Abnormality Score Calculation for Root-Cause Estimation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing APM tools can detect failure occurrence services but require manual analysis of metrics, traces, and logs to investigate root-causes, which is inefficient and does not utilize metrics effectively for root-cause estimation.

Innovation Solution

An estimation device that calculates an abnormality score from metrics and traces to estimate the service with a failure and identify the root-cause, utilizing a multivariate time-series model to analyze the deviation from normal activity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If manual analysis of metrics, traces, and logs is performed to investigate root-cause, then accuracy of root-cause identification can be maintained, but operator workload increases and efficiency decreases

Engineering Contradiction:

Improvefailure investigation efficiencyVSAvoidoperator workload

Core Design Contradiction:

ProductivityVSEase of operation

Solution Approach 1:

The system performs automated root-cause analysis using the abnormality score calculation unit and root-cause estimation unit, enabling the system to serve itself in identifying failures and their causes without requiring manual operator intervention for analysis

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical manual analysis process with an automated computational system that calculates abnormality scores from metrics and traces using a multivariate time-series model, substituting human operator effort with algorithmic processing

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If metrics are not utilized in failure detection, then failure occurrence service can be detected using traces, but root-cause estimation becomes impossible

Engineering Contradiction:

Improveroot-cause estimation accuracyVSAvoidmetrics utilization

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent merges metrics and traces into a unified analysis framework where the abnormality score calculation unit processes both data types together, combining their information to enable both failure detection and root-cause estimation simultaneously

Inventive Principle:

Principle #5Merging (Combining)

3Extent of automation

If only failure occurrence service estimation is performed using metrics and traces, then automated failure detection is achieved, but root-cause estimation capability is lost

Engineering Contradiction:

Improvefailure detection automationVSAvoidroot-cause information

Core Design Contradiction:

Extent of automationVSLoss of information

Solution Approach 1:

The patent creates a multi-functional system where the same abnormality score calculation unit and multivariate time-series model serve dual purposes: both estimating failure occurrence services and estimating root-causes, maximizing the utility of the automated analysis

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20250068502A1Estimation apparatus, estimation method, and program

Publication Date: 2025.02.27 NIPPON TELEGRAPH & TELEPHONE CORP

AI summary

An estimation device includes an abnormality score calculation unit that calculates an abnormality score indicating a degree of deviation from a normal time from a metric obtained by quantifying an activity of each of the plurality of services and a trace in which time information and a call order of processing of each of the plurality of services are recorded, a failure occurrence service estimation unit that estimates a service in which a failure has occurred based on the abnormality score, and a root-cause estimation unit that estimates a root-cause based on an abnormality score of a metric of the service in which the failure has occurred.