Creating a secure data center how to prevent the server from overheating

Although the industry has tried to unify the data center cooling methods, such as standardized hot or cold channel cabinet layout, Vuf (Verticalunderfloor) and Voh (verticaloverhead) ventilation system, etc., but we have to do a lot.

To help minimize risk and further refine these standard practices so that they can function over time, some specific strategies are described below:

Flexible design: Given the fact that most mission-critical devices undergo 4 to 5 complete replacement of computer equipment during the life cycle, we need to design them flexibly to ensure that the cost and risk are the lowest in the upgrade process and that continuous uptime is maintained. The key to flexible design is manifold, including installing large pipes, leaving extra space for additional computer and mechanical electronic equipment, etc.

Communication in the design process: in planning the room layout, IT staff must communicate with the design engineers so that they can better understand how to apply the hot and cold channel strategy layout equipment racks, how to decorate the porous board to optimize the air-conditioning distribution. In addition, communication can also help IT staff understand the long-term space constraints.

Using published guidelines: In order to assist in space planning, estimating load growth and mechanical electronic http://www.aliyun.com/zixun/aggregation/9344.html > System design, some industry groups have begun to publish global standards , such as Anglo powering of heating,refrigerating and air-conditioning Engineers (ASHRAE), 7x24 Exchange and uptime Cato. These guides take the first step towards a higher standard of data center design.

Using Computational fluid dynamics (CFD): CfD models can be applied in the following ways: A design engineer describes a computer room by inputting a physical space description, which includes a wall, ceiling, movable floor, equipment layout, and equipment load. CFD software then divides the room space into millions of "small units". Then, CFD Software calculates the energy, mass and momentum conservation equations of each unit by iterative process, and the output results in the temperature, pressure and velocity of each small unit. People are very clear about the physics and mathematics of the process and can help us to provide good predictions for practical applications.

The CFD model can not only play an important role in the initial design phase of the data center, but also be applied in the whole life cycle of the equipment-from initial design to various upgrades and expansions to proactively prevent the equipment from overheating and downtime.

Development of alternative system design and layout: In order to reduce hot spots in the cold channel (hot-spot), data center design must also have alternative design and layout, such as in addition to CFD analysis should also have a top air distribution (over the head distribution) and the combination of airflow treatment equipment. To compare the design of each alternative system, it is advisable to use several measurement indicators to measure the effectiveness of the cooling system.

Hewlett-Packard's SHI (Supply Heat index) and RHI (return Heat index) can help you record the heat flow recirculation into the cold channel and the amount of air flowing back into the air-conditioning unit. In addition, the RCI (Rack cooling Index) developed by Magnus Herrlin can also help us assess whether the cooling system is suitable for a particular environment.

Although the threat of device overheating persists in the IT field, the five steps described in this article are indispensable in the design of more standardized and cost effective data centers.