Predictive Scheduler for DOCSIS Upstream Latency Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional DOCSIS upstream scheduling in Hybrid Fiber-Coaxial networks experiences high latency, exceeding three times the propagation delay, which is problematic for delay-sensitive services, especially in Remote-PHY deployments where distances can exceed 1,250 miles, leading to latency of over 30 ms.

Innovation Solution

A predictive scheduler is introduced that anticipates incoming data traffic and issues grants before data arrival, reducing upstream latency to one propagation delay, and incorporates machine learning for real-time and long-term traffic dynamics prediction to optimize grant allocation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If conventional request-grant scheduling is used in DOCSIS upstream, then the system maintains simplicity and compatibility with existing protocols, but upstream latency exceeds three times the propagation delay, which is unacceptable for delay-sensitive services

Engineering Contradiction:
Improveupstream latencyVSAvoidscheduling mechanism complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

The predictive scheduler performs preliminary actions by analyzing traffic patterns and predicting future bandwidth requests before they actually occur. The system uses machine learning models to forecast traffic dynamics and proactively allocates grants in advance, thereby reducing upstream latency without requiring complex real-time negotiations during data transmission

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements feedback mechanisms where the predictive scheduler continuously monitors actual traffic patterns compared to predictions, and uses this feedback to refine its machine learning models. This closed-loop approach allows the system to adapt to changing traffic conditions while maintaining the simplified grant-based structure, resolving the contradiction between latency reduction and system complexity

Inventive Principle:
Principle #23Feedback

2Adaptability or versatility

If the unsolicited grant size is fixed, then the bandwidth allocation is simple to manage, but it cannot adapt to varying traffic dynamics, leading to inefficient channel utilization

Engineering Contradiction:
Improvebandwidth allocation adaptabilityVSAvoidbandwidth allocation complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system transforms the static unsolicited grant size into a dynamic parameter that adapts to traffic conditions. The predictive scheduler uses machine learning to continuously adjust grant sizes based on predicted traffic patterns, allowing the bandwidth allocation to be flexible and responsive while maintaining systematic control through automated decision-making algorithms

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The invention changes the parameter of grant size from a fixed value to a dynamically adjustable parameter based on traffic predictions. By using machine learning models to determine optimal grant sizes under different traffic conditions, the system achieves adaptability without requiring complex manual configuration or real-time negotiation protocols

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3688938B1Predictive scheduler
Publication Date: 2022.05.18 CISCO TECHNOLOGY INC
  • EP3688938B1 patent drawingFigure 1
  • EP3688938B1 patent drawingFigure 2~3
  • EP3688938B1 patent drawingFigure 4~5

AI summary

Predictive scheduling may be provided. First, a first device may identify when a service flow is expected to become active. The first device may estimate an initial traffic profile in response to identifying when the service flow is expected to become active. The first device may then grant allocation based on the initial traffic profile of the service flow. Next, the first device may collect feedback to later update the traffic profile estimate. The first device may then update the traffic profile estimate.