绿色数据中心环境下服务器热感知再定位（英文）

发布时间：2018-10-25 15:53

【摘要】：研究目的:数据中心服务器入风口温度升高将导致出风口温度随之升高,同时某些正在运作的服务器在高温环境下排放气体,可能会导致热点或硬件损坏。高出风口和热点造成制冷机制的负担。因此服务器可以被用于入风口温度灵敏度分析,且导致热点的位于高入风口区域的服务器可以被重新定位。创新要点:预测出风口温度作为入风口温度的参考。根据预测的出风口温度重新定位服务器,从而降低最大出风口温度并节省用于制冷的能量消耗。研究方法:根据能量守恒原则,提出服务器再定位算法(算法1),用于测试一组异构服务器。讨论不同测试组别下异构服务器再定位前后出入风口温度的时间响应以及CPU使用率(图2-28)。重要结论:所提热感知再定位方法应用于数据服务中心服务器可实现节能2.1 k W·h。再定位之后的服务器出风口温度同构化程度更高。对于每对再定位服务器,可以减少77%热点产生可能性。
[Abstract]:The purpose of this study is that the temperature of the air outlet will rise with the increase of the inlet temperature of the data center server, and some operating servers will emit gas in the high temperature environment, which may lead to hot spots or hardware damage. High air outlet and hot spot cause the burden of refrigeration mechanism. Therefore, the server can be used for the temperature sensitivity analysis of the air inlet, and the server located in the high air inlet area that leads to the hot spot can be relocated. Innovation points: predict outlet temperature as a reference for inlet temperature. The server is relocated according to the predicted outlet temperature, which reduces the maximum outlet temperature and saves energy consumption for refrigeration. Research methods: according to the principle of energy conservation, a server relocation algorithm (algorithm 1) is proposed to test a group of heterogeneous servers. This paper discusses the time response and CPU utilization rate of different test groups before and after repositioning of the heterogeneous servers (Fig. 2-28). Important conclusion: the proposed heat sensing repositioning method can be used in data service center server to realize energy saving 2.1 k W h.. After the repositioning of the server outlet temperature isomorphism is higher. For each pair of relocatable servers, the probability of hot spot generation can be reduced by 77%.
【作者单位】： Department
【分类号】：TP308;TP368.5

【相似文献】