Eight energy-saving methods for data centers

Eight energy-saving methods for data centers

Some of the most effective energy-saving technologies require a large amount of front-end investment, which takes years to get a return on investment. Server leasing, but some technologies that are often ignored, have very little cost. These technologies are often ignored because they seem impractical or too radical. The eight energy-saving methods listed below have already been tried and tested in a real data center environment, which has proved to be effective. Some of the technologies you use in the production environment can immediately get a return on investment. Other technologies may require capital investment, but the rate of return on investment is faster than that of traditional IT Capital expenditures.

The energy efficiency measurement standard of the data center is the energy utilization rate (PUI). a lower ratio indicates a better utilization rate. 1.0 is the ideal target for energy utilization. The energy utilization rate is the comparison between the total power cost of the data center and the number of tasks converted into effective computing tasks. The common value of the 2.0 ratio means that the two-watt Energy originally consumed during a server test in the data center is reduced to one-the loss is that the energy is converted to heat, in this case, energy needs to be eliminated through a traditional data center cooling system.

From a simple calculation standard, you must look at and define the energy utilization rate as a method to calculate the power efficiency. This method does not take other energy sources into account, such as ambient temperature or hydrogen fuel units. Many of these energy sources can be used to reduce overall energy consumption. The techniques described below may or may not reduce your measurable energy utilization, but you can simply evaluate their effectiveness by reviewing the monthly electricity bill. In any case, this is the real concern of enterprise users.

You will not find solar, wind, or hydrogen energy in this article. These optional energy sources require a large amount of investment before the technology is put into practical use, which will greatly reduce the cost savings advantage achieved in the context of the current economic crisis. None of the following eight technologies needs to be more complex than fans and pipelines.

These eight technologies are as follows:

1. Increase the temperature of the data center

2. Disable unused servers

3. Use free external air for cooling

4. Use Data Center heat to heat the office area

5. Use SSD to run highly active read-only Datasets

6. Use DC in the data center

7. Import heat to the ground

8. Heat is discharged to the sea through pipelines

One of the basic energy-saving methods: increase the temperature of the data center

The easiest way to save energy is to increase the temperature of the data center temperature regulating device. Generally, the temperature in the data center is set to 68 degrees Fahrenheit or lower, which can logically prolong the life of the device and allow users to give feedback when the cooling device fails.

Experience shows that when the server hardware is faulty, especially when the hard disk is faulty, the operating temperature of the system is increased. However, in the recent year, the IT economy has crossed an important critical point: the current server operating costs generally exceed the procurement costs. This may make it more practical to retain hardware than to cut operating costs.

At last year's GreenNet conference, Google's energy manager BillWeihl introduced Google's experience in improving the temperature of data centers. Weihl said 80 degrees Fahrenheit could be safely used as a new value, however, the data center must meet a simple prerequisite: try to separate the cold air flow from the hot air flow, and use curtains or solid barrier if needed.

Although 80 degrees Fahrenheit is a safe temperature upgrade, Microsoft's experience shows that you can increase the temperature. Microsoft's data centers in Dublin and Ireland operate in a "cooling insufficiency" model, using free external air cooling, with the server entry temperature of about 95 degrees Fahrenheit. However, when you increase the temperature, the server fan speed required to increase the energy consumption will become higher, and the Server rental will cause a reduction in return.

Basic energy-saving method 2: Disable unused servers

Virtualization technology shows the benefits of combining unused processors with hard disks and memory. So why not shut down the entire server? Will this increase the "enterprise flexibility" that matches their energy costs? If you can find that the server is shut down, you can minimize the energy consumption of these servers to zero. However, you must first consider the opinions of opponents.

First of all, we usually think that because the pressure is accumulated on non-domain switching components such as the motherboard adapter, so the energy cycle reduces the expectations of the server life. Then we come to the conclusion that in practical applications, servers are built from the same components that are used in everyday devices through frequent energy loops (such as automotive and medical devices. There is no evidence that any reduced MTBF (that is, the average failure interval) is a result that the energy cycle server can bear.

The second objection is that the server took too long to start. However, you can usually disable unnecessary import time diagnostic tools to speed up server startup. You can directly import and use the hot boot feature provided by some hardware from a running snapshot image.

The third objection is: if we have to restart a server to adapt to the enhanced load, no matter how fast the import is, users cannot wait. However, most application software architecture is unfamiliar to new users, because the process needs to be simplified more slowly, so users do not know that they are waiting for the server to start. Application Software has indeed reached the limit on the number of users. If users know that "we start more servers to speed up the response to requirements", they may be willing to do so.