Normalized Graph Baseline Bands for Multi-Metric Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to provide an effective and easily understandable visual display of multiple system anomalies occurring simultaneously, as traditional graphical representations with multiple baselines can confuse users when tracking metrics over time.
Innovation Solution
A system and method that normalize actual value measurements and baseline bands using historical data to create a common scale, allowing for the display of multiple metrics on a single graph with a baseline band representing normalized measurement ranges, facilitating the identification of anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple metrics are displayed on separate graphs with individual baselines, then each metric can be tracked independently, but it becomes difficult to visually track multiple metrics against a common point in time and users may be confused
Solution Approach 1:
The patent combines multiple separate graphs displaying individual metrics into a single unified graph. Multiple metrics are plotted on the same coordinate system with a common time axis, allowing users to visually track all metrics against a common point in time. This merging eliminates the confusion of switching between multiple separate graphs while maintaining the ability to track each metric independently through distinct visual representations.
2Device complexity
If a baseline is displayed as a single line, then the graph is simple, but it is unsatisfactory when the baseline needs to represent a range of values
Solution Approach 1:
The patent transforms the baseline representation from a one-dimensional single line to a two-dimensional baseline band or shaded region. This baseline band has vertical extent to represent the range of expected values, while maintaining horizontal alignment with the time axis. This dimensional expansion allows the baseline to convey both the central expected value and the acceptable range simultaneously, preserving information about measurement variability without excessive complexity.
3Adaptability or versatility
If traditional graphical representations with multiple baselines are used, then each metric has its own baseline, but the display becomes confusing and hard to understand when multiple anomalies occur simultaneously
Solution Approach 1:
The patent creates a universal graph framework that can display multiple different metrics simultaneously using a common coordinate system and baseline representation. Each metric is plotted on the same graph with its own data series, but all share the common time axis and baseline band structure. This multi-functional display allows users to monitor various system metrics (CPU usage, memory usage, network traffic, etc.) in a single unified view, making it easier to understand overall system behavior and identify correlated anomalies.
Data Source
AI summary
In one example of the disclosure, an expected measurement range is identified for each of a set of metrics. Actual value measurements are obtained for each of the metrics. Each of the identified expected measurement ranges and the actual value measurements are normalized according to a common scale. A graph is caused to be displayed. The graph includes a baseline band representative of the normalized measurement ranges, and includes the normalized actual value measurements plotted relative to the baseline band.