I have a DSL WAN which is usually fastest, and then a starlink WAN as backup.
For 3 hours per night, the DSL connection is slowing down to a fraction of its rated speed to around 2 mbps. The rest of the time, it’s perfect.
Speedfusion is set with
DSL WAN as Priority 1
Starlink WAN as Priority 2
so it’s going over DSL for lowest latency and consistency and only failing over to stalrink if dsl is down, which is what I want 21 hours of the day.
What is the best way to have speedfusion switch to the starlink wan as primary during these “slow” 3 hours from 8:30-11:30 pm and then switch back to the dsl wan?
Ping on the dsl looks like this during the congested 3 hours each day:
Reply from 8.8.8.8: bytes=32 time=27ms TTL=60
Reply from 8.8.8.8: bytes=32 time=26ms TTL=60
Reply from 8.8.8.8: bytes=32 time=21ms TTL=60
Reply from 8.8.8.8: bytes=32 time=21ms TTL=60
Request timed out.
Request timed out.
Reply from 8.8.8.8: bytes=32 time=20ms TTL=60
Reply from 8.8.8.8: bytes=32 time=20ms TTL=60
Request timed out.
Reply from 8.8.8.8: bytes=32 time=23ms TTL=60
Reply from 8.8.8.8: bytes=32 time=33ms TTL=60
Reply from 8.8.8.8: bytes=32 time=21ms TTL=60
Reply from 8.8.8.8: bytes=32 time=20ms TTL=60
Reply from 8.8.8.8: bytes=32 time=20ms TTL=60
Reply from 8.8.8.8: bytes=32 time=22ms TTL=60
Reply from 8.8.8.8: bytes=32 time=20ms TTL=60
Reply from 8.8.8.8: bytes=32 time=21ms TTL=60
Reply from 8.8.8.8: bytes=32 time=23ms TTL=60
Request timed out.
Reply from 8.8.8.8: bytes=32 time=20ms TTL=60
Reply from 8.8.8.8: bytes=32 time=21ms TTL=60
Reply from 8.8.8.8: bytes=32 time=20ms TTL=60
Reply from 8.8.8.8: bytes=32 time=20ms TTL=60
Reply from 8.8.8.8: bytes=32 time=20ms TTL=60
You could create a custom schedule to disable the DSL WAN during the “bad hours”.
This is a bit of a blunt tool but would get you the end result you are looking for, unfortunately there aren’t more granular health checks that consider things like latency / loss for disabling WANs or lowering their priority.
On the Peplink navigate to System > Schedule (you can also manage the schedules from Ic2).
Make a new schedule that would disable during the “bad” time window:
Thanks very much. I didn’t realize there was a scheduler and this looks very interesting.
I first thought that I could use the latency cutoff on speedfusion… oddly when the dsl wan has a lost packet, speedfusion has the next hop ~160-170 ms. I’m not sure if I can use that though to set a cutoff latency or if I would introduce problems in trying…
The scheduler looks much simpler since the dsl issue occurs every night during the same hours. The only thing I would lose would be the redundancy, since I only have the dsl wan plus the starlnk wan: I take it with this if I set the dsl to a schedule, then if the starlink had an outage I’d be down? It would be even neater if I could schedule the connection priority on a schedule so it would then become a backup wan 4 hours a day, and primary 20 hours a day.
One other question: when the DSL wan ping looks like this:
Why does the ping through the speedfusion tunnel look like this
What exactly causes the ~170ms ping every time following the lost packet over the speedfusion tunnel? (FEC is off rightnow, I have DSL in priroirty 1 and starlink in priority 2 right now on the speedfusion tunnel).
Not sure and honestly I wouldn’t get too hung up on it.
I’d be more inclined to try and understand why your line predictably / regularly goes bad for those hours instead - is it capacity in the ISPs local network, or something else?
DSL lines - assuming that is in the traditional sense delivered over copper to your property - could be something environmental, what happens every day at 8:30pm that might explain this?
For example - I used to have an ADSL line to my house, in the winter months the performance would often degrade and then at some point it just outright broke, that was due to water ingress in a cable join that was worse in the winter due to heavier rainfall.
Unfortunately it wasn’t until it outright broke that I could convince the ISP to take the complaint seriously but they did end up replacing a segment of cable to my house along with repairing a joint box and the problems went away entirely.
8:30pm though I guess is probably prime time and people hammering away at streaming services and the like, so it could also just simply be congestion, either way I’d be calling the ISP and asking them what might be going on.
It’s backhaul congestion – there was a government grant to install a new fiber backhaul last year, but they have a 5-year window to complete, so I will likely have to work around this condition for a couple years until it’s complete.
Most of the day right now, the dsl works flawlessly and noticeably faster (latency and consistency) than starlink, except for the hours when the backhaul is congested when it now loses 98% of the download speed and gets 12% packet loss during those hours. Then when people stop streaming, it returns to perfect.
So I had a little thought about a better way to automate this where you are adjusting the priority of the WAN vs outright disabling it on the schedule, but it will require a bit of work outside of what Peplink provides in their GUI.
According to the API docs the WAN config is possible, and you can set the priority of a WAN using the API.
The TST Duo Pro can also run Docker containers, so I wonder if it might be possible to bring up a small linux environment that can talk to the routers API.
A little script (Python probably would be my choice here) could then be run on a timer or cron job to raise and lower the WAN priorities.
