Parallel Network Topology Discovery Using Distributed Agents
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As networks grow in size, discovering their topology becomes increasingly challenging due to the time required for graph traversal and the number of probe messages needed, especially when using single discovery agents, which can be inefficient and time-consuming.
Innovation Solution
A method and system utilizing multiple node discovery agents operating in parallel, where nodes are partitioned using hash operations to distribute the discovery process, allowing each agent to send probe messages independently and collect connection information, thereby reducing overall discovery time and probe messages sent.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If multiple discovery agents operate in parallel to reduce discovery time, then the discovery time is reduced, but the number of probe messages sent increases
Solution Approach 1:
The patent divides the network discovery task into multiple independent segments by creating multiple discovery agents, each responsible for a specific subset of network nodes. This segmentation allows parallel execution of discovery operations, reducing overall discovery time while managing probe message throughput through controlled distribution of discovery tasks across agents
Solution Approach 2:
The patent introduces a discovery manager as an intermediary component that coordinates multiple discovery agents. The manager assigns specific node subsets to each agent, monitors probe message usage, and manages the parallel discovery process to optimize both discovery time and probe message efficiency through centralized control and coordination
2Measurement precision
If a single discovery agent performs graph traversal to ensure accurate topology discovery, then measurement precision is improved, but the discovery time increases significantly
Solution Approach 1:
The network node set is segmented into multiple disjoint subsets, with each discovery agent assigned to traverse and discover topology information for its specific subset. This segmentation enables parallel graph traversal operations while maintaining accurate topology discovery for each segment, which is then aggregated to form the complete network topology
Solution Approach 2:
The patent combines the results from multiple parallel discovery agents by aggregating their individual topology discoveries. The discovery manager merges the partial topology information from each agent into a unified complete network topology, maintaining measurement precision through systematic combination of parallel discovery results
3Productivity
If multiple discovery agents are used to traverse the network graph in parallel, then productivity is improved, but the coordination complexity increases
Solution Approach 1:
A central discovery manager acts as an intermediary to coordinate multiple discovery agents, assigning node subsets to agents and managing their parallel operations. This intermediary structure enables high productivity through parallel processing while controlling coordination complexity by centralizing management functions in a single coordinating component
Solution Approach 2:
The patent changes the operational parameters of discovery agents by assigning each agent a specific subset of nodes to discover, rather than having all agents traverse the entire network. This parameter change (node subset assignment) enables parallel productivity improvement while reducing coordination complexity through clear division of discovery responsibilities
Data Source
AI summary
A method, system and computer-usable medium are disclosed for discovering the topology of a network by using multiple discovery agents operating in parallel. A NodeList containing information about known nodes in a target network is received. The number ('K') of discovery agents to be used is determined, where 1<=K<=TotalNumberOfNodes). The nodes are partitioned into node groups, each of which has a group identifier respectively assigning it to a discovery agent 'M' where 1<=M<=K. A GraphList is created to hold information about known and discovered nodes. Each entry in the NodeList list is processed by its assigned discovery agent 'M', which sends probe messages to the target node to determine the node identifiers of discovered neighboring nodes. If the discovered nodes are not listed in the NodeList, then they are assigned to a node group for processing. If the connections between the target and the discovered neighboring node nodes are not listed in the GraphList, then they are added. Once all node and connection information is collected by the 'K' node discovery agents 'M' and listed in the GraphList, a network topology is generated.