Network Node Controller Latency Optimization via In-Band Telemetry

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing network solutions fail to effectively address congestion in egress port queues of network nodes, leading to latency issues, as they lack real-time monitoring and adaptive rerouting capabilities.

Innovation Solution

A network node with a controller that determines congestion levels in egress port queues through telemetry packets and reroutes data flows to alternative queues based on congestion thresholds, using Differentiated Services Code Point (DSCP) fields in packet headers to manage priority levels.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If traditional networking equipment is used without real-time monitoring, then device complexity is reduced, but network latency increases due to undetected queue congestion

Engineering Contradiction:
Improvenetwork latencyVSAvoidnetwork monitoring system complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

The patent implements feedback by having network nodes continuously monitor queue congestion levels and send telemetry data back to a centralized controller. The controller processes this feedback and dynamically adjusts data flow routing decisions, creating a closed-loop system that reduces latency through real-time congestion detection and response

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent introduces a centralized controller as an intermediary between network nodes and the routing decision-making process. This intermediary collects telemetry data from multiple nodes, processes congestion information centrally, and coordinates rerouting decisions across the network, enabling complex latency optimization without increasing complexity at individual node levels

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If data flows are statically routed through fixed queues, then device complexity is minimized, but network adaptability decreases when congestion occurs

Engineering Contradiction:
Improvenetwork routing adaptabilityVSAvoidqueue management complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent applies dynamics by transitioning from static queue routing to dynamic queue selection. Network nodes continuously monitor congestion levels and adaptively reroute data flows between different queues based on real-time conditions. The system dynamically adjusts routing decisions without requiring complex manual configuration or hardware changes

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes routing parameters dynamically by modifying which queue is selected for data flow based on congestion thresholds. When congestion is detected in one queue, the system changes the routing parameter to direct traffic through alternative queues, enabling adaptability through parameter adjustment rather than structural complexity

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If congestion is not monitored in real-time, then measurement precision is reduced, but loss of time increases due to delayed congestion detection

Engineering Contradiction:
Improvecongestion detection precisionVSAvoidcongestion detection time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent implements preliminary action by continuously monitoring queue congestion levels before they reach critical thresholds. The system proactively detects congestion conditions and triggers rerouting decisions in advance, preventing severe latency issues rather than reacting after problems occur

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent uses feedback mechanisms where network nodes continuously report queue status to the centralized controller, enabling precise real-time measurement of congestion levels. This feedback loop ensures accurate detection and immediate response to congestion conditions, minimizing both detection time and latency impact

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS20210306267A1Optimized network latency using in-band telemetry
Publication Date: 2021.09.30 LENOVO GLOBAL TECHNOLOGIES SWITZERLAND INTERNATIONAL GMBH
  • US20210306267A1 patent drawing
  • US20210306267A1 patent drawing
  • US20210306267A1 patent drawing

AI summary

An apparatus for reporting node congestion of a queue of an egress port and changing to a different queue is disclosed. The apparatus includes a network node that includes a controller. The controller is configured to determine a level of congestion of a designated queue of an egress port of the network node in a communication pathway between a sending host sending data packets to a receiving host in response to receiving a telemetry packet seeking telemetry data for packets being transmitted between the sending host and the receiving host. The designated queue is designated for queuing data packets for the egress port. The controller is configured to add the level of congestion to the telemetry packet in response to determining that the level of congestion indicates that the designated queue is congested and to transmit the telemetry packet to a next destination on the communication pathway.