Centralized Log Database for SDN Fault Tracing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In massively distributed complex systems like software-defined networks (SDNs), identifying faulty components after a failure occurs is challenging due to the vast number of interconnected components and the potential for log data to be overwritten, making it difficult to determine the root cause of failures in a timely manner.
Innovation Solution
A method is introduced to automatically trace back faults by subdividing the system into tiers, tagging reports with unique keys, and storing them in a central database for later querying, allowing for the identification of likely faulty processes and components, and taking preemptive measures such as restarting or isolating them.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If log data is stored locally in distributed systems, then storage capacity is distributed and accessible, but log data can be overwritten and lost making fault identification difficult
Solution Approach 1:
The patent merges distributed log storage with a centralized database system. Local agents continue to store logs distributed across the system, while simultaneously transmitting logs to a centralized database that preserves them indefinitely. This combination maintains the accessibility benefits of distributed storage while eliminating the data loss problem through centralized preservation.
Solution Approach 2:
The system performs preliminary actions by continuously collecting and archiving log data in the centralized database before faults occur. When a fault happens, the pre-stored logs are already available for immediate analysis, eliminating the need to wait for log generation after the fault occurs.
2Device complexity
If the system is subdivided into many tiers and components, then system complexity is managed and fault isolation is improved, but identifying the root cause among many components becomes more difficult
Solution Approach 1:
The patent implements feedback mechanisms where log data from all tiers includes contextual information about parent-child relationships between components. When a fault is detected at any tier, the system queries the centralized database for logs from related tiers using correlation IDs, automatically tracing the fault propagation path back to the root cause without manual investigation of each component.
Solution Approach 2:
The centralized database acts as an intermediary between distributed log sources and fault analysis. It stores, manages, and provides unified access to logs from all tiers, enabling cross-tier correlation and simplifying root cause identification by mediating the complex relationships between distributed components.
3Productivity
If traditional monitoring systems are used, then real-time monitoring is provided, but automatic identification of faulty components and predictive analysis is not achieved
Solution Approach 1:
The system enables self-service automation where the centralized database automatically queries, correlates, and analyzes logs from multiple tiers without human intervention. When faults are detected, the system autonomously traces fault propagation paths, identifies root causes, and even predicts potential failures by analyzing patterns in the stored log data.
Solution Approach 2:
The patent replaces manual fault investigation mechanisms with automated information processing. Instead of operators manually examining logs from multiple sources, the system uses automated querying and correlation algorithms that process log data, identify patterns, and pinpoint root causes computationally.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In one example, a controller device includes one or more network interfaces communicatively coupled to one or more devices of a virtual network, and a processor configured to determine, for the virtual network, a set of two or more related processes executed by respective devices in the virtual network, receive via the network interfaces data for the set of two or more related processes, and aggregate the data for the set of two or more related processes to form aggregated data for the set of two or more related processes.