Fabric Control Protocol Extensions for Data Center Congestion

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current data center networks face challenges in efficiently managing congestion within the network fabric, leading to potential bottlenecks and reduced performance.

Innovation Solution

The proposed fabric control protocol (FCP) and its reliable extensions (rFCP) implement congestion control mechanisms that dynamically determine the degree of congestion at egress interfaces within the network fabric, allowing for the modification of send window sizes to mitigate congestion.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If packets are always forwarded along a single path through the network fabric, then routing simplicity is maintained, but network performance and reliability deteriorate due to congestion bottlenecks

Engineering Contradiction:

Improverouting complexityVSAvoidnetwork performance

Core Design Contradiction:

Device complexityVSProductivity

Solution Approach 1:

The patent segments the network fabric into multiple parallel data paths between source and destination servers. Instead of using a single routing path, the system divides traffic into multiple streams that can traverse different physical or logical paths simultaneously, thereby distributing congestion and improving overall network performance while maintaining manageable routing complexity through structured path selection

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent implements dynamic path selection and packet spraying mechanisms that adapt to real-time network conditions. The system dynamically determines the degree of congestion at egress interfaces and modifies send window sizes accordingly, allowing the network to flexibly respond to changing traffic patterns and congestion conditions rather than relying on static single-path routing

Inventive Principle:

Principle #15Dynamics

2Productivity

If congestion control mechanisms dynamically adjust send window sizes, then network performance improves, but protocol complexity increases

Engineering Contradiction:

Improvenetwork performanceVSAvoidprotocol complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent implements feedback-based congestion control where the destination server monitors the degree of congestion at its egress interface and communicates this information back to the source server. The source server uses this feedback to dynamically modify send window sizes, creating a closed-loop control system that automatically adapts to network conditions. This feedback mechanism improves network performance while keeping protocol complexity manageable through standardized congestion indicator messages

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The congestion control mechanism operates autonomously without requiring external intervention. The protocol enables the network endpoints to self-regulate traffic flow by monitoring congestion conditions and automatically adjusting transmission parameters, eliminating the need for manual configuration or complex centralized control while maintaining optimal network performance

Inventive Principle:

Principle #25Self-service

3Reliability

If multiple parallel data paths are used for packet spraying, then network reliability improves, but congestion management difficulty increases

Engineering Contradiction:

Improvenetwork reliabilityVSAvoidcongestion detection difficulty

Core Design Contradiction:

ReliabilityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent implements a universal congestion detection and control mechanism that operates across all multiple parallel data paths simultaneously. The congestion control protocol provides multi-functional capabilities including congestion detection, path selection, and send window management through a unified framework, making the system scalable and manageable despite the complexity of multiple paths

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent uses the destination server's egress interface congestion status as an intermediary indicator to simplify congestion detection across multiple paths. Instead of requiring the source server to independently monitor congestion on each individual path, the destination server aggregates congestion information from all paths and communicates a unified congestion degree back to the source, significantly simplifying the detection and measurement process while maintaining reliability through diversified path usage

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS12212495B2Reliable fabric control protocol extensions for data center networks with unsolicited packet spraying over multiple alternate data paths

Publication Date: 2025.01.28 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

A fabric control protocol (FCP) is a data transmission protocol that enables spraying of individual packets for a given packet flow across a data center from an ingress interface of the source data processing unit (DPU) across a plurality of parallel data paths of a logical tunnel in the network fabric to the egress interface of the destination DPU. The FCP has congestion control mechanisms used to determine a degree of congestion at the egress interface of the destination DPU and to modify a send window size at the source DPU based on the degree of congestion. Reliable FCP (rFCP) extensions provide reliability enhancements and improved failure resilience within the data center. The rFCP extensions provide an unsolicited mode for low latency operation with enhanced reliability mechanisms. The rFCP extensions provide failure resilience mechanisms to identify and avoid failed paths among multiple parallel data paths within the logical tunnel.