Cellular Bonding Performance Problems


We have an HD4 MBX and are experimenting with bonding the 4 cellular modems. I’m using iPerf with 50 TCP/IP connections to our server.

I’m comparing Direct to Cell 1, Weighted Balance Across 4 Cells, SFC to London, Belgium, Miami and Speed Fusion Hub Solo running on Google Cloud Services in various locations as well.

I’ve tried Bonding and Dynamic Weghted Bonding.

No matter what I do though I don’t see any any benefit in any of the bonding solutions. I don’t expect to get 4x the speed of a single SIM. I understand there are VPN overhead etc. And the benefits of bonding for reliability. But usually the bonding actually gives me less speed than the Direct to Cell 1 connection.

I was hoping for a doubling or tripling of speed across 4 modems maybe. What am I missing?


Just following this topic as I’m working on this again now too.

Question: if you connect only one cell wan, how does your speedfusion throughput on the one cell compare to speed over the that connection directly (non speedfusion)? For some reason on one cellular I have I’m seeing poor performance this year of vpn traffic while a direct speedtest.net is performing great.

Second, are you seeing the poor bonding performance both up and down? For some reason, I’ve been able to bond upload bandwidth, but not download bandwidth over cellular this year. (bonding worked well for me with dsl connections in contrast.)

You might post the speedfusion graph while dong a test so others can see what is happening (packets out of order, packets lost, latency differences, how each cell wan is being utilized, etc.).

What’s the difference in throughput when you are using weighted local balaning across 4 cellular connections compared to speedfusion bonding?

1 Like

I too have not seen the ability to truly bond multiple cell phone modem… If there’s somebody out there that is getting true aggregated connectivity between multiple WAN’s please share the secret.