Network Resiliency System for Dynamic Routing Updates

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Massively parallel processing systems face challenges in dynamically updating routing information without causing deadlocks and ensuring packet ordering when link errors occur, as existing methods often require reinitializing the entire network, which is inefficient and disruptive.

Innovation Solution

A network resiliency system that detects link errors and generates new routing information to bypass failed links, entering a quiescent state to prevent packet mixing and then loads new routing information into routers, allowing for dynamic updates without reinitializing the network, ensuring deadlock avoidance and packet ordering.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If routing information is updated dynamically after a link error, then network resiliency and adaptability are improved, but deadlocks may occur and packet ordering may be violated during the transition

Engineering Contradiction:

Improvenetwork resiliencyVSAvoidpacket ordering

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The system performs preliminary actions by suppressing new request packets before updating routing information and by validating the new routing information to ensure it maintains deadlock freedom and packet ordering properties. This prevents mixing of packets using old and new routing information, thereby avoiding deadlocks and ordering violations during the transition.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system introduces an intermediary validation mechanism that checks the new routing information against deadlock freedom criteria and packet ordering requirements before applying it. This intermediary layer ensures that only safe routing updates are applied, maintaining reliability while enabling adaptability.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If the entire network is reinitialized after a link error, then routing consistency is restored, but network downtime and loss of time increase significantly

Engineering Contradiction:

Improverouting consistencyVSAvoidnetwork downtime

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system segments the routing update process into localized changes rather than requiring full network reinitialization. Only the affected routing tables and components are updated, allowing the rest of the network to continue operating normally. This maintains routing consistency in affected areas while minimizing overall network downtime.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Instead of performing a complete network reinitialization (excessive action), the system applies partial updates only to the necessary routing components affected by the link error. This achieves sufficient routing consistency restoration with minimal disruption and time loss.

Inventive Principle:

Principle #16Partial or excessive action

3Productivity

If new routing information is loaded while packets are in transit, then network continuity is maintained, but packet mixing occurs causing deadlocks and ordering violations

Engineering Contradiction:

Improvenetwork continuityVSAvoiddeadlocks

Core Design Contradiction:

ProductivityVSObject-generated harmful factors

Solution Approach 1:

The system suppresses new request packets as a preliminary action before loading new routing information. This prevents new packets from being injected into the network with outdated routing information, eliminating the source of packet mixing that causes deadlocks and ordering violations.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system allows in-transit packets to complete their journey using the old routing information without interruption, then quickly transitions to the new routing information once those packets are cleared. This rushing through the transition period minimizes the window for packet mixing while maintaining network continuity.

Inventive Principle:

Principle #21Skipping (Rushing through)

Data Source

PatentUS9154400B2Dynamically updating routing information while avoiding deadlocks and preserving packet order after a configuration change

Publication Date: 2015.10.06 CRAY INC

AI summary

A system for allowing dynamic changing of routing information of a network interconnect while avoiding deadlocks and preserving packet ordering. A network resiliency system detects when an error in the network interconnect occurs and dynamically generates new routing information for the routers that factors in the detected error. The network resiliency system then generates new routing information that factors in the failure. The network resiliency system then directs the network interconnect to enter a quiescent state in which no packets are transiting through the network interconnect. After the network interconnect enters the quiescent state, the network resiliency system directs the loading of the new routing information into the routing tables of the network interconnect and then directs the network interconnect to start injecting request packets into the network interconnect.