Automatic Visual Monitoring System for Data Center Device Management
Misheel Enkhbaatar, Tatsuya Yamazaki · 2022
One of the core responsibilities of a data center is to provide stability and availability for online services. But the data center frequently faces infrastructure failures and unpredicted irregularities. In order to minimize hardware related failures, data centers must monitor the housing devices for 24 hours and 365 days. Due to various physical and software related failure causes, monitoring has always been a challenging task. A visual inspection by human operators is a reliable method which is widely used for checking device's status. However, growth of number of devices in data centers causes exponential cost growth for visual checks and it is almost impossible to detect anomaly occurrences in real time. Therefore, we have proposed an automatic visual monitoring system for data centers to substitute human visual monitoring. In this paper, we present the overview of the proposed system and show experimental results to evaluate LED detection performance of the system.