Computation-Aware Data Packet Routing for Faster Result Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data packet routing methods solely rely on current network status to determine the next-hop routing node, which may not necessarily be optimal for executing computation tasks, leading to prolonged wait times for computation results.
Innovation Solution
A data packet routing method that considers both current network status and computing performance of each node capable of executing a computation task, determining a target node based on computing performance and link status to ensure rapid task completion and result feedback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If routing is determined solely based on current network status, then network status optimization is improved, but computation task execution efficiency deteriorates
Solution Approach 1:
The patent changes the routing decision parameters from solely network status parameters to a composite parameter set that includes both network status and computing performance metrics. This allows the system to optimize for both network conditions and computation capabilities simultaneously, resolving the contradiction between network status optimization and computation task execution efficiency.
Solution Approach 2:
The patent introduces a new dimension to the routing decision process by adding computing performance as an independent evaluation criterion alongside network status. This multi-dimensional approach enables the system to select target nodes that satisfy both network optimality and computation efficiency requirements, rather than optimizing for a single dimension.
2Reliability
If any computing node capable of executing the task is selected, then task completion is guaranteed, but wait time for results increases
Solution Approach 1:
The patent changes the node selection criteria from a binary capability check (can execute or cannot execute) to a performance-based ranking system that considers computing performance metrics. This allows the system to prioritize nodes with better computation capabilities while still ensuring task completion, thereby reducing wait time without sacrificing reliability.
Solution Approach 2:
The patent implements a feedback mechanism where computing performance information is collected and used to guide routing decisions. By continuously gathering and utilizing performance data from various nodes, the system can dynamically select the most efficient target nodes, reducing computation wait times while maintaining task completion guarantees.
Data Source
AI summary
The present disclosure relates to data packet routing methods and apparatuses. One example method includes when a data packet of a computation task type is received, determining a first computation task type corresponding to the data packet, determining, based on a pre-obtained first correspondence between a computation task type, other nodes, and computing performance, at least one other node corresponding to the first computation task type and computing performance corresponding to the at least one other node, determining a target node from the at least one other node based on the computing performance corresponding to the at least one other node and based on a link status between a local node and each of the at least one other node, determining an address of the target node as a destination address of the data packet, and forwarding the data packet based on the destination address.