Outbound Path Reroute via Forward Progress Hashing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current path failure detection methods, such as BGP, are error-prone due to reliance on timeouts, leading to significant packet loss and network congestion when rerouting packet flows over outbound paths.
Innovation Solution
A reroute process that monitors forward progress on connections, generates a hash value using an identifying value and a failure counter, and selects a new outbound path to quickly reroute packet flows, avoiding failed paths and distributing traffic across multiple routes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If BGP is used to detect path failures, then path failure detection is achieved, but significant packet loss and network congestion occur due to reliance on timeouts
Solution Approach 1:
The patent applies preliminary action by pre-establishing backup outbound paths before failures occur. The system maintains multiple candidate paths in the routing table and pre-computes hash-based path selections, so when a failure is detected at the connection layer, traffic can be immediately switched to a pre-prepared alternative path without waiting for BGP timeouts or performing complex real-time rerouting calculations.
Solution Approach 2:
The patent replaces the mechanical timeout-based detection mechanism of BGP with a connection-layer monitoring mechanism that actively tracks forward progress of data packets. Instead of relying on periodic hello packets and timeout thresholds, the system substitutes a more sensitive detection method that monitors actual data flow acknowledgments, enabling faster and more accurate failure detection.
2Adaptability or versatility
If BGP reroutes packet flows after path failure detection, then traffic is redirected to alternative paths, but significant traffic loss occurs during the rerouting process
Solution Approach 1:
The system performs preliminary routing setup by pre-computing hash-based path selections for multiple failure scenarios and storing these in the routing table. When a failure occurs, the router simply switches to a pre-computed alternative path using a different hash value, eliminating the need for complex real-time path computation and ensuring continuous traffic flow without interruption.
Solution Approach 2:
The patent changes the routing parameter from a single static path to multiple dynamic paths selected via hash functions. By varying the hash input parameter (using different failure counters or connection identifiers), the system can quickly select from multiple pre-prepared alternative paths, enabling rapid adaptation to failures while maintaining continuous traffic flow.
3Reliability
If retransmission attempts are made after path failure, then data reliability is maintained, but network congestion increases due to unnecessary traffic on compromised paths
Solution Approach 1:
The system implements feedback by monitoring connection-layer acknowledgments to detect path failures in real-time. When failure is detected, the system immediately stops sending traffic on the compromised path and switches to an alternative path, preventing unnecessary retransmission attempts that would only increase network congestion. The feedback loop continuously monitors path health and dynamically adjusts routing decisions.
Solution Approach 2:
The patent extracts the retransmission function from the routing layer and handles it at the connection layer instead. By separating these functions, the system can detect path failures at the routing layer and switch paths immediately, while the connection layer handles actual data retransmission only on successfully switched paths, eliminating wasteful retransmissions on failed paths.
4Productivity
If multiple outbound paths are used for load balancing, then traffic distribution is improved, but path failure detection and rerouting complexity increases
Solution Approach 1:
The patent uses parameter changes by varying the hash function input based on the failure counter or connection identifier to select different outbound paths. This simple parameter variation enables automatic load balancing across multiple paths and facilitates rapid failure detection and rerouting without requiring complex routing logic or manual configuration.
Solution Approach 2:
The system applies self-service by using hash-based automatic path selection that requires no manual intervention. The hash function automatically distributes traffic across multiple paths and self-adapts when failures occur by selecting alternative paths based on the same hash mechanism, eliminating the need for complex manual rerouting configurations or centralized control.
Data Source
Figure 1
Figure 2
Figure 3A~3C
AI summary
A server detects a failure of an outbound path based on at least a measure of forward progress made on a connection between the server and an end point. In response to the failure, the server generates a hash value based at least on an identifying value of the connection and a failure counter associated with the measure of forward progress made on the connection. The server then selects a next outbound path for the packet flow based on at least the hash value generated in response to the failure. The server also sends the packet flow over the next outbound path to the end point.