WAN Load balancing

WAN Load balancing

WAN Load balancing
WAN Load balancing
3 weeks ago - last edited 2 weeks ago
Tags: #Load Balancing
Model: OC200  
Hardware Version: V1
Firmware Version:

Hi everyone,


I'm facing a problem with load balancing on a dual-WAN setup using Omada Controller 5.14.26.23 and have a few questions:

 

1. When one WAN connection that was down comes back online, the router seems to disrupt the connection of the other. In other words, I always lose ping when one WAN returns, even if the other is working fine. My router is ER605 V2 and my current configuration is:

  • Load Balancing: 1:1
  • Application Optimized Routing: checked
  • Link Backup: unchecked

 

 

2. I’ve noticed that when a WAN connection drops, the router doesn’t always register it as "Link Down." For example, if I disconnect the PON from the ONU, the router keeps trying to balance traffic over that connection instead of marking it as down. How can I specify an IP address for ping checks, and how many pings need to fail before it flags the link as down?

 

3. My Online Detection Interval is set to a custom value of 1 second. I want quick switching, but I’m wondering if this short interval could cause issues. The default minimum is 1 minute, so I’m concerned if 1-second intervals affect some performance. Both ISPs have low latency, so this shouldn't impact performance, but I wanted to confirm.

 

 

Thanks!

  0      
  0      
#1
Options
1 Accepted Solution
Re:WAN Load balancing-Solution
2 weeks ago - last edited 2 weeks ago

Hi  @Miletos 

 

1. How long will the disruption last? Will it recover to normal quickly? Here are the possible reasons and some suggestions:

The 1:1 load - balancing configuration you have set might be causing the problem. When one WAN connection comes back online, the router may be trying to re - balance the traffic in a way that disrupts the existing connection. In a 1:1 load - balancing scenario, the router distributes traffic evenly between the two WAN links. When a link that down came back, it might trigger a re - evaluation of the traffic distribution algorithm.
Suggested Solution: You could try adjusting the load - balancing ratio. For example, if you have two WAN connections and you want to prioritize one over the other, you can set a different ratio like 7:3 or 8:2, depending on your needs. This way, when the secondary link comes back online, it won't have as much of an impact on the primary, more stable connection.


The "Application Optimized Routing" feature might also play a role. With Application Optimized Routing enabled, the router will consider the source IP address and destination IP address (or destination port) of the packets as a whole and record the WAN port they pass through. Then packets with the same source IP address and destination IP address (or destination port) will be forwarded to the recorded WAN port. This feature ensures that multi-connected applications work properly. However, when a link changes state (from down to up), it could potentially misroute traffic for a short period.
Suggested Solution: Temporarily disable the "Application Optimized Routing" to see if the problem persists. If the disruption stops, it means that this feature is likely causing the issue, and you may need to fine-tune its settings. You could also check if there are any specific application - routing rules that might be interfering and adjust them accordingly.

 

2. You can configure Echo server:

 

3. Setting the online detection interval to 1 second can significantly increase the router's CPU and memory usage. The router has to constantly send out detection packets (such as ping requests) to both WAN connections to check their status. This frequent polling can put a strain on the router's resources, especially if the router's hardware is not powerful enough. For example, if the router's CPU is already handling other tasks like NAT (Network Address Translation), QoS (Quality of Service), and routing for multiple connected devices, the additional load from the frequent detection can lead to slower overall performance.

 

What's more, a very short detection interval can also lead to false positives. For instance, a momentary network fluctuation or a single dropped packet might trigger the router to think that a WAN connection is down, even though it's just a transient issue. This can cause unnecessary switching between WAN connections and introduce network jitter. Network jitter is the variation in latency over time, and it can be detrimental to applications that require a stable connection, such as VoIP (Voice over Internet Protocol) calls or online gaming.

Recommended Solution
  0  
  0  
#2
Options
1 Reply
Re:WAN Load balancing-Solution
2 weeks ago - last edited 2 weeks ago

Hi  @Miletos 

 

1. How long will the disruption last? Will it recover to normal quickly? Here are the possible reasons and some suggestions:

The 1:1 load - balancing configuration you have set might be causing the problem. When one WAN connection comes back online, the router may be trying to re - balance the traffic in a way that disrupts the existing connection. In a 1:1 load - balancing scenario, the router distributes traffic evenly between the two WAN links. When a link that down came back, it might trigger a re - evaluation of the traffic distribution algorithm.
Suggested Solution: You could try adjusting the load - balancing ratio. For example, if you have two WAN connections and you want to prioritize one over the other, you can set a different ratio like 7:3 or 8:2, depending on your needs. This way, when the secondary link comes back online, it won't have as much of an impact on the primary, more stable connection.


The "Application Optimized Routing" feature might also play a role. With Application Optimized Routing enabled, the router will consider the source IP address and destination IP address (or destination port) of the packets as a whole and record the WAN port they pass through. Then packets with the same source IP address and destination IP address (or destination port) will be forwarded to the recorded WAN port. This feature ensures that multi-connected applications work properly. However, when a link changes state (from down to up), it could potentially misroute traffic for a short period.
Suggested Solution: Temporarily disable the "Application Optimized Routing" to see if the problem persists. If the disruption stops, it means that this feature is likely causing the issue, and you may need to fine-tune its settings. You could also check if there are any specific application - routing rules that might be interfering and adjust them accordingly.

 

2. You can configure Echo server:

 

3. Setting the online detection interval to 1 second can significantly increase the router's CPU and memory usage. The router has to constantly send out detection packets (such as ping requests) to both WAN connections to check their status. This frequent polling can put a strain on the router's resources, especially if the router's hardware is not powerful enough. For example, if the router's CPU is already handling other tasks like NAT (Network Address Translation), QoS (Quality of Service), and routing for multiple connected devices, the additional load from the frequent detection can lead to slower overall performance.

 

What's more, a very short detection interval can also lead to false positives. For instance, a momentary network fluctuation or a single dropped packet might trigger the router to think that a WAN connection is down, even though it's just a transient issue. This can cause unnecessary switching between WAN connections and introduce network jitter. Network jitter is the variation in latency over time, and it can be detrimental to applications that require a stable connection, such as VoIP (Voice over Internet Protocol) calls or online gaming.

Recommended Solution
  0  
  0  
#2
Options