How to enable mid-session failback to higher priority WANs without WAN Smoothing?

Product Discussion
,
1

Hi everyone,

Setup: FusionHub, Max BR2 Pro, 3 WAN Emulators on NetEm basis.

We are running SpeedFusion sub-tunnels ([FusionHub] - Latency & Capacity) with a strict 3-tier WAN hierarchy:

  • Tier 1: 5G

  • Tier 2: OneWeb (OW)

  • Tier 3: VSAT

  • Outbound Policy has 5100 traffic routed via the Latency tunnel with above hierachy.

Problem:
During failover testing with a continuous UDP stream (iperf3 -u on port 5100), failover works seamlessly (5G → OW → VSAT). However, when OW and 5G recover, the active UDP stream stays stuck on VSAT. It refuses to preempt back to Tier 1/2, even when the PepVPN status for those WANs returns to “Established”.

Constraints:

  1. No Session Restarts: Active streams cannot be restarted as the test protocol dictates so.
  2. NO WAN Smoothing / FEC: WAN Smoothing is intentionally DISABLED and NOT an option due to severe bandwidth overhead costs on satellite links.

Questions:

  1. Is there a way to force active UDP flows (UDP flow pinning) to dynamically fail back to higher-priority links upon recovery without WAN Smoothing?
  2. Which PepVPN, Reordering Buffer, or UDP Timeout settings control this behavior during link restoration?

Appreciate any insights on how to achieve hitless UDP failback under these bandwidth-constrained conditions!

Thanks and Best Regards,

Lars

1 Like