Decoupled Sub-Flow Synchronization for Multipath Data Transfer
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multipath transmission control protocol (MPTCP) schemes, such as those using the minimum round trip time (MinRTT) algorithm, often result in sub-optimal performance due to non-simultaneous completion of packet sub-flows over multiple network paths, leading to increased download times for mobile devices under fluctuating WiFi/cellular network conditions.
Innovation Solution
The approach involves decoupling sub-flows to optimize data transfer by selecting stop times for each sub-flow to ensure simultaneous completion at the receiver, using calculated timing offsets based on network latency and bandwidth differences between paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional MPTCP schemes (e.g., MinRTT algorithm) are used to distribute data over multiple network paths, then data can be transmitted in parallel, but sub-flows complete at different times leading to increased download times
Solution Approach 1:
The patent calculates and determines the completion times of all sub-flows before data transmission begins. By preliminarily analyzing network path characteristics and predicting sub-flow completion times, the system can pre-plan the data distribution strategy to ensure all sub-flows complete simultaneously, avoiding the need for waiting periods that occur in conventional approaches.
Solution Approach 2:
The patent dynamically adjusts the data distribution parameters across sub-flows based on calculated completion times. By changing how data is allocated to different sub-flows (rather than using fixed algorithms like MinRTT), the system optimizes the transmission to achieve simultaneous completion of all sub-flows, thereby reducing overall download time.
2Power
If data is distributed across multiple sub-flows simultaneously, then parallel transmission is achieved, but non-simultaneous completion occurs causing inefficiency
Solution Approach 1:
The patent employs a feedback mechanism where the system continuously monitors and calculates the completion times of sub-flows during transmission. Based on this feedback information, the multipath scheduler adjusts data distribution in real-time to ensure that all sub-flows complete simultaneously, maximizing effective data delivery efficiency while maintaining high transmission throughput.
Solution Approach 2:
The patent transforms the static data distribution approach into a dynamic one. Instead of fixed allocation rules, the system continuously adapts data distribution across sub-flows based on real-time network conditions and predicted completion times, allowing the transmission system to remain optimal despite fluctuations in network performance.
Data Source
AI summary
A request for a chunk of data is received from an application executing on a remote device. The chunk of data is split into a plurality of packets including a first packet and a second packet. A first sub-flow is selected to transport the first packet to the remote device and a second sub-flow is selected to transport the second packet. The first sub-flow is decoupled from the second sub-flow. A first stop time is selected for the first sub-flow and a second stop time is selected for the second sub-flow. The first stop time determines when the first sub-flow will complete at the remote device and the second stop time determines when the second sub-flow will complete at the remote device. The first stop time and the second stop time are selected so that the first sub-flow and the second sub-flow will complete simultaneously at the remote device.