Packet Queue Scheduling for Deterministic Delay and Jitter
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing scheduling methods fail to meet the requirements for deterministic delay and jitter in data packet transmission, particularly in large-scale IP networks, due to burst accumulation and contention between data packets.
Innovation Solution
A method for sending data packets that involves determining a second reference moment based on time information carried in the packet, allowing the packet to be added to a target queue in a second queue system, ensuring deterministic delay and jitter by accounting for the remaining processing time on both the first and second network devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing scheduling methods (weighted fair queue, CAF) are used, then network devices can forward data packets, but deterministic delay and jitter requirements cannot be met due to burst accumulation and contention
Solution Approach 1:
The patent applies preliminary action by calculating and embedding the first remaining processing time in the data packet before transmission. This allows the second network device to determine the second reference moment in advance, ensuring deterministic delay requirements are met before actual packet forwarding occurs. The scheduling decision is prepared beforehand based on time information carried in the packet.
Solution Approach 2:
The patent introduces time information (first remaining processing time, first reference moment, first output moment) as an intermediary carrier between network devices. This intermediary data structure enables the second network device to calculate the second reference moment and make scheduling decisions without requiring complex real-time coordination, thus achieving deterministic delay while controlling system complexity.
2Manufacturing precision
If the second network device calculates the second reference moment based on multiple time parameters, then deterministic delay can be ensured, but calculation complexity increases
Solution Approach 1:
The first network device performs preliminary calculations to determine the first remaining processing time, first reference moment, and first output moment before transmitting the data packet. By preparing these time parameters in advance and embedding them in the packet, the second network device only needs to perform straightforward calculations using the provided values, reducing its computational burden while maintaining precise delay control.
3Productivity
If data packets are scheduled based on real-time conditions, then network utilization is improved, but deterministic delay guarantees become difficult to maintain
Solution Approach 1:
The patent resolves this contradiction by making the scheduling decision in advance through the second reference moment calculation. The second network device determines which queue to use and when to forward the packet before actual transmission occurs. This preliminary scheduling locks in the delay bound guarantee while still allowing the network to handle multiple packets efficiently, as the scheduling decisions are made autonomously based on pre-calculated time information.
Data Source
AI summary
Embodiments of this application provide a method for sending a data packet. The method includes: A second network device may determine a remaining processing time of a data packet on a first network device based on a predefined first delay. When the remaining processing time is greater than 0, a moment at which the data packet enters a queue on the second network device serves as a start moment, and after a period of time, the remaining processing time ends at a second reference moment corresponding to the data packet. The remaining processing time is consumed on the second network device.