MiRack is equipped with independent air conditioners that can be flexible deployed in non IDC or semi-outdoor environments without considering heat dissipation issues.
MiRack has combined 3 intelligent systems including MiSeeR－failure prediction and abnormal detection system, smart energy-saving cooling system, and environmental control system, which included failure prediction, energy saving, information security monitoring, and etc. lead MiTAC to a new benchmark for 5G AIoT Edge Data Center.
MiRack is exclusive with huge 42 inches touch screen on the door. The LED lights on the screen display the health status of the cabinet instantly which proactively give an alarm before damage occurs to avoid major losses.
The smart energy-saving cooling system helps retain temperature inside the cabinet which is setting based on the external environment to achieve the most energy-saving effect. With smart power distribution unit (PDU) and AI technology, can know the health status of air conditioner and servers to do the prediction and maintenances in advance. Adopt in one-class learning AI algorithm to monitor power circuit and the characteristics of power consumption curves, so as to avoid information security risks such as implanted mining programs, worms, and even abnormal packet attacks.
Product Features
 |
Measures chiller current with PDU sensors and server temperature differentials using temperature sensing modules. Calculates heat dissipation using 𝑄=𝑆×𝑉×(𝑇2−𝑇1)Q=S×V×(T2−T1) to analyze the power required to remove unit heat. Recommends optimal cabinet air conditioning temperatures and actively controls temperature for smart energy savings.
Features an anomaly detection and degradation prediction system with self-training capabilities. Creates a mapping model based on chiller settings, chiller power consumption, and server power consumption to forecast server temperature. Compares predicted and actual temperatures to assess chiller health and predicts when conditions will trigger an alert based on trend analysis.
Monitors for cybersecurity risks such as mining software, worm viruses, or abnormal packet attacks, which are often indicated by unusual current fluctuations. Uses intelligent PDU and AI technology to monitor power usage of each device, analyzing server power consumption trends to detect potential mining software or worm infections.
Evaluates internal and external cabinet humidity to detect any abnormal moisture and assess leakage risk. Uses water detection units to identify water accumulation inside the cabinet, ensuring safe operation.
Inputs energy consumption requirements to generate optimal operational parameters for the chiller. Establishes validation standards with IoT-measured temperatures, recommending parameter adoption if standards are met or suggesting model updates if they are not. |
System Components
- MiSeeR－Failure Prediction and Abnormal Detection System
The anomaly and damage detection system with self-training function has passed the new patent (M603111) and the invention patent (No. 1758824). It adopts a uniquely developed algorithm deep learning architecture and automated feature extraction project, which can simulate extremely complex functions. It also supports high-performance parallel computing, and unknown and rare data including high-risk data can be effectively detected. If the server, air-conditioning equipment, etc. in the cabinet is faulty or abnormal, it will be displayed on the screen outside the cabinet and identified by LED light color, so that managers can still understand the current status of the equipment even when the screen is dormant. and abnormal warning. - Smart Energy-saving Cooling System
Compared with traditional cabinets, which require a good heat dissipation environment, MiRack AI cabinet only needs enough power to ensure the normal operation of the servers in the cabinets. Not only the industrial-grade air conditioning equipment with frequency conversion function is used, but the power consumption of the equipment collected through temperature sensors and intelligent PDU is combined with climate data through the AI learning mechanism, which can be used under different climate conditions or different operating states of the server. The recommended air conditioning temperature in the cabinet is estimated, and the template deployment is fully automated, highly flexible and convenient, and the temperature is actively controlled to achieve smart energy saving. - Environmental Control System
Common information security risks, such as servers being implanted with mining modes, worms, or even abnormal packet attacks, will be accompanied by abnormal current changes. Through intelligent PDU and AI technology monitoring, the independent circuit recording of each device is monitored. Electricity behavior: Analyze the change curve characteristics of equipment's electricity consumption behavior and the number and general categories of characteristics of normal electricity consumption behavior to establish an AI assessment model, which can effectively detect anomalies and estimate equipment security risk values based on outliers. If the system detects a high risk value, an alarm message will be provided on the display panel to facilitate users to quickly grasp the status of the equipment and eliminate high-risk behaviors.
DM Download
