Network Port Failover via Routing Protocol and Loopback IP
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computing environments like VMware ESX or ESXi hosts struggle to detect and manage failover between network ports due to limitations in detecting forwarding failures, leading to IP packet loss and service faults, especially in environments with a single network stack and limited physical ports.
Innovation Solution
Implementing an IP routing protocol such as BGP, combined with loose reverse path forwarding and assigning an externally visible IP address to a loopback interface, allows for the detection of forwarding failures and seamless failover between physical network ports, even with only two available ports, while enabling full routing between switches and enhancing security and performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If beacon probing is used to detect forwarding failures, then forwarding failure detection capability is improved, but the requirement for at least three physical ports connected to different switches increases device complexity and limits applicability
Solution Approach 1:
The patent introduces an intermediary IP address (not directly assigned to any physical port) that acts as a mediator for failure detection. Routing protocol messages are sent through this intermediary address via different physical ports, allowing the system to detect forwarding failures without requiring beacon probing's three-port configuration. The intermediary IP address decouples the detection mechanism from direct port-to-port requirements.
Solution Approach 2:
The patent changes the detection parameter from physical port connectivity (layer 2) to IP routing reachability (layer 3). By using routing protocol messages directed to an intermediary IP address, the system transitions from detecting physical link status to detecting logical forwarding path status, thereby reducing the number of physical ports needed while maintaining reliable failure detection.
2Device complexity
If a single network stack with one default route is used, then device complexity is reduced, but the ability to detect and manage failover between network ports deteriorates
Solution Approach 1:
The intermediary IP address serves as a mediator that enables failover detection within a single network stack. By directing routing protocol messages to this intermediary address through different physical ports, the system can monitor the health of each port independently and trigger failover when a port becomes unreachable, all while maintaining a single default route configuration.
Solution Approach 2:
The single network stack with one default route is made multi-functional by using the intermediary IP address to serve multiple detection purposes. The same default route configuration is used for both normal data traffic and failure detection traffic, allowing the simplified network stack to simultaneously maintain low complexity and provide robust failover management through the intermediary address mechanism.
3Reliability
If loose reverse path forwarding is enabled, then failover between physical ports is improved, but the network stack configuration complexity increases
Solution Approach 1:
The intermediary IP address acts as a mediator that simplifies the configuration of loose reverse path forwarding. By using this intermediary address in routing protocol announcements, the system can enable loose reverse path forwarding with a single configuration statement rather than complex per-port settings, as the intermediary address serves as a universal target for all failover detection messages.
Data Source
AI summary
Using a routing protocol for network port failover. A network stack, which does not perform reverse path forwarding, includes a first interface as a loopback interface that is assigned a first internet protocol (IP) address, a second interface to a first network port, and a third interface to a second network port. The routing protocol is configured to communicate over the second interface with a first port at remote network device(s), and to communicate over the third interface with a second port at the remote network device(s). A route comprising the first IP address is announced to the remote network device(s), and default route of the network stack is configured to egress IP packets via the second interface. Later, the routing protocol indicates that there is a forwarding failure associated with the second interface, and the default route is configured to egress IP packets via the third interface.