What factors need to be considered for data center availability

1. Factors affecting availability in the data center

In many current data center environments, the uptime is generally set to "5 9s", which is 99.999% availability, which is equivalent to about 5 minutes of downtime per year. Generally, the downtime lasts one hour or more, which is equivalent to about one downtime every 10-20 years. Under normal circumstances, due to the interaction of many data centers from the system, it is very difficult to achieve this goal. 99.999% data center availability means that the availability level of each subsystem is much higher than this value. The reason is that the downtime of all subsystems must be equal to or less than 5min. Therefore, these factors need to be carefully considered, especially when determining the cost of reducing downtime.

When building, upgrading or reviewing a data center, each subsystem should be considered. The reason is that they have a huge impact on the normal operation and downtime of the data center. The solution provided for the data center should not be "complete", but should pay attention to the quality of the products provided. The availability of a certain subsystem may be able to reach five nines, but the availability of some other subsystems is not high, which will reduce the availability level of the entire system.

In order to achieve high reliability of the overall system, it is necessary to "equip" the availability according to the subsystem. Since the actual availability level of each subsystem is different, the availability goal of each subsystem is also different. For example, the availability of water-cooled transformers may be very high, but the availability of multiple moving parts of pumping systems is usually low. In this way, the design availability standard of the transformer may be 7 9s, and the pump may be limited to 6 9s, etc. By assigning availability levels to different subsystems to design availability targets, the solution can be applied to the data center to make the overall The system achieves availability goals.

2. Several major issues to consider when designing a data center

Physical structure

The physical structure of a data center is the doors, windows, walls, floors, and ceilings that can install data center equipment and protect it from environmental conditions. When considering where and how to construct the file, appropriate precautions need to be taken, such as floor load-bearing, the earthquake resistance level of the building, the surrounding environment (grid conditions, seasonal thunderstorms, interference conditions), etc., leaving sufficient room, Once the data center is built, it is extremely difficult to modify it. Therefore, careful investigation, accurate prediction, and careful planning must be done to determine the relevant data bureaus in the required data center. In addition, the building structure is generally used for decades. If possible, the design of the computer room structure should be able to allow simple renovation and expansion in the future.

power distribution

1. Input power supply in computer room (commonly known as primary power supply). The power source is the driving force for the operation of all equipment in the data center, from lighting, heating, cooling, dehumidification, to all IT equipment. In order to maintain the normal operation of the data center, it is necessary to seriously consider how to distribute power. The important components of this subsystem include: mains power input interface, cable, terminal block, distribution board, transformer, circuit breaker, transfer switch, socket board, etc. If the system requires very high availability, some or all of them need to be redundant.

2. Power backup. Data center power backup usually refers to the high-quality power supply used to protect the load from the inferior power supply and provide uninterrupted power supply to the load when the power grid is powered off. Usually batteries are used as temporary power supply, which can provide backup time from 5min to several hours. At a glance with other key subsystems, if high availability is required, redundancy is also essential.

3. Power generation. If the mains power outage lasts for a long time, it is not enough to use only the UPS standard configuration battery. Therefore, most data centers have adopted on-site backup power generation. These power generation systems are powered by diesel, natural gas or other hydrocarbon fuels and can provide high-quality power for long periods of time on site. In the event of a long-term power outage, the timely supply of generator fuel or large storage containers must have sufficient capacity to maintain the operation of critical systems. Similarly, when the availability of the system is high, the generator must also be redundantly backed up. With the development of technology, fuel cells will be put into use soon, which can save generator noise.

air conditioning

The heating and cooling subsystems provide basic environmental conditions for the data center. Since the heat generated by computer equipment is very large in most cases, heat dissipation must be seriously considered. In particular, a large number of high-density servers with small footprints are used to rapidly increase the power density of power sources in data centers, which may form many "hot spots" where heat is difficult to dissipate. The heat dissipation method, airflow direction and path must be considered in planning future power density levels and data center development. Of course, the cooling process will be accompanied by an increase in humidity, so it is necessary to add dehumidification means.

In the cold winter, the machine room needs to be heated, and the temperature is often accompanied by drying. Therefore, humidification measures need to be added. Whether it is humidification or dehumidification, you must master the scale, otherwise it will cause machine failure.

