Deadline-Aware TSN Flow Scheduling with Distributed Latency Budgets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing TSN network scheduling methods face challenges in scalability, computational complexity, and inefficiency, particularly in dynamic industrial environments with increased network diameters and varying network configurations, leading to excessive configuration times and reduced performance.
Innovation Solution
A method for scheduling data packets in a deterministic network that involves determining latency budgets for intermediate nodes, selecting transmission periods within transmission windows based on these budgets, and using an Active and Accurate Latency Prediction (AALP) scheme to ensure timely delivery while optimizing network resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If centralized network controller computes and distributes gate control lists for scheduled traffic, then end-to-end latency can be bounded, but configuration computation time increases exponentially with network size
Solution Approach 1:
The patent divides the centralized scheduling problem into distributed per-node scheduling decisions. Each node independently determines transmission times based on locally received latency budgets, eliminating the need for centralized gate control list computation and distribution across the entire network.
Solution Approach 2:
The patent introduces a new dimension of latency budget distribution, where the centralized controller only needs to compute and distribute latency budget parameters to nodes rather than computing complete schedules. This transforms the problem from centralized schedule computation to distributed execution based on simple local parameters.
2Reliability
If gate control lists are distributed to all switches for scheduled traffic, then deterministic latency can be guaranteed, but network complexity and configuration overhead increase
Solution Approach 1:
The patent extracts the complex gate control list computation from individual switches and centralizes only the latency budget parameter computation. Each switch then independently determines its transmission schedule based on simple latency budget rules, eliminating the need for complex distributed gate control list management.
Solution Approach 2:
Each network node autonomously determines its own transmission schedule based on received latency budgets without requiring centralized control or coordination with other nodes. This self-service approach simplifies network configuration while maintaining deterministic latency guarantees.
3Reliability
If transmission time is selected as earliest free period, then latency requirement is satisfied, but network resource utilization decreases
Solution Approach 1:
The patent introduces dynamic selection of transmission periods within the latency budget constraint. Nodes can adaptively choose transmission times based on current network conditions and latency budgets, rather than always selecting the earliest possible time, thereby improving resource utilization while meeting latency requirements.
Solution Approach 2:
The patent changes the parameter selection from fixed earliest free period to flexible period selection within latency budget constraints. This allows nodes to optimize transmission timing based on latency budgets and network conditions, improving overall resource utilization while maintaining latency guarantees.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for scheduling transmission of a data packet from a talker node to a listener node via one or more intermediate nodes with a maximum allowed end-to-end, E2E, latency. The method comprises for each of the one or more intermediate nodes: determining a latency budget such that the sum of the latency budgets for all the intermediate nodes is less than or equal to the maximum allowed E2E latency, and scheduling a time for transmission of the data packet from the intermediate node to a next node based on the latency budget. A computer- readable medium, a device for scheduling a transmission of a data packet from a talker node to a listener node, and a device for operating as an intermediate node in a network are also disclosed. The method allows scheduling a data packet in a way that a latency of the data packet will normally not exceed the maximum allowed E2E latency.