Small data center planning and design principles
1、 Definition of small data center
The Data Center is a large-scale collaborative network of specific devices, which is used to accelerate the transmission of information on the Internet network infrastructure. It can also be subdivided into enterprise level data centers, other data centers, etc.
Enterprise Data Center (EDC) achieves the goal of data sharing and utilization by realizing unified data definition and naming standards and centralized data environment. Enterprise data centers can be divided into department level data centers, enterprise level data centers, Internet data centers, and hosting data centers. Among them, the Internet Data Center (IDC) is a center that provides users with professional and standardized data storage services and other related services with carrier class computer room equipment. Users can build their own Internet platform by hosting, leasing, virtual hosting and other services in the data center, and renting the technical strength of the data center.
Among them, as time goes by, the definition of indicators for small data centers is also changing. 2-3 years ago, those with an area of less than 100 square meters could be defined as small data centers. However, with the gradual centralization and enlargement of computer rooms, the current definition of computer rooms with an area of less than 200 square meters as small data centers is more appropriate. See the following table for specific division:
2、 Several core elements of building a small data center
Overall design
1. Location problem The data center has extremely high security requirements. Once problems occur, they will affect the normal production and life of society. Therefore, even in the event of natural disasters such as earthquakes, floods, lightning strikes and accidental events such as machine equipment failures and fires, the data center must also have high security and reliability to ensure that business will not stop. Improving safety and reliability can be achieved by various means, such as building redundant systems and systems that are easy to repair faults. However, the reasonable planning of the geographical location of the data center is extremely important to improve the "disaster avoidance". When investigating the project related to the location of the land and the properties of the land itself, it is necessary to conduct a multifaceted study of the natural and social conditions from the past to the future.
Among them, the Code for Design of Electronic Computer Room (GB 50174-93) stipulates that the location of electronic information system room should meet the following requirements: ◆ sufficient water, stable and reliable power, convenient transportation and communication, and clean natural environment; ◆ Keep away from factories, warehouses, storage yards, etc. that produce dust, oil smoke, harmful gases, and produce or store corrosive, flammable, and explosive materials; ◆ Keep away from areas with potential flood hazards; ◆ Keep away from strong vibration and noise sources; ◆ Avoid strong electromagnetic field interference.
In addition, the domestic Code for Design of Electronic Information System Room (GB 50174 - Draft for Comments) should soon have a clear result, which will give more detailed description of site selection, and will not be repeated here.
Relevant reference on site selection in Appendix F of ANSI/TIA-942-2005 standard: ◆ The data center should not be close to flight routes of nearby airports.
◆ The distance from the railway and expressway is more than 800m to reduce the risk of chemical substances splashing.
◆ It is more than 800m away from the military base and more than 1600m away from the nuclear facilities and national defense facilities. ◆ Other precautions: pollution risk/approaching the police station, fire brigade, hospital, regional planning, multiple access, vibration problems, environmental problems.
2. The data center planning, design and construction of the data center should meet the current and future needs. When planning, pay attention to the following requirements:
A. To solve the problem of system overheating, local overheating is a common problem in traditional data centers. The cooling of traditional data centers cannot be distributed according to the needs of actual equipment due to unreasonable design, resulting in high overall energy waste and local overheating and downtime. The air conditioner in the machine room is set unreasonably. Instead of using the precision air conditioner dedicated to the machine room, the family comfort air conditioner is directly used. The power cable is laid too carefully, which has a major potential safety hazard. The safe operation of the equipment in the machine room cannot be guaranteed without a power supply.
The cost of scientific data center power configuration will increase at the initial stage of construction. Compared with the business reliability brought by the stable operation of the system, it also reflects the low cost value brought by energy conservation and consumption reduction.
B. To solve the problem of system downtime, a large number of enterprises and institutions will build their own computer rooms or have their own computer rooms through telecom service providers and IDC operators. However, due to various reasons, the data center will be shut down and temporarily closed, which will bring immeasurable losses to enterprises. First, establish its own disaster recovery center. In addition to IT support, disaster recovery should include business impact analysis, strategy formulation, business recovery plan, personnel structure, communication support, third-party cooperation agencies, etc., to become a business continuity plan (BCP). Second, perform N+x redundancy backup for key equipment to minimize problems in the equipment itself.
（2） Product selection and layout
1. For small data centers, UPS and other power systems should adopt compact space saving design, with built-in bypass maintenance switch and expandable battery, so as to minimize the total cost of ownership of users. It shall also have user-friendly intelligence and system management functions.
The cost and performance of the product should be taken into account when adopting the product to meet the total cost of ownership requirements of IT managers. In some small data centers, IT managers often have to manage and maintain some power solutions themselves. There are no power professionals around, so the purchased products should have good performance and be easy to manage.
The selected product shall also be equipped with a wide range of optional accessories, such as network management card with environmental monitoring function. With this network management card, users can remotely or web-based manage UPS. It can be integrated with the system manager to achieve easy management of the entire data center.
3、 Several problems to be noticed in building a small data center
1. The design and installation of infrastructure and power environment should be reasonable first. The data center should be configured with guaranteed power supply, at least N+x UPS power supply, to ensure that the equipment can operate normally after the mains power interruption. If possible, another generator should be configured to ensure the impact caused by the long interruption time of mains power.
Second, the data center should be equipped with precision air conditioners dedicated to the computer room, and do a good job in overall monitoring. It can adjust the operation of air conditioners at any time according to the changes in the environment and humidity of the computer room.
Third, the data center should build a dedicated air supply channel. Because the equipment in the data center has large energy consumption and heat generation, the downward air supply mode has the best effect, so the floor must be set in the machine room to establish a dedicated downward air supply channel.
Fourth, special planning should be carried out. We should make a good plan according to the company's development. It is best to make a comprehensive plan for the computer room, and then implement it step by step according to the needs of business development. Only in this way, the construction of the computer room will be normal, and the power cable will not generate heat.
2. Focus on each part of energy consumption, and strictly prevent exceeding the standard. First, IT equipment system. The power consumption generated by servers, storage, network communication and other equipment accounts for about 50% of the total power consumption required by the data center computer room. The server accounts for about 40% of the total power consumption, and the other 10% is shared equally by storage devices and network communication devices.
Second, air conditioning system. The power consumption generated by it accounts for about 37% of the total power consumption required by the data center computer room. About 25% of the power consumption comes from the power consumption generated by the air-conditioning refrigeration system, and about 12% of the power consumption comes from the power consumption generated by the air-conditioning supply and return air systems.
Third, UPS power supply system. Their power consumption accounts for about 10% of the total power consumption of the computer room. Among them, about 7% comes from the power consumption generated by the UPS power supply system, and about 3% comes from the power consumption generated by the UPS input power supply system.
Fourth, lighting system. It accounts for about 3% of the power consumption required by the data center computer room.
3. Pay attention to the overall cost of the data center. To achieve maximum utility, we should pay more attention to energy conservation and sustainable development than building a small data center, because with the development of business, the computer room may be rebuilt and expanded at any time. The construction of the data center is not only about the cost performance of servers, storage and other equipment, but also the construction of the data center itself should follow environmental protection standards. Generally speaking, when it comes to green environmental protection measures in data centers, people first think of reducing energy consumption, increasing energy efficiency, reducing the number of air conditioners to save energy, etc. The measures are basically focused on the transformation of hardware equipment. Software is also the main means to achieve green, energy saving and environmental protection in the data center. The biggest contribution of software to the green data center is that it can change the architecture of the data system in the data center, thus reducing the number of servers used and various hardware devices in the server, thus realizing the goal of the green data center.
