AI Inference Engine for Server Failure Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for achieving high availability in cellular networks fail to predict server failures in advance, leading to unplanned downtime and delayed diagnosis of hardware failures, which affects user services and is inefficient in handling big data flows.
Innovation Solution
A scalable tree-based artificial intelligence (AI) inference engine is employed to predict server failures by identifying leading indicators from server parameters, allowing for proactive load balancing and reducing downtime through a heat map visualization of server health scores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional reactive monitoring methods are used to detect server failures, then failure diagnosis can be performed, but server downtime occurs before action is taken and availability is reduced
Solution Approach 1:
The system performs preliminary actions by identifying leading indicators and predicting server failures before they occur. The AI model analyzes server parameters and traffic patterns to detect early symptoms of failure, allowing operators to take preventive measures such as load shifting or server maintenance during planned downtime windows, thereby avoiding unplanned outages and improving overall server availability
Solution Approach 2:
The system implements continuous feedback by monitoring server parameters and traffic patterns in real-time. The AI model processes this feedback data to update failure predictions and adjust risk assessments dynamically. This closed-loop feedback mechanism enables the system to adapt to changing server conditions and maintain high availability through proactive interventions
2Measurement precision
If real-time analysis of large data flows from multiple servers is performed, then failure prediction accuracy improves, but processing complexity and computational resources increase
Solution Approach 1:
The system extracts only the most relevant features from large data flows, focusing on leading indicators such as specific server parameters and traffic pattern characteristics. Rather than processing all raw data, the AI model identifies and analyzes key features that most strongly correlate with failure predictions, reducing computational complexity while maintaining high prediction accuracy
Solution Approach 2:
The system transforms raw server parameters and traffic data into meaningful statistical features and metrics that the AI model can process efficiently. By changing the parameter representation from raw data to derived features (such as rates of change, thresholds, and patterns), the system reduces data dimensionality and processing complexity while preserving the information needed for accurate failure prediction
Data Source
AI summary
Server hardware failure is predicted, with a probability estimate, of a possible future server failure along with an estimated cause of the future server failure. Based on the prediction, the particular server can be evaluated and if the risk is confirmed, load balancing can be performed to move a load (e.g., virtual machines (VMs)) off of the at-risk server onto low-risk servers. High availability of deployed load (e.g., VMs) is then achieved. A flow of big data may be on the order of 1,000,000 parameters per minute. A scalable tree-based AI inference engine processes the flow. One or more leading indicators are identified (including server parameters and statistic types) which reliably predict hardware failure. This allows a telco operator to monitor cloud-based VMs and perform a hot-swap on virtual machines if needed by shifting virtual machines VMs from the at-risk server to low-risk servers. Servers having a health score indicating high risk are indicated on a visual display called a heat map. The heat map quickly provides a visual indication to the telco person of identities of at-risk servers. The heat map can also indicate commonalities between at-risk servers, such as if the at-risk servers are correlated in terms of protocols in use, if the at-risk servers are correlated in terms of geographic location, server manufacturer, server OS load, or the particular hardware failure mechanism predicted for the at-risk servers.