Load Distribution Table for Network Flow Balancing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing load distribution techniques, such as ECMP and hashing algorithms, struggle to evenly distribute traffic across equal cost paths in asymmetrical network topologies and fail to adapt dynamically during network failures, leading to imbalances and correlation issues.

Innovation Solution

A method using a load distribution table populated with next hop discriminators, where each discriminator uniquely identifies a next hop path, and a content addressable memory (CAM) is used to determine the forwarding path for protocol data units (PDUs), allowing for a more even distribution and minimizing correlation effects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If hashing algorithms are used for load spreading in ECMP, then traffic can be distributed across multiple paths, but load distribution becomes uneven in asymmetrical topologies and correlation issues arise

Engineering Contradiction:

Improveload distribution capabilityVSAvoidload distribution uniformity

Core Design Contradiction:

Adaptability or versatilityVSManufacturing precision

Solution Approach 1:

The patent pre-calculates and stores next hop discriminators in a lookup table before actual traffic forwarding occurs. This preliminary action allows the system to avoid correlation issues and achieve uniform load distribution without performing complex hashing operations during real-time traffic flow, directly resolving the contradiction between distribution capability and uniformity.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces the mechanical hashing algorithm with a lookup table approach using next hop discriminators. This substitution eliminates the correlation problems inherent in hashing while maintaining the ability to distribute traffic across multiple paths, achieving both adaptability and precision in load distribution.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If hashing algorithms are used for path selection, then flows can be grouped together, but implementation complexity increases and forwarding delays occur

Engineering Contradiction:

Improveflow ordering preservationVSAvoidhashing algorithm complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent replaces complex hashing algorithms with a simpler lookup table mechanism that uses next hop discriminators. This substitution maintains flow ordering reliability while significantly reducing implementation complexity and forwarding delays, as the lookup operation is computationally lighter than hashing.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent uses simple, easily computable next hop discriminators instead of complex hashing functions. These discriminators are derived from basic packet header fields and can be quickly generated and looked up, providing reliable flow ordering with minimal computational overhead and complexity.

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

3Productivity

If ECMP is implemented with fixed hashing schemes, then initial load distribution can be achieved, but the system cannot adapt dynamically to network failures

Engineering Contradiction:

Improvetraffic forwarding efficiencyVSAvoidnetwork failure adaptability

Core Design Contradiction:

ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent implements a dynamic lookup table that can be updated in response to network changes and failures. This allows the system to adapt its load distribution strategy dynamically, switching between different next hop discriminators based on current network conditions, thereby maintaining both forwarding efficiency and adaptability.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent pre-configures multiple next hop discriminators in the lookup table, allowing the system to quickly switch between different path sets when network failures occur. This preliminary preparation enables rapid adaptation without disrupting ongoing traffic flow, maintaining productivity while improving failure resilience.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS8467298B2Applying a table-lookup approach to load spreading in forwarding data in a network

Publication Date: 2013.06.18 TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)

AI summary

A method provides load distribution that selects one of a set equal cost paths for each flow that minimizes imbalance in the distribution of flows across the set of equal cost paths. The method includes calculating the set of equal cost paths to each destination node, generating a set of next hop discriminators, each next hop discriminator to uniquely identify one of the equal cost next hops, populating the load distribution table by storing, according to a distribution mechanism, in each load distribution table entry one of the next hop discriminators from the set of next hop discriminators, receiving a first PDU from a first flow and performing a lookup of one of the next hop discriminators in the load distribution table by using values at the bit locations of the first PDU directly as an index into a CAM.