Hello!
We are having an issue with a client and are relatively new to Peplink. The client’s primary internet has intermittent packet loss, but the WAN Smoothing option only seems to protect SIP traffic from one direction…
Please forgive the ugly diagram below but hopefully it helps.
Hello @seb82,
Have you enabled WAN smoothing on both ends, as this only applies to traffic being sent from the device, across the WAN connections. Do you know what the latency levels are like on each of the connections?
Which devices are being used at Site 1 and Site 2 and are they using Firmware 7.1.0?
Thanks,
Steve
Hi Steve, thanks for replying so quickly.
We are using Max BR1 Ent devices with fw 7.1.0 at both locations.
Both locations have WAN Smoothing enabled on their PepVPNs.
We also tried turning it on WAN Smoothing on the Fusion Hub, but since the Fusion Hub only has one NIC, does it make a difference?
Our current fix for now is that we are only using LTE for the VPN to the Fusion Hub from Site 1 which causes things to run properly, but we are going through LTE data very quickly.
The latency we see on LTE is about 50ms and on Comcast is about 25ms, but Comcast is dropping packets every 5 or 10 pings from Windows CMD. Comcast said they have an outage and are working on it, but who knows when that will happen.
Hopefully this helps!
So couple of things. Since WAN smoothing only affects outbound traffic, you need wan smoothing turned on everywhere so that all VoIP traffic in all directions is replicated - and it sounds like you have this set up - so all good.
You then need to look at bandwidth usage of non VoIP traffic. Are your links getting congested with non VoIP traffic perhaps? Do you indeed have any non VoiP traffic going over the VPN?
If you do, and I assume you do since you say that LTE data is being used up quickly, then I would look at breaking out internet traffic direct on the comcast links and just using SpeedFusion Wan smoothing for VoIP traffic alone.
If you have non VoIP traffic that needs to go over the VPN too (so site to site data traffic rather than direct to internet traffic) then I would suggest you use a subtunnel within your VPN profile specifically for VoIP, then just turn on WAN smoothing for that VoIP sub tunnel, so you are not replicating all the data traffic over the LTE connection.
You could also add in an additional receive buffer just for VoIP. User experience on a VoIP call with as much as 240ms will be quite acceptable, so you could add in a little additional inline buffering to improve the call quality too.
Did you make sure your voip traffic goes via the vpn? Check the active sessions under the status tab.
We are pushing RDP and voip only thought the vpn and need both to be smooth.
Daily usage is 2.x GB will only give us about 10 days before slowdown.
When I had wan smoothing enabled everywhere there was packet loss pinging though vpn and voice issues.
As soon as I disable wan at site 1 and turn off wan smoothing and only have cell turned on though vpn all works fine.
Phone server is at Site 1 with PRI also more phone equipment at site 2 with another PRI however phone calls mostly go into site 1 then though vpn to site 2.
Sounds like you’ll need to fault find this in stages. Set up WAN smoothing between site 1 and the fusionhub. Ping the fusionhub from two or three devices on site 1 continiously and see if you have any packet loss. If not turn on wan smoothing for site 2. If you do have packet loss in this first step then monitor latency and packet loss in the speedfusion status view on the device webui (In Status → SpeedFusion) to identify where its coming from.
Click the little graph button to get a more detailed view over time.
Post a screen grab of the graphs and we’ll see what is to be learnt there…
Keeping an eye on the SpeedFusion Status graphs on all devices at the same time will ultimately show you where the issue lies I think.
Hi @seb82. I’ve read all the thoughtful comments which follow and I have no substantive additions. Indeed, I’d be quite reticent to correct or add much to the responses of the experts who have responded. 
But may I made a “side” comment? We are fighting Comcast over the exact same issue. Their network seems to periodically drop packets. We;ve verified this with PingPlotter and other tools. This is not a “Peplink issue.” We’ve tried WAN health checks via ping and DNS look-up and no matter what we do vis-a-vis frequency, wait time, etc, periodically the Comcast WANs at two locations are marked off-line by the Peplink routers as a result of failing health checks.
Complaints to Comcast have resulted in – predictably – no relief. No “outages” and no help from them. We’re going to continue fighting but we’re also looking for alternatives to this very shaky ISP. (We’ll see how this search goes --given that the US Government is fine with monopolies. We’re not optimistic we’ll find a reasonable, cost-effective solution.)
