Network Adapter Load Balancing With End-to-End Congestion Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional congestion control algorithms focus on managing traffic flow between endpoints without considering local load on output ports, while load balancing algorithms on switches do not control when packets arrive, leading to inefficient network utilization and potential congestion.
Innovation Solution
A combined mechanism that integrates congestion control and load balancing to optimize traffic spreading across multiple local output ports by considering both local and end-to-end load, using E2E congestion rates to determine the best time and port for packet transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional congestion control algorithms are used to manage traffic flow between endpoints, then traffic flow management is achieved, but local load on output ports is not considered leading to inefficient network utilization
Solution Approach 1:
The patent merges congestion control and load balancing into a unified mechanism. The network adapter integrates both functions by determining E2E congestion rates and local port states simultaneously, then using both factors together to select output ports and determine transmission timing. This combination resolves the contradiction by achieving both efficient network utilization and local load management without requiring separate complex systems.
Solution Approach 2:
The network adapter performs multiple functions: it monitors E2E congestion rates, tracks local port states, selects optimal output ports, and determines transmission timing all within a single congestion control mechanism. This multi-functional approach improves network utilization while avoiding the need for additional dedicated load balancing hardware or software layers.
2Productivity
If load balancing algorithms on switches are used to spread traffic across multiple routes, then traffic spreading is achieved, but control over packet transmission timing is lost
Solution Approach 1:
The patent combines load balancing port selection with congestion-controlled transmission timing in a single unified decision process. The network adapter selects output ports based on both local load and E2E congestion, then determines transmission timing based on the same congestion rate information. This integration maintains transmission control while achieving effective traffic spreading.
3Productivity
If traffic is concentrated on fewer paths, then simpler routing is achieved, but network congestion increases and utilization decreases
Solution Approach 1:
The patent applies local quality by making routing decisions specific to each packet based on current congestion conditions. The network adapter evaluates E2E congestion rates and local port states for each packet, then selects the most appropriate output port individually. This per-packet adaptive approach optimizes network utilization while keeping the routing mechanism manageable through localized decision-making.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Technologies for optimizing the spreading of traffic across multiple local output ports while considering both local load and end-to-end (E2E) load are described. One device has multiple outgoing ports and a network adapter that determines, for a first flow of packets, a first end-to-end (E2E) congestion rate of at least some of the outgoing ports. The network adapter determines a port state of at least some of the outgoing ports. The network adapter receives a first packet associated with the first flow of packets. The network adapter determines, using a first desired rate for the first flow, the first E2E congestion rates, and the port states, i) a first time at which the first packet is to be transmitted and ii) a first outgoing port on which the first packet is to be transmitted. The first packet is sent on the first outgoing port at the first time.