VXLAN Link Detection via Controller-Managed Path Traversal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies fail to traverse all paths in a VXLAN network, leading to incomplete detection of link failures, as they rely heavily on hash path selection algorithms and cannot ensure that all paths are traversed, especially in cascading equal-cost path scenarios.
Innovation Solution
A method and apparatus that utilize a controller to send detection messages with unique identifiers across multiple equal-cost paths, allowing network devices to replicate and forward these messages through all paths, ensuring comprehensive link status information is collected and recorded.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If detection messages are sent sequentially with increasing source ports to traverse all paths, then path coverage is improved, but in cascading equal-cost path scenarios, some paths are still missed due to hash path selection algorithms
Solution Approach 1:
The detection message is segmented into multiple copies, each sent through a different equal-cost path. The controller divides the detection task by creating replicated messages that traverse different network paths simultaneously, ensuring complete path coverage without relying on sequential source port changes that may be filtered by hash algorithms.
Solution Approach 2:
The controller acts as an intermediary that manages the detection process by sending detection messages to the starting VTEP and collecting responses from all network devices. This centralized coordination ensures that all paths are traversed and monitored, overcoming the limitations of distributed hash-based path selection.
2Reliability
If multiple detection messages are sent to traverse all paths, then path detection capability is improved, but message complexity and processing overhead increase
Solution Approach 1:
A single detection message format is designed to serve multiple functions: it can be replicated by the starting VTEP for path traversal, forwarded by intermediate network devices for path monitoring, and responded to by the destination VTEP for link status confirmation. This multi-functional message design reduces overall system complexity while maintaining comprehensive path detection.
Solution Approach 2:
The detection mechanism incorporates feedback loops where the destination VTEP sends detection responses back through the same path to the controller. This feedback mechanism allows the controller to verify path status and link health without requiring separate detection messages for each direction, reducing message complexity.
3Reliability
If detection messages are replicated at the starting VTEP for all equal-cost paths, then path traversal completeness is improved, but network traffic overhead increases
Solution Approach 1:
Detection messages are sent periodically rather than continuously, with the controller initiating detection at intervals. This periodic approach maintains accurate path detection capability while significantly reducing the overall volume of detection traffic in the network compared to continuous or event-driven detection schemes.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Link detection method, apparatus, network device and controller are provided in the present application. The method includes sequentially sending a detection message from a starting network device to each network device until an ending network device on a detected link, the detection message including a first source IP address and a first destination IP address on the detected link, and a first identifier of the detection message; receiving link status information reported by a network device on the detected link in response to the detection message, the link status information including the first source IP address, the first destination IP address, a device IP address of the network device, and the first identifier; and performing a matching between the detection message and the detected link based on the first source IP address, the first destination IP address and the first identifier included in the link status information, and locating a network position on the detected link in which the network device is located based on the device IP address. The technical solutions of the present application can enable a controller to know link status information of all links on a detected link for identifying and locating a failed link in a quick manner.