Scaled Parallel Hashing for Weighted Balancing in Uneven Networks

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current weighted load balancing methods in multistage networks, such as data center networks, are complex to implement and lack scalability, particularly in networks with uneven inter-stage topologies, leading to inefficient traffic distribution across unevenly distributed links.

Innovation Solution

A method for weighted load balancing using scaled parallel hashing, where independent hash values are generated for each egress port based on packet fields, and a scaling factor is applied to achieve balanced traffic distribution by selecting egress ports based on weighted hash values, allowing for efficient traffic routing even in networks with varying link capacities.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If current weighted load balancing methods are used in multistage networks, then traffic routing can be performed, but the implementation becomes complicated and lacks scalability in large networks with uneven inter-stage topologies

Engineering Contradiction:
Improvetraffic distribution adaptabilityVSAvoidload balancing implementation complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent transforms the complex weighted load balancing problem into a simpler form by changing the parameter representation. Instead of using traditional weighted algorithms that require complex configuration and calculation, the invention uses hash function scaling factors that can be directly applied to hash values. This parameter transformation simplifies the implementation while maintaining adaptability to uneven topologies, as the scaling factors can be easily adjusted to match network conditions without complicating the overall algorithm.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent segments the load balancing function into independent hash function operations for each egress port. Each port has its own scaled hash function that operates independently, allowing parallel computation and simplifying the overall system architecture. This segmentation enables each port to be configured and optimized separately, improving scalability in large networks while reducing the complexity of managing centralized weight configurations.

Inventive Principle:
Principle #1Segmentation

2Productivity

If traditional load balancing is used in networks with uneven inter-stage topologies, then routing can occur, but traffic distribution becomes inefficient with link underutilization

Engineering Contradiction:
Improvetraffic distribution efficiencyVSAvoidlink utilization
Core Design Contradiction:
ProductivityVSLoss of energy

Solution Approach 1:

The patent applies local quality by allowing each egress port to have its own customized hash function with specific scaling factors tailored to the local network topology and link capacities. Instead of using a uniform load balancing approach across all ports, the invention enables each port to optimize its traffic distribution based on local conditions, thereby improving overall traffic distribution efficiency and preventing link underutilization in uneven topologies.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent introduces dynamics by making the hash function scaling factors adjustable and adaptable to changing network conditions. The scaling factors can be dynamically modified to reflect changes in link capacities, traffic patterns, or topology variations, allowing the load balancing system to continuously optimize traffic distribution efficiency and maintain high link utilization as network conditions evolve.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS12363190B1Weighted load balancing using scaled parallel hashing
Publication Date: 2025.07.15 GOOGLE LLC
  • US12363190B1 patent drawing
  • US12363190B1 patent drawing
  • US12363190B1 patent drawing

AI summary

A method for weighted data traffic routing can include receiving a data packet at data switch, where the data switch includes a plurality of egress ports. The method can also include, for each of the egress ports, generating an independent hash value based on one or more fields of the data packet and generating a weighted hash value by scaling the hash value using a scaling factor. The scaling factor can be based on at least two traffic routing weights of a plurality of respective traffic routing weights associated with the plurality of egress ports. The method can further include selecting an egress port of the plurality of egress ports based on the weighted hash value for each of the egress ports and transmitting the data packet using the selected egress port.