4 Cellular Connections with Two SpeedFusion Engines for High Reliability

Hi, I recently purchased a SpeedFusion Engine. I am wondering if it is possible to combine two SpeedFusion Engines on the same LAN (or VLAN) and link them both to FusionHub via a PepVPN connection to effectively have a bonded connection consisting of 4 cellular WANs. For example, in some daisy-chaining or master/slave type arrangement.

Is this possible? If so how would this be configured on the client (Speedfusion Engine) side?

Thanks!

There is only a single ethernet port on the SFE. So the only way you could do this is set the first SFE’s ethernet port to be a WAN and plug the 2nd SFE into it. Your computer would have to connect over USB to the first SFE.

That configuration would work ok for load balancing, but bonding will be messy. If you set up a bonding profile on the first SFE it would use its two embedded cellular conections and then only one of the cellular connections of the 2nd SFE on the WAN.

Do you really need to use a SFE? Why do you need to bond 4 cellular connections? How much bandwidth / what number of mobile networks do you need to support?

2 Likes

Thanks, @MartinLangmaid!

I guess it is a better idea to state my problem rather than my naive solution :smiley:

Actually “bonding 4 cellular connections” is misleading, I’ve updated the title.

My objective is in-fact to have as reliable a mobile connection as possible for video streaming in an urban environment. I’m trying to create a mobile video streaming setup with as small a form factor as possible. Thus SpeedFusion Engine was chosen due to its small size.

So, bandwidth is really not a huge limiting factor here but instead reliability of the connection is (the video data is traveling over a WebRTC channel).

I’m planning to connect to FusionHub and enable Forward Error Correction and Latency cut-off (at around 150ms) but likely not WAN smoothing to save on data costs.

4 Cellular connections would be preferred to 2 so as to have the maximum possible level of reliability (in the deployment area I should be able to find 4 carriers operating different network APs).

Given the above requirements, would do you have any further recommendations w.r.t. the Hardware to use on the client side of FusionHub setup?

Thanks so much for any help!

I see. by mobile do you mean moving whilst in use, or do you mean portable to where it needs to be used?

Speedfusion engine is a fantastic bit of kit but its size dictates its limited connection options. I find video streaming in the field to be most successful using the MAX transit. It is bigger, but its a more useful product for in production use. We build these into portable enclosures with batteries which works very well.

If I need to use this rig on the move (eg in the back of a car), I’ll add a BR1 or another transit to its WAN to give me more of a chance of maintaining connectivity.

If it is size you’re after you can’t beat the SFE, but its not something I’d likely end up using for this kind of job to be honest.

3 Likes

Hello @AJCarter,
Can you use the SpeedFusion Engine Cam?

We have been doing some work with this unit in Australia, and as it has a wired WAN connection & both 2.4Ghz + 5Ghz Wi-Fi, this can be used to offer additional connectivity redundancy and bandwidth.

In our typical deployment, the two internal modems get configured in the following way:

  • Modem #1, SIM A with Telstra, SIM B with Optus
  • Modem #2, SIM A with Optus, SIM B with Telstra

We use SpeedFusion to combine the two modems offering primarily redundancy.

The SpeedFusion Engine Cam also offers Wi-Fi WAN, so you can connect wirelessly to additional systems to give you that additional redundancy & bandwidth.

We have used the SpeedFusion Engine Cam with professional JVC & Panasonic cameras in Sydney.
Happy to Help,
Marcus :slight_smile:

2 Likes

Good idea Marcus. I forgot the SFE CAM has a WAN port.

1 Like

Thanks for the feedback and support @MartinLangmaid, @mldowling!

To clarify, I meant mobile as in moving while in use.

I’ll consider some other units. I think the MAX HD4 will be too big given my constraints, but a MAX Transit Duo or Max HD2 MINI + a MAX BR1 MINI sounds promising.

@MartinLangmaid, @mldowling

Apologies for raising this question from the dead but I have some new information and questions pertinent to this proposed setup.

Firstly, I want to clarify that in this deployment the concern is around connection reliability/ packet-loss compensation rather than bandwidth bonding, so the plan is to use PepVPN tunnel(s) with FEC or WAN Smoothing enabled.

My follow up question is whether or not it would be possible to continue to use Ethernet to supply internet access to the compute node (Linux) if a Network Switch was introduced to connect the two SpeedFusion Engines on the same LAN. I’m envisioning one of the SpeedFusion Engines (A) configured as a WAN node and the other SpeedFusion Engine (B) making use of A in its configuration to use it as another link for the purposes of redundancy in the PepVPN and/or Outbound Policy configuration. The USB data connection may be an option, but the issue is there are potentially other devices on the LAN that need to access the internet as well, other than the main compute node.

Unfortunately, I am stuck with two SpeedFusion Engines for the foreseeable future so I’m trying to work out the best way to utilize both to achieve the most reliable connection.

Thanks as always!