The temperature of many data centers will immediately exceed the operating temperature limit of the equipment after the cooling function fails. Therefore, it is often necessary to use system redundancy to avoid the opportunity. General precision air conditioning (HvAc) systems are usually connected to generator backup power to minimize downtime caused by power failure.

Computer room security

For any computer room where critical systems are installed, safety issues must be resolved. Access to the site is limited to those who have a pass and are properly trained to be able to use the relevant equipment correctly. Otherwise, it may cause unexpected failure. On-site security often includes armed guards, intrusion prevention, video surveillance, and advanced identification technologies.

Use dry cables for routing and elevated floors for air ducts

Elevated floors are currently used in most computer rooms. Elevated floor is a kind of special floor, which is divided into anti-static and ordinary type. It supports the equipment at a position of 12 ~ 48in above the floor. This makes it easy to route data and power cables under the device. Raised floors are also often used as air-cooled aisles to provide cooling for specific points in the data center. The type, quality and location and size of the raised floor also need to be carefully considered.

On-site monitoring

On-site monitoring is similar to the dashboard on a car. A timely understanding of the main parameters helps to carry out predictive maintenance. This monitoring must be continuous. Monitoring can be done locally or remotely. In this way, the data center can be monitored from any location.

Emergency power off

The emergency power-off system is also a subsystem of the data center, but everyone hopes never to use it. But in case of emergency, in order to avoid expanding the scope of the accident, the system was completely stopped and all other power systems in the data center were disabled. In the event of a fire or natural disaster, the system allows rescuers to enter the room without worrying about electrical safety.

Fire Fighting System

When carrying out capital construction, it is necessary to understand the regulations on fire protection and follow them. Because there are so many data and power cables in and out of the data center, fire prevention is an important issue, but this problem is often overlooked. In addition to the fire protection system, the data center is also equipped with a fire suppression system. Not only does the fire suppression system need to be integrated with the data center, but the sensors must also be very accurate. Only in this way can the accidental start of the fire protection system be avoided.

Design of physical interface

The physical interface between the machine room and the machine includes vents, vents, cable passages, etc., often because people do not think about these interfaces, which leads to the bite of the cable due to the mouse drilling under the elevated floor, the invasion of reptiles or flying insects It is not uncommon to have a short circuit inside the aircraft.

Two-way mains solution

There are always two views on how to use the mains and the secondary AC power supply (UPS) more effectively in the computer power room with two-way mains that requires high reliability. One view is that two mains powers are connected to the UPS at the same time (one-way Connect to the rectifier and bypass all the way); Another view is that the two-way mains should be converted into one-way mains in the input distribution cabinet and provided to the UPS. Both of these design concepts are in use, but their usability is very different. When designing the computer room, you must carefully compare its advantages and disadvantages and make the right decision.

Lighting of computer rooms and related environments

This problem is often overlooked. Especially large data centers often occupy a large area. For example, a northern bank data center not only has 160OkVA UPS capacity, but also is spatially distributed on five floors. The lighting time of the entire data center is 7x24h. This brings several problems:

· The frequent maintenance and replacement of such a large number of lamps will bring unsafe factors to the data center.

· The glare of the lamp will cause damage to people who have been working in this environment for a long time.

· Long-term uncontrolled power consumption of 365x24h will increase operating costs.

· The frequent flashing of lamp lights near the end of life is a serious pollution to the central environment and machines.

There is a power supply that can solve these problems. It can save 30%-40% of energy while ensuring that the brightness of the lamp is basically unchanged. The life of the lamp is extended 3-5 times, which will bring great benefits to the computer room.

Seamless Belt is made from PTFE coated fiberglass by Jiangsu Youngsun New Materials Co.,Ltd, its surface is smooth specifically engineered for fusing machines.
Its surface has thermal transfer and high temperature resistance characteristics.
Ensure the flow of woven and non-woven materials through the fusing and lamination process.
We have developed a full range of fusing machine belting products specially for garment industries.
Our Fusing Machine Belt are excellent compatible with many well-known fusing machines in the world,such as Hashima,Oshima,etc.

Seamless Belt For Fusing Machine

Seamless Belt, Hashima Endless Belt, Oshima Endless Belt

Jiangsu Esone New Material Co,.Ltd , https://www.cookware-manufacturer.com

Posted on