and the load rating of the floor tile.

References:

1: CDCP Preparation Guide, page 22, section 2.4.1.1 2: Raised Floor Design Considerations for Data Center3, page 1, section 1 4: [SOLVED] server rack on casters or feet5, page 1, section 1

According to the CDCP Preparation Guide1, risk can be defined as the product of impact and probability. Impact is the measure of the negative consequences or losses that may result from a risk event, such as downtime, data loss, or damage to the data centre. Probability is the measure of the likelihood or frequency of a risk event occurring, based on historical data, expert judgment, or statistical analysis. By multiplying impact and probability, risk can be quantified and compared, which helps in prioritizing and mitigating the risks. For example, a risk event that has a high impact but a low probability may have the same risk level as a risk event that has a low impact but a high probability.

References:

1: CDCP Preparation Guide, page 25, section 2.5.1 2: Data center risk assessment: A decision-making tool3, page 1, section 1 4: Accounting for Risk in Your Data Center Design/Build Strategy5, page 1, section 1 6: A Novel Framework for Data Center Risk Assessment7, page 1, section 1

The current recommended temperature for ICT equipment as described in the ASHRAE TC 9.9 guideline is 18-27 C (64.4 - 80.6°F). This is the recommended range for the dry-bulb temperature at the inlet of the servers, which is the most critical parameter for ensuring the optimal performance and reliability of the ICT equipment. The recommended range is based on the thermal specifications of the majority of the ICT equipment in the market, as well as the energy efficiency and environmental considerations of the data centre cooling systems. The recommended range is suitable for Classes A1 to A4 of the ASHRAE thermal guideline classes, which cover different types and generations of ICT equipment.

References:

1: 2021 Equipment Thermal Guidelines for Data Processing Environments - ASHRAE2, page 1, Table 2.1 3: ASHRAE TC9.9 Data Center Power Equipment Thermal Guidelines and Best Practices4, page 10, section 2.1 5: Does your Data Center Follow ASHRAE TC 9.9 Thermal Guidelines for Data Processing Environments?6, page 1, section 1

When having two non-synchronized power sources, the ATS / STS need to be of the type break before make, which means that the switch disconnects from one source before connecting to the other source. This prevents any short circuit, back feed, or phase mismatch that could occur if the two sources were connected simultaneously. Break before make switches are also known as open transition switches, because they create a brief interruption in the power supply during the switching process. This interruption is usually acceptable for most ICT equipment, as they have internal power supplies or batteries that can handle the transient. However, if the interruption is not acceptable, then the two power sources need to be synchronized before switching, which requires a make before break switch, also known as a closed transition switch. Make before break switches connect to the second source before disconnecting from the first source, which ensures a seamless transfer of power without any interruption. However, make before break switches require that the two sources have the same voltage, frequency, and phase, which can be achieved by using a synchronization module or a phase-locked loop.

References:

1: CDCP Preparation Guide, page 17, section 2.3.1 2: STS in data centres - Borri3, page 1, section 1 4: Using Static Transfer Switches to Enhance Data Center … - Donwil5, page 1, section 1 6: What is an Automatic Transfer Switch (Power)? | Ethan Banks7, page 1, section 1

VESDA (Very Early Smoke Detection Apparatus) or HSSD (High Sensitivity Smoke Detection) systems are the fastest smoke sensors among the options listed. These systems use a network of pipes to draw air samples from the protected area and analyze them using a laser-based detection chamber. VESDA/HSSD systems can detect smoke at very low concentrations, typically in the range of 0.005 to 20 percent obscuration per meter. This means they can provide early warning of a fire before it becomes visible or spreads. VESDA/HSSD systems are ideal for data centers and other critical facilities that require high levels of fire protection and minimal downtime.

References: VESDA Smoke Detection Systems - Xtralis, HSSD Smoke Detection Systems - Fire Protection Online, Smoke Detection in Data Centers - Siemens.

Class C fires involve energized electrical equipment, such as computers, servers, switches, cables, and wiring. These fires require the use of non-conductive extinguishing agents, such as carbon dioxide, dry chemical, or clean agent, to prevent electrical shock and damage to the equipment. Water-based extinguishers, such as Class A or K, are not suitable for Class C fires, as water can conduct electricity and cause electrocution or short circuits.

References: EPI Data Centre Training Framework, CDCP Preparation Guide, ABCs of Fire Extinguishers