Cognitive Server Prediction via Decision Trees
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for managing and predicting problematic servers in unknown server groups are either speculative or intrusive, requiring extensive investigation and diverting expert resources, which is undesirable and inefficient.
Innovation Solution
A method that computes profile parameters for an unknown group of servers by selecting similar known groups from a historical repository, constructing decision trees based on these parameters, and predicting the number of problematic servers, allowing for non-intrusive and accurate identification and reduction of issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If extensive and intrusive investigation is performed on a contemplated set of servers, then the speculative portion of the service level agreement commitment is minimized, but expert resources are diverted from other tasks, systems are disrupted, and system performance is adversely affected
Solution Approach 1:
The patent creates a virtual copy of the server environment through simulation, allowing investigation and analysis without touching the actual servers. The simulation model replicates server behaviors, interactions, and failure modes, enabling experts to conduct thorough investigations while the real servers continue operating normally, thus resolving the contradiction between measurement precision and productivity
Solution Approach 2:
The patent performs preliminary investigation and analysis through simulation before committing to service level agreements. By conducting virtual investigations in advance, the system gathers necessary data and insights without disrupting actual server operations or diverting expert resources from production tasks, thereby maintaining both accuracy and productivity
2Measurement precision
If extensive and intrusive investigation is performed on a contemplated set of servers, then the speculative portion of the service level agreement commitment is minimized, but system performance is adversely affected
Solution Approach 1:
The simulation creates a virtual replica of the server system that can be investigated thoroughly without affecting the actual servers. All intrusive measurements, monitoring, and analysis are performed on the simulation copy, ensuring the real system maintains its performance and reliability while still enabling precise predictions for service level agreements
3Productivity
If a simulation-based approach is used to predict problematic servers, then expert resources are not diverted and systems are not disrupted, but the prediction accuracy may be reduced compared to intrusive investigation
Solution Approach 1:
The simulation creates a faithful virtual replica that preserves the predictive accuracy needed for service level agreements while eliminating the need for intrusive real-system investigation. The simulation model is calibrated to match actual server behaviors and failure patterns, maintaining prediction accuracy without requiring expert diversion or system disruption
Solution Approach 2:
The simulation acts as an intermediary between the need for accurate prediction and the constraint of not disrupting real systems. It mediates by providing a virtual testing ground that delivers prediction accuracy equivalent to intrusive investigation while keeping actual servers operational and expert resources available for production tasks
Data Source
AI summary
A set of profile parameters to characterize an unknown group of servers is computed. A set of known groups of servers is selected from a historical repository of known group of servers. A subset of known group is selected such that each known group in the subset has a corresponding similarity distance that is within a threshold similarity distance from the unknown group. A decision tree is constructed corresponding to a known group in the subset, by cognitively analyzing a usage of the set of profile parameters of the unknown group in the known group. Using the decision tree a number of problematic servers is predicted in the unknown group. When the predicted number of problematic servers does not exceed a threshold number, a post-prediction action is caused to occur on the unknown group, which causes a reduction in an actual number of problematic servers in the unknown group.