No abstract

No abstract

We describe an inexpensive autonomous robot capable of navigating previously unseen data centers and monitoring key metrics such as air temperature 1 . The robot provides real-time navigation and sensor data to commercial IBM software, thereby enabling real-time generation of the data center layout, a thermal map and other visualizations of energy dynamics. Once it has mapped a data center, the robot can efficiently monitor it for hot spots and other anomalies using intelligent sampling. We demonstrate the robot's effectiveness via experimental studies from two production data centers.

In this paper we address the high level of validation of commercially available numerical predictive tools, namely CFD models, for data center applications. Experimental data at a real world facility is compared with computational results. Specifically, a recently developed temperature measurement tool is used to capture three dimensional temperature profiles of the facility with very fine spatial granularity. These detailed contours based on actual measurements reveal hot air recirculation patterns in the room as well as variable utilization levels of the room air-conditioning units. We compare these experimental temperature distributions measured using a novel 3D mapping tool with CFD thermal modeling results. An algorithm for generating CFD models for real world systems is proposed and demonstrated. It is found that due to the extensive inaccuracies in rapidly gathered input data as well as inherent limitations of the model, there can be significant discrepancies between predicted and actual temperatures. In addition to steady state measurements, transient data was also collected and are presented. Knowledge of the transient temperature profiles at different parts of the room allows an estimation of the temporal behavior of the data center. Transient temperature fluctuations are presented which capture the real variations in the system boundary conditions, for example the temperature of the chilled water to the air conditioning units, or the power dissipation of the servers over time.

No abstract