B One SpeedFusion Hot Failover configuration questions (lots and lots of questions...)

Hi all,

I just got a Peplink B One (to replace a TP-Link ER605 router)

I am using both Starlink and Verizon Home Internet (unlimited data) as WAN1 and WAN2 respectively.

I was able to do the basic/default configuration, which was actually fairly easy, and it seems that the B One sends connections to either WAN based on their relative speed as defined in the upload and download bandwidth in WAN Connection Settings

but then I found out that for the default outbound policy rule

When Auto is chosen, lowest latency algorithm will be applied.

yet the description for upload WAN speed says:

This value is referenced when default weight is chosen for outbound traffic and traffic prioritization. A correct value can result in effective traffic prioritization and efficient use of upstream bandwidth.

if indeed “lowest latency” is used for the default, then what are the upload and download WAN speeds used for?

I assume that when/if either WAN fails, the default policy will automagically route via the remaining active WAN, and if so a new session will have to be established with the remaining active WAN?

To save on SpeedFusion data transfer $$$, I would like only my wife’s laptop and her VOIP phone to use SpeedFusion, these are the “critical devices” and need to to be up and connected at all times without interruptions or dropping sessions (the main reason I purchased the B One)

though it would be highly desired to exclude streaming videos (from any source, youtube, x.com, amazon videos, etc, etc, etc) from SpeedFusion for the critical laptop, again to save on SpeedFusion data $$$

all other non-critical devices should continue to use the current basic configuration i.e. routing through either WAN1 or WAN2 (and when either one fails, though the remaining WAN)

but I had no idea how to setup SpeedFusion for Hot Failover…

the latest manual I could find was for the 8.4.0 firmware, and I have 8.5.0 and some sections/functionality/terminology in the manual does not match my router’s interface…

the phrase “Hot Failover” is not found at all in the 8.4.0 version of the manual???

After spending several hours, I believe that I figured it out, but still have questions, lots and lots of questions.

maybe once they are all answered, we can make a configuration guide for new B One users, esp since the B One will probably be purchased by home users such as myself who have little to no experience setting up a complex dual wan router.

in the B One interface the only place where I could see a “Hot Failover” logo is in the “Advanced > SpeedFusion VPN” screen, so I initially started there. but clicking on “New Profile”, the “Speed Fusion VPN Profile” configuration setup form does not have “Hot Failover” selection/option and requires a “Remote ID” and “Pre-shared Key” which I have no idea how to fill. If I have to guess, this is for connecting to another corporate main office Peplink device (which is why the “Remote ID”, not for connecting from my home to the internet. is that correct?

the other option I could see was the “SF Connect” menu, which initially was very puzzling. but I started from the top down.

in “Client Mode” it was set up for “automatic” which I understand chooses the best/nearest SFC location.

clicking on the “SFC” pops up a configuration form

but I could also not find any configuration selection/option for “Hot Failover” selection anywhere…

and I have no idea how to set the rest of the settings on this form :frowning:

so I started making educated guesses and assumptions:

I turned on “WAN Smoothing” to “Normal” as it said that it will use twice the bandwidth, and I assume that this is how it would do the “hot failover” by sending double the packets one through each of the two active WAN connection. is that correct? or will it consume my SpeedFusion data and $$$ twice as fast?

but if I do not have “Smoothing” turned on, will I have the uninterruptible connection protection, i.e. the illusive “Hot Failover” I want to have???

or will disabling “Smoothing” result in “plain old bonding” of the two WAN connection which will only give me more throughput but not “hot failover”???

I then set Forward Error Correction to “Adaptive”, as it seems like a good thing to have, is it? or is not not worth the extra bandwidth it consumes?

I left Traffic Distribution Policy at “Dynamic Weighted Bonding” but set Packet Jitter Buffer to 25ms (since I wanted to keep latency low, but not disable the buffer), was this a reasonable thing to do?

and finally under WAN Connection Priority, I set both “Starlink” and “Verizon” to priority 1 and turned off all the rest (as none are active/configured)

question: what the additional tunnel (clicking on the “+”) would be used for? clicking on the “+” duplicates the configuration settings of the original/default. yet I can make changes to the config options. what might be a use case for this additional tunnel? i.e. why/when/how would I use another tunnel with a different configuration?

