Autonomous Network Element Metric Discovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network metric collection methods, such as those using SNMP, require remote servers to discover and collect data from network elements, which can be time-consuming and burdensome, leading to increased network overhead and latency.
Innovation Solution
Network elements are empowered to autonomously discover and collect their own object instances and metric values, then transmit this data to a network information server, shifting the burden from the server to the element.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If remote servers discover and collect data from network elements using SNMP, then metrics can be collected from network elements, but the process is time-consuming and increases network overhead
Solution Approach 1:
The patent inverts the traditional SNMP architecture by enabling network elements to autonomously discover their own object instances and collect their own metric data, then transmit it to remote servers. This reversal eliminates the time-consuming discovery phase performed by remote servers and reduces network overhead, while maintaining complete metric collection capability.
Solution Approach 2:
Network elements perform the discovery of object instances and collection of metric values before being queried by remote servers. By pre-discovering and storing their own object instances locally, network elements eliminate the need for remote servers to perform time-consuming discovery operations, thereby reducing metric collection time while maintaining accuracy.
2Loss of information
If remote servers perform discovery and data collection, then comprehensive metrics can be gathered, but network overhead and server burden increase
Solution Approach 1:
Network elements autonomously perform self-discovery of their object instances and self-collection of metric data without requiring remote servers to initiate discovery operations. Each network element maintains a local repository of its object instances and actively pushes or makes available its metric data, thereby reducing network traffic and server processing burden while ensuring complete metric coverage.
Solution Approach 2:
The patent segments the metric collection function by distributing the discovery and data collection capabilities to individual network elements rather than concentrating these functions in remote servers. Each network element independently manages its own object instances and metric data, reducing the computational burden on servers and minimizing network overhead through localized processing.
3Productivity
If network elements autonomously discover and collect their own metrics, then collection time is reduced, but device complexity increases
Solution Approach 1:
The patent implements a universal object instance repository structure that can store multiple types of management information (scalars, tables, etc.) in a unified format. This multi-functional data structure allows network elements to handle diverse metric collection requirements using the same autonomous discovery and collection mechanism, improving efficiency without proportionally increasing complexity.
Data Source
AI summary
In an embodiment, a computer-implemented method collects metrics on a network element. The method includes receiving, on the network element, a specification of the objects on the network element to monitor. The network element queries an object data structure representing management information of the network element to identify instances of each of the specified objects. For respective instances identified, the network element queries the object data structure for metric values associated with the respective instance. Finally, data representing the instance and the associated metric value is transmitted from the network element to a network information server over a network.