Multi-class Data Transport via Dynamic Sub-flow Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network technologies face challenges in maintaining Quality of Service (QoS) during periods of network congestion or faults, as they often rely on all packets of a TCP flow having the same QoS class, which can lead to inadequate service when congestion occurs.
Innovation Solution
The method involves dividing a data flow into multiple sub-flows with different DiffServ classes, allowing for dynamic adjustment of the proportion of data units sent as higher or lower priority based on measured flow characteristics, such as congestion or buffer levels, to ensure optimal QoS by reallocating traffic between sub-flows.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If all packets of a TCP flow are assigned the same QoS class, then the network management is simplified, but the Quality of Service cannot be maintained during network congestion
Solution Approach 1:
The patent segments a single TCP flow into multiple sub-flows, where each sub-flow is assigned a different DiffServ class (e.g., EF, AF, BE). This allows different portions of the same data stream to receive different QoS treatments, enabling the system to maintain reliable QoS during congestion while managing network traffic through structured classification
Solution Approach 2:
The patent implements dynamic adjustment of sub-flow proportions based on network conditions. The sender monitors flow characteristics (such as throughput, delay, or loss) and dynamically reallocates traffic between different DiffServ classes, transitioning from static single-class assignment to adaptive multi-class distribution
2Reliability
If dynamic adjustment of sub-flow proportions is implemented, then the Quality of Service is optimized during congestion, but the system complexity increases
Solution Approach 1:
The patent employs feedback mechanisms where the sender monitors flow characteristics and uses this information to dynamically adjust the proportion of traffic in each sub-flow. This closed-loop control optimizes QoS by responding to actual network conditions while maintaining manageable complexity through algorithmic automation
Solution Approach 2:
The patent changes the DiffServ class parameter for different sub-flows within the same TCP connection. By manipulating the DSCP (Differentiated Services Code Point) values assigned to packets in different sub-flows, the system achieves flexible QoS optimization without requiring fundamental architectural changes
3Productivity
If multiple DiffServ classes are used for different sub-flows, then traffic prioritization during congestion is improved, but the network configuration complexity increases
Solution Approach 1:
The patent applies local quality by assigning different DiffServ treatments to different sub-flows of the same TCP connection based on their specific requirements or current network conditions. This localized differentiation allows efficient traffic prioritization where needed while maintaining standard handling for other traffic, reducing overall configuration complexity
Data Source
AI summary
Methods and apparatus for controlling the forwarding of a flow (12) of data units across a network from a sender node (11) to a receiver node (19) via at least one intermediate node (15), the data units each having one of a plurality of different types of class indication associated therewith, each type of class indication providing, to an intermediate node currently processing the data unit with which the class indication is associated, an indication denoting a particular class of treatment, selected from a plurality of different classes of treatment, according to which the data unit is to be processed by that intermediate node, such that different proportions of the data units of the flow are forwarded as different sub-flows each comprising data units having class indications of different types thereby providing indications denoting different classes of treatment.


