On Reliability And Uptime Of It In Contained Solution

2016 15th IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm)(2016)

引用 0|浏览4
暂无评分
摘要
In this investigation, we utilize a new method of characterization in which IT airflow systems are ranked from a 1RU ToR switches, xRU servers to 9RU Blade Center. The free delivery (FD) and the critical pressure (Pc) are unique properties of each IT airflow system, satisfying its active flow curve (AFC). Those curves are used in previously validated models to investigate the interaction of IT with different air systems in a confined space. A validated model of containment is studied of a CAC aisle from Binghamton University data center lab. It is shown that reverse flow can take place in an operating piece of IT with weaker air system; this can endanger IT reliability and up time at different data center events. This can also change previous conclusions about the ride through time in well-sealed fully contained systems. Another important finding is that design containment systems based on the assumption of uniform load in the aisle/cabinet is a misconception that can endanger facility operation. In general, the reduction in airflow rate upon events in the data center (DC) is a function of each IT air system. To our knowledge, this is the first time in literature where both recent models of IT active flow curves and black box thermal inertia are combined to yield realistic failure analysis of contained solutions.
更多
查看译文
关键词
Mission Critical,DCIM,Reliability,Experimental Validations,containment,Predictive models,Data Centers
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要