Packet Scheduling Algorithm for Heterogeneous Network Paths
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for transmitting data between source and destination nodes via multiple network paths fail to minimize in-order delivery delay due to unpredictable, time-varying transmission delays, leading to packet re-ordering and increased processing delays.
Innovation Solution
A method that models transmission delays as random variables to optimize packet scheduling across paths, using a scheduling algorithm that minimizes expected in-order delivery delay by selecting the most suitable path for each packet based on its variability and loss rate, and employs a pruning-tree algorithm to reduce computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If packets are transmitted across multiple paths with different transmission delays, then throughput and resilience are improved, but packet re-ordering occurs causing increased delivery delay
Solution Approach 1:
The patent applies dynamics by making the scheduling decision adaptive rather than static. The source node dynamically selects which path to use for each packet based on real-time or near-real-time observations of path conditions (transmission delays, packet losses). This dynamic adaptation allows the system to exploit multiple paths for improved throughput while minimizing re-ordering delays by choosing the most appropriate path for each packet based on current network state.
Solution Approach 2:
The patent changes the parameter of path selection from a fixed deterministic choice to a variable decision based on observed network parameters. By monitoring transmission delays and packet loss rates on each path and using these parameters to guide scheduling decisions, the system optimizes the trade-off between throughput (using multiple paths) and in-order delivery (selecting paths with favorable characteristics for each packet).
2Loss of time
If deterministic scheduling is used to ensure in-order packet arrival, then re-ordering delay is minimized, but the system cannot adapt to unpredictable, time-varying transmission delays
Solution Approach 1:
The patent implements feedback by having the source node observe transmission delays and packet loss rates on each path, then use this feedback information to make informed scheduling decisions. The system continuously monitors path performance and adapts its packet transmission strategy based on these observations, allowing it to respond to time-varying network conditions while maintaining relatively low re-ordering delays through intelligent path selection.
Solution Approach 2:
The scheduling mechanism transitions from a static deterministic approach to a dynamic adaptive approach. The source node dynamically adjusts its path selection based on observed network conditions, making the system versatile enough to handle unpredictable transmission delays while still minimizing re-ordering delays through intelligent, condition-based decision making.
Data Source
AI summary
A method is disclosed for transmitting data between a source node and destination node connected via multiple paths of a heterogeneous network, at least one of the paths delivering packets with a non-deterministic delivery time. Data is divided into frames, each frame comprising a number of packets, where processing by the destination node of an information packet p is conditional on receipt of the data for any information packet i where i<p. A number of sequential transmission slots s, each for transmitting a given packet p of a frame, are allocated for each network path k. A set D of possible assignments of packets to transmission slots for a frame are provided. The method comprises determining an assignment x with minimum in-order delivery delay of said packets based on expected arrival times of packets to their destination and expected reordering delay for packets. Packets are then transmitted in accordance with the determined assignment x.