Physical Medium Dependent Layer Bonding for Modem Throughput

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing bonding techniques for physical lines at modems face limitations such as differential delay compensation requirements, slow addition of new lines, restricted operation modes, and increased bonding overhead, which affect throughput and flexibility.

Innovation Solution

Bonding of protocol stacks at the Physical Medium Dependent layer allows for dynamic switching between modes, reducing the need for large buffers and enabling flexible operation, including time-synchronized communication across multiple physical lines.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If bonding is implemented above the physical layer, then bonding flexibility is improved, but differential delay compensation buffers are required increasing device complexity

Engineering Contradiction:
Improvebonding flexibilityVSAvoiddifferential delay compensation buffers
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent inverts the conventional bonding approach by implementing bonding at the Physical Medium Dependent layer instead of above the physical layer. This reversal eliminates the need for differential delay compensation buffers while maintaining bonding flexibility, as the PMD layer operates at a lower level where such compensation is not required.

Inventive Principle:
Principle #13The other way round (Inversion)

Solution Approach 2:

The patent changes the dimensional level of bonding operation from the data link layer or higher (traditional approach) down to the Physical Medium Dependent layer. This dimensional shift in the protocol stack enables bonding without requiring large compensation buffers, fundamentally changing how bonding is achieved.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Adaptability or versatility

If another physical line is added to a bonding group in existing implementations, then bonding capacity is improved, but switching speed deteriorates

Engineering Contradiction:
Improvebonding capacityVSAvoidswitching speed
Core Design Contradiction:
Adaptability or versatilityVSSpeed

Solution Approach 1:

The patent implements dynamic bonding at the PMD layer that allows rapid addition and removal of physical lines from bonding groups during Showtime operation. The PMD layer's position in the protocol stack enables fast reconfiguration without the delays associated with upper-layer bonding implementations.

Inventive Principle:
Principle #15Dynamics

3Ease of operation

If bonding overhead is introduced for fragmentation and sequence numbers, then data distribution is improved, but throughput deteriorates

Engineering Contradiction:
Improvedata distributionVSAvoidthroughput
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent extracts and removes the bonding overhead mechanisms (fragmentation and sequence numbers) from the bonding implementation. By operating at the PMD layer, the patent achieves data distribution across bonded physical lines without requiring these overhead elements, thereby preserving throughput.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentUS10476995B2Physical medium dependent layer bonding
Publication Date: 2019.11.12 INTEL GERMANY GMBH & CO KG
  • US10476995B2 patent drawing
  • US10476995B2 patent drawing
  • US10476995B2 patent drawing

AI summary

A first protocol stack for communication on a first physical line is implemented. At least parts of a second protocol stack for communication on a second physical line are implemented. The first protocol stack and the second protocol stack are bonded at the Physical Medium Dependent layer of the first protocol stack and the Physical Medium Dependent layer of the second protocol stack (172). In some scenarios, the bonding may be at an upper edge of the Physical Medium Dependent layer, i.e., at the δ interface.