Session-Based Load Balancing for Network Link Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network systems face challenges in managing network traffic across multiple links, leading to reduced throughput, increased latency, and network traffic loss due to interference and asymmetrical link capacities, as they distribute session packets without considering the specific session or link capacity, causing degradation across all links.
Innovation Solution
Implementing session-based load balancing, where network traffic is distributed based on the session to which packets belong, using link weights determined by the capacity ratio of individual network links to optimize transmission across multiple links, thereby minimizing interference and ensuring reliable data transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If session packets are distributed across multiple network links without considering link capacity, then network traffic can be transmitted in parallel improving throughput, but interference and asymmetrical link capacities cause degradation across all links
Solution Approach 1:
The patent applies local quality by assigning different treatment to different sessions based on their characteristics. Sessions are classified into first sessions (with packet loss tolerance) and second sessions (without packet loss tolerance), and are routed to different network links accordingly. This allows the system to optimize each session's transmission based on its specific requirements rather than treating all traffic uniformly, thereby maintaining high throughput while preventing degradation of critical sessions.
Solution Approach 2:
The patent segments network traffic into different categories (first sessions and second sessions) based on packet loss tolerance requirements. By dividing the traffic stream into distinct segments with different quality requirements, the system can apply link weightings and routing decisions tailored to each segment, enabling parallel transmission across multiple links while protecting critical sessions from interference.
2Productivity
If link weights are not considered in packet assignment, then packet distribution is simple, but throughput is reduced and latency increases due to inefficient link utilization
Solution Approach 1:
The patent introduces link weight as a new parameter that characterizes the capacity or quality of each network link. By determining and utilizing link weights in the packet assignment process, the system optimizes throughput by directing traffic to appropriate links based on their capacities. Although this adds some complexity to the device, the weight determination can be performed through standard link negotiation protocols, making the complexity manageable while achieving significant productivity improvements.
3Reliability
If all session packets are treated equally, then processing is simplified, but interference affects all links uniformly causing increased latency and packet loss
Solution Approach 1:
The patent applies local quality by assigning different treatment to different sessions based on their characteristics. Sessions are classified into first sessions (with packet loss tolerance) and second sessions (without packet loss tolerance), and are routed to different network links accordingly. This allows the system to optimize each session's transmission based on its specific requirements rather than treating all traffic uniformly, thereby maintaining high throughput while preventing degradation of critical sessions.
Solution Approach 2:
The patent introduces session classification as an intermediary mechanism between the packet source and the network link selection process. By classifying sessions into different categories based on their packet loss tolerance requirements, the system creates an intermediate layer that mediates between simple packet processing and complex link optimization, enabling differentiated service without requiring complete redesign of the routing architecture.
Data Source
AI summary
A method includes obtaining a first session packet of multiple session packets belonging to a session, where the session may be queued to be transmitted between a first network device and a second network device. The method also includes assigning the first session packet to a first network link based on a link weight determined for the first network link between the first network device and the second network device. The method further includes queuing the first session packet to be transmitted from the first network device to the second network device via the first network link. The method also includes obtaining a second session packet of the plurality of session packets. The method further includes assigning the second session packet to the first network link.


