Network State Monitoring via Local Evaluation Messages
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Critical systems face challenges in reliably monitoring the communication state of interconnected computing nodes due to silent malfunctions, non-homogeneous latencies, and asynchronous operations, leading to potential false alarms and operational inconsistencies.
Innovation Solution
A method where each calculation node periodically transmits and receives local evaluation messages about the network state, allowing for collaborative reporting of malfunctions and synchronization, ensuring a coherent functional context and consistent time management across all nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If local tests (loopback tests) are implemented at each computing node to report network state, then the complexity of each function is reduced, but the network state monitoring becomes incomplete as only local parts of the network are tested
Solution Approach 1:
The network monitoring function is segmented into local components (each node performs local loopback tests) and global coordination (centralized collection of local results). Each computing node independently performs local monitoring tasks, and the results are aggregated to form a complete network state picture, resolving the contradiction between local simplicity and global completeness
Solution Approach 2:
The patent combines local loopback test results from multiple computing nodes with centralized monitoring data to create a comprehensive network state view. By merging distributed local tests with centralized coordination, the system achieves both reduced function complexity and complete network state monitoring
2Measurement precision
If integrated monitoring system is implemented within each function to detect message loss, then the network state detection capability is improved, but the function complexity increases and functions become highly dependent on deployment topology
Solution Approach 1:
The monitoring logic is extracted from individual functions and implemented as a separate, dedicated monitoring system. This allows functions to remain simple while a specialized monitoring component handles the complex task of detecting message loss and reporting network state, separating concerns between functional operations and monitoring operations
Solution Approach 2:
A dedicated monitoring system acts as an intermediary between the network communication and the functions. This intermediary handles the complexity of message loss detection, timing synchronization, and topology-independent monitoring, allowing functions to interact with the network through a simplified interface without needing to understand deployment topology details
3Reliability
If time dissemination is ensured by a master computing node with redundancy management, then time synchronization is achieved, but system time jumps may occur particularly in the event of a failure
Solution Approach 1:
The system performs preliminary actions by establishing redundant time reference sources before failures occur. Each computing node maintains multiple potential time references and pre-configured switching logic, allowing seamless transition to backup time sources when the master node fails, preventing time jumps without requiring complex real-time failure analysis
Solution Approach 2:
The system changes the time synchronization parameter from a single master-node dependency to a multi-source reference model. By allowing nodes to switch between different time reference sources (master node, backup nodes, local oscillators) and adjusting synchronization parameters dynamically based on available references, the system maintains time continuity while preserving synchronization reliability
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method of communicating the state of a network comprising several computing nodes, implemented cyclically by each node, and comprising: - transmission (48), by a current node, to each other node, of a message comprising: - a local view corresponding to the list of nodes from which the current node has received a message, - a local synthesis corresponding to the intersection of the local view of the current node and the local views of the other nodes, - a final global synthesis corresponding to a local synthesis common to all nodes, - reception (46) of such a message emitted by each other node, - every other cycle, processing (50) of the last received data generating an associated message, the message emitted at each cycle corresponding to the last message generated by processing.