Study on Automatic LED Monitoring System for Data Center Devices

Monitoring is an essential process of data center operation. There are some common methods to detect data center device corruption or any type of intrusions in the operation. The best practices of monitoring data center are ensuring physical features of data center environment such as room temperature, humidity, and usage of the central processing unit (CPU) and storage of the machines which is used for hosting user services. Due to exceeding varieties of physical feature combination in data center environment, data center managing organizations cannot fully monitor it and experience unexpected conditions from time to time. Many organizations who manage data center also apply visual monitoring of Light Emission Diode (LED) indicators of devices with human interaction which takes large amount of cost and labor to reduce service interruption occurrences. In this paper, we present the concept and experimental results of an automatic monitoring system for data center devices. The concept of the automatic monitoring system is based on monitoring frequency of flickering LEDs of the device method. To evaluate performance of the automatic monitoring, we applied it to video recording of flickering LED of server device. During the implementation stage, we successfully located LED position from frame images and visualized series duration time of flickering LEDs. The experimental results present evaluation of the proposed method.