I then clicked on “route by LAN Client” and added the two Ethernet MAC addresses for my wife’s laptop and her VOIP phone. I assume that these two MAC addresses will be routed via SFC

I created a new SSID called “mySSID4SFC”

I then clicked on “Route by Wi-Fi SSID” and selected “mySSID4SCF” and overwrote the default SSID for SFC with “mySSIDSpeedFusion”. I assume that any client connected to the “mySSIDSpeedFusion” will be routed via SFC, but what about clients connected to “mySSID4SFC” which is the “reference” SSID? default routing?

But wait, there’s more…

I then found out about Outbound Policy, which seems to be able duplicate the configuration above. i.e. I set up one outbound policy per each MAC address and an outbound rule for “Client’s Associated SSID” to which I assigned “mySSID4SFC” (since I could not assign “mySSIDSpeedFusion”. for all these new rules I chose “Any” Destination and “Any” Protocol. I selected the “Priority” Algorithm and moved SFC to highest priority followed by Starlink and Verizon.

so now I believe that clients connected to either mySSID4SFC or mySSIDSpeedFusion will connect via SFC. is that correct?

but what about the two sets of MAC address rules, the two MAC addresses I entered in “route by LAN Client” and the two Outbound Policy Rules for the same MAC address? will they conflict? which one would take priority? which one is better to have?

my gut feeling says that it might be best to delete the “route by LAN Client” and “route by Wi-Fi SSID” from the SF Connect configuration and just use the Outbound Policy rules, as it would avoid the “duplicated” i.e. the extra mySSIDSpeedFusion and just leave mySSID4SFC as the SSID which is used to route via Speed Fusion. it also seems that the Outbound Policy Rules are more configurable (i.e. more configuration options), for example, I can set Priority for when SFC might run out of data and fall through to the next rule, etc.

and final question: how would I exclude streaming videos (youtube, twitter/x, prime video, etc) from SpeedFusion? as I rather not pay for SpeedFusion data for video streaming to my wife’s laptop. and I also read that most might be blocked since they are considered “VPN”?

The least used Algorithm.

Yes if a WAN fails sessions on that WAN will terminate and need to be rebuilt, new sessions get sent over the other active WANs.

This is used to connect to your own remote appliances. So between your device and another physical device at another location, or between your physical appliance and a virtual FusionHub appliance you can install in the cloud.
I have an old FusionHub setup video here that you might find enlightening https://www.youtube.com/watch?v=ZxZBWloVizM

Hotfailover is the consequence of having more than one active WAN link in priority 1 on the dashboard, AND then having those healthy WANs used at the same time inside of a Speedfusion VPN tunnel (or SFC tunnel), AND then sending your traffic over that VPN link.

If speedfusion has more than one active WAN in simultaneous use, traffic flowing over it can hot failover between those WANs.

Have more than one WAN, enable SFC and you’ll have hot failover.

WAN smoothing duplicates traffic, it is not hot failover. You consume double the bandwidth to ensure a sent packet gets to the other side. Used when WAN links have lots of packetg loss.

If a WAN fails it will be seamless, but when WANs are healthy you are burning data on both WANs and so reducing your total available bandwidth for user traffic by half. You don’t want this. Although you might enable it for VoIP traffic alone since that is very low bandwidth and it ensures the traffic gets through.

Adaptive FEC is the best thing in the whole wide world. It helps with Starlink + Cellular connections. Turn it on.

I always leave it at 150ms

Excellent. That enabled hot failover between the two P1 WANs.

You might have a primary tunnel with Dynamic Weighted bonding and adaptive FEC, then a subtunnel just for VoIP with WAN smoothing enabled. On Outbound policy the subtunnel is there as a WAN option so you can send VoIP traffic via it.

Or you might have a subtunnel that just uses Starlink which you send you smart TVs over so that they don’t consumed cellular bandwidth.

Yes. This is an automated way to set up an outbound policy for those MAC addresses.

Yes default routing. Or whatever you have in our outbound policy list.

Outbound policy gives you full granular policy based routing control. What you did in SFC is like a wizard to achieve that for SFC tunnels and LAN clients that use it.

That is what I do. I like the full control I have in outbound polices.

Take a look at SaaS steering overview videos for guidance. https://www.youtube.com/watch?v=rcqh84KP8lk

1 Like