Inference Engine for Automated Network Failure Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network engineers face challenges in diagnosing and repairing network failures due to the complexity of unstructured data in support tickets, which can be lengthy and difficult to navigate, even for experienced professionals.
Innovation Solution
A technique that involves extracting phrases from support tickets, mapping them to an ontology model, and using an inference engine to infer concepts and trends, facilitating automated detection of failure causes and improving troubleshooting efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If network engineers manually analyze support tickets to diagnose network failures, then they can identify root causes through experience and training, but the process takes a relatively long time (e.g., days) and is difficult even for skilled engineers due to large amounts of unstructured data
Solution Approach 1:
The patent replaces the mechanical manual analysis process with an automated natural language processing system. The inference engine automatically extracts concepts from unstructured support ticket text, performs phrase extraction, maps phrases to ontology classes, and generates structured outputs without human intervention, thereby reducing diagnosis time from days to minutes while maintaining accuracy through systematic analysis of all ticket content.
Solution Approach 2:
The system enables self-service diagnosis by automatically processing support tickets and generating concept extractions without requiring network engineers to manually review the data. The inference engine independently performs phrase extraction, filtering, ontology mapping, and concept identification, allowing the system to serve itself in the diagnosis process and freeing engineers from time-consuming manual analysis.
2Loss of information
If support tickets include detailed unstructured data (free form text, device logs, automated messages, email conversations), then comprehensive information is available for diagnosis, but the tickets grow very large (e.g., one million words) and become difficult to navigate
Solution Approach 1:
The patent extracts only the essential information from large volumes of unstructured support ticket data. The inference engine performs phrase extraction to identify meaningful terms, filters out noise using frequency-based and length-based criteria, and extracts only relevant concepts while discarding redundant information. This extraction process condenses millions of words into manageable concept sets that retain diagnostic value while improving navigability.
Solution Approach 2:
The patent introduces an intermediary inference engine that acts as a mediator between the raw unstructured support ticket data and the network engineer. This intermediary automatically processes the unstructured text through phrase extraction, filtering, and ontology mapping to generate structured concept extractions, making the information accessible and navigable without requiring engineers to directly handle the raw unstructured data.
3Reliability
If network engineers rely on experience and formal training to handle network failures, then they can deal with particular types of failures, but even experienced engineers may take days to diagnose and repair certain failures
Solution Approach 1:
The patent performs preliminary action by pre-building an ontology model with classes and relationships relevant to network failures before diagnosis is needed. The system pre-extracts phrases from historical support tickets, maps them to ontology classes, and stores this structured knowledge in advance. When a new failure occurs, the inference engine can quickly query this pre-processed knowledge base rather than analyzing raw data from scratch, enabling faster diagnosis while maintaining reliable handling of various failure types.
Solution Approach 2:
The patent creates a universal inference engine that can handle multiple types of network failures through a single system. The ontology model is designed to be domain-specific yet flexible, allowing the same inference engine to process diverse failure scenarios by mapping different phrases to appropriate ontology classes. This universal approach eliminates the need for separate expert systems for each failure type while maintaining comprehensive handling capability and improving overall repair speed.
Data Source
AI summary
The described implementations relate to processing of electronic data. One implementation is manifested as a system that can include an inference engine and at least one processing device configured to execute the inference engine. The inference engine can be configured to perform automated detection of concepts expressed in failure logs that include unstructured data. For example, the inference engine can analyze text of support tickets or diary entries relating to troubleshooting of an electronic network to obtain concepts identifying problems, actions, or activities. The inference engine can also be configured to generate output that reflects the identified concepts, e.g., via a visualization or queryable programming interface.