• MESHdesk and APdesk now offer multi-WAN support.
• This allows you to ensure that your customers are always connected.
• The multi-WAN feature includes the following options
  • Support for a number of Ethernet ports that can be used as a WAN connection.
  • Support for up to three radios that can be used for WAN over WiFi.
  • Support for up to six LTE modems.
• You can use any combination of these connection types and customise them to your needs.
• We offer two modes:
  • Load balancing
  • Failover
• In load balancing mode, the active and standby connections can in turn be grouped and weighted (to distribute traffic).
• Next, we will look at the Multi-WAN Profiles applet and finally see how easy it is to apply these profiles to devices.

• The Multi-WAN Profiles applet is located under the Other selection.

• A functional multi-WAN profile consists of the following:
  • A name, which can be thought of as a find of folder, under which the interfaces are grouped. There is also a site-wide option that makes the multi-WAN profile available for all clouds in RADIUSdesk.
  • One or more interfaces that are used for WAN access.

• There are two gentle reminders with instructions on how to make the multi-WAN profile usable.
• The first reminds you that you need to add at least one interface.

• The second reminds you to apply a policy after you have added an interface to the Multi-WAN profile.

• Each Multi-WAN profile has an associated policy.
• Here you define the mode of the multi-WAN profile and the weighting of the individual interfaces as well as the role in the event of failover mode.

• Once you have applied the policy, the multi-WAN profile is ready for use.
• Next, we will look at the settings that are triggered when an interface is marked as down or up.

• For each interface you define, there is a separate area where you can specify the conditions under which the respective interface is considered unavailable / offline.
• Here are some things to keep in mind when customising some values.
  • Active monitoring, where the hosts are pinged, is not mandatory
  • If it is not enabled, the multi-WAN configuration will still monitor whether the interface itself is available or unavailable and act accordingly.
  • Some of the more popular hosts used for ping testing may limit the number of responses or not respond at all.
  • This can lead to a false trigger and also to the device being unreachable on the specific interface (e.g. for accessing the device via ssh or Luci).

• When you attach a device to a mesh or access point profile there is the option to select the type of Internet connection.
• If you select the Multi-WAN option, another select control will show where you can select which multi-wan profile to apply.
• Remember that the multi-wan profile and hardware has to be compatible. For instance you cant use a multi-WAN profile that has a LTE interface defined on hardware that does not have LTE capability.

• Multi-WAN Internet connections are visually distinctive on the overviews of mesh nodes and APs.

• In the following screenshot we have an Access Point with a multi-WAN profile in failover mode.
  • The WAN interface is active.
  • Traffic allocation is 100%
  • Two LTE interfaces are in the standby role.
  • They are actively monitored and thus marked in blue as on standby.
  • The WiFi interface are not actively monitored but the interface itself is up.

• Next we introduce a scenario which triggers a failover.

#We block ping packets on the host to which the access point is configured to send.
#This will trigger  a failover 
sudo iptables -A INPUT -s 197.64.14.100 -p icmp --icmp-type echo-request -j DROP
#We remove the block 
#This result in a fall back.
sudo iptables -D INPUT -s 197.64.14.100 -p icmp --icmp-type echo-request -j DROP

• Failover triggered

• Ping block removed and active role link is restored.