Network Fault Detection via Distributed Event Automata
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems face challenges in efficiently detecting and analyzing causal faults in complex networks, leading to bottlenecks in data processing and storage due to the high volume of event data, which can result in avalanches of subsequent faults and difficulties in identifying the root cause.
Innovation Solution
A system and method that extracts event sequences ending with alarm events, constructs reduced event sequences, and forms automatons to detect these sequences, optimizing them into a deterministic automaton distributed across the network, using a publish/subscribe protocol to minimize data traffic and processing burden.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all event data is transmitted to a central processing system, then complete fault detection is achieved, but data transmission bandwidth and processing capacity become bottlenecks
Solution Approach 1:
The system divides the network into multiple domains, with each domain having its own event processing unit that independently processes events. This segmentation prevents all events from being transmitted to a central system, reducing overall data transmission volume while maintaining comprehensive fault detection coverage across all domains.
Solution Approach 2:
Event processing units perform preliminary processing of events at the domain level before transmission. By filtering and preprocessing events locally, the system reduces the volume of data that needs to be transmitted to central management, while still capturing all critical fault information for complete detection.
2Loss of information
If a central processing system handles all event data, then unified fault analysis is achieved, but storage and computing capacity requirements increase significantly
Solution Approach 1:
The system segments processing functions between domain-level event processing units and central management. Each unit handles local event filtering and initial analysis, while central management receives only processed event data for unified fault analysis. This distribution reduces the storage and computing burden on any single system while maintaining analysis completeness.
Solution Approach 2:
Event processing units act as intermediaries between raw event sources and central management. These units perform preliminary filtering and processing, transforming raw high-volume event data into reduced sets of significant events for central analysis, thereby reducing the capacity requirements of the central system.
3Quantity of substance
If event data is preprocessed in distributed units, then data transmission volume is reduced, but system complexity increases
Solution Approach 1:
Event processing units are designed as universal components that can be deployed in each domain with identical functionality. Each unit performs the same set of processing operations (filtering, pattern recognition, event reduction), making the distributed architecture manageable through standardization while reducing data transmission volume.
Solution Approach 2:
The system changes the parameter of event data representation by transforming raw events into reduced event sequences and automaton states. This parameter transformation reduces data volume while the standardized transformation process across distributed units manages the complexity of the architecture.
4Measurement precision
If all events are transmitted for analysis, then accurate root cause identification is achieved, but processing time increases due to data volume
Solution Approach 1:
Event processing units perform preliminary filtering and pattern recognition before events reach central management. By pre-identifying and filtering significant events locally, the system reduces the time required for central analysis while maintaining the accuracy needed for precise root cause identification.
Solution Approach 2:
The system extracts only the most significant events and patterns from raw event data at the domain level. By taking out and transmitting only relevant event information rather than all raw events, the system maintains root cause analysis accuracy while reducing processing time at the central level.
Data Source
AI summary
A system for error detection and troubleshooting analysis in a network consisting of a plurality of network components having software modules and communication interfaces and network nodes, which are connected to the network components by means of communication connections. The network components and/or the network nodes are designed to generate data, which are stored as a quantity of historical data, and to form event sequences from the quantity of historical data consisting of a sequence of events is provided. The system is designed to extract those event sequences from the event sequences which end with an alarm event (a), to extract relevant events in turn from said event sequences having an alarm event (a) for an error analysis and to construct reduced event sequences from the relevant events, and, for each reduced event sequence, to construct an automaton for detecting said reduced event sequence.