Probing Behaviour - WAN Monitoring
Overview
WAN Monitoring is a method used in routers to check whether each internet connection (WAN interface) is working properly. It helps the router decide when to switch between internet links during failover or load balancing.
This document explains how WAN Monitoring technique checks the connection status and how it updates the state of each WAN interface.
What WAN Monitoring Does
WAN Monitoring tools continuously watch each internet link connected to the router. Its main job is to:
Detect when an internet link is working (UP)
Detect when an internet link stops working (DOWN)
Switch to another link if needed
Avoid wrong decisions caused by temporary network issues
Checking an Internet Link
Every 2 seconds, WAN Monitoring tools perform a ping test on each WAN interface.
Ping Test
Sends a ping to known internet servers.
Default servers (Probe IPs):
13.92.16.93
40.118.248.237
These servers are Kognitive hosted cloud services with full reliability.
Ping Logic
First pings the first IP.
If it gets a reply within 2 seconds, the test is marked as successful.
If not, it tries the second IP.
If neither responds, the test is marked as failed.
This ensures accurate checking even if one server doesn’t respond.
How to Decide the Link Status
The status is not changed after just a single test. It waits for multiple results to avoid false alarms, especially when the ping fails.
Marking a Link as UP
If the system gets 5 successful test results in a row, it marks the internet link as UP (working).
Marking a Link as DOWN
If the system gets 5 failed test results in a row from both servers, it marks the internet link as DOWN (not working).
These are called:
Link Up Counter: 5
Link Down Counter: 5
This method ensures stable and accurate detection.
Updating the Status: Time Taken
Concurrent Testing
WAN Monitoring tools check all WAN interfaces at the same time (concurrently). It does not wait for one interface’s test to finish before starting another.
Why is it important:
It avoids delays when multiple internet links are connected.
Each interface gets tested independently.
Status updates happen faster and more accurately.
When the Internet Goes DOWN
The overall delay depends on how often tests run and how many results the WAN Monitoring tool waits for.
Every 2 seconds, the tool tests the connection by sending a small ping.
It waits for 5 failed tests in a row before declaring the link as DOWN.
Timing
2 seconds × 5 failures = about 10 seconds
Additional internal processing and scheduling overhead increases this to around 30 seconds + concurrency delay
In short:
It takes approximately 30 seconds for the tool to confidently decide that the internet on that WAN link is down.
When the Internet Comes UP
When the internet starts working again, the tool needs 5 successful tests in a row.
Timing
2 seconds × 5 successes = about 10 seconds
Actual time including processing = around 20 seconds + concurrency delay
In short:
It takes approximately 20 seconds for the tool to declare the link as working again.
When the Cable Is Unplugged
If the physical Ethernet cable is removed:
The router detects the link break immediately.
The WAN Monitoring tool instantly marks the interface as DOWN.
No waiting time or repeated tests are required.
In short:
Physical disconnection is detected instantly, while internet-related failures take longer because confirmation is needed.
Parameters / Attributes
The parameters / attributes in the configuration file are listed below:
Probe IPs
Probe method (ICMP)
Probe timeout
Probe frequency
Link up/down counters
Configurability of these parameters from the UI will be supported in subsequent releases.
Summary
This method provides a smart and reliable way for the router to:
Continuously check each internet link
Avoid mistakes caused by network fluctuations
Quickly switch between multiple internet connections
Keep your network stable and connected
By regularly testing and confirming results before changing the status, the system ensures smooth and uninterrupted internet connectivity.