Layer 2 VPN - Speed Fusion packet loss

We implemented the following topology

Switch–Balance 710----MPLS ---- Balance 380 ---- Switch

Where there is a Speedfusion VPN between both balance devices that is being used to extend 3 VLANs between both sites across the MPLS network however, we are seeing constant packet loss only for packets going through the PepVPN, this is between PCs at both ends on the same VLAN.

Balance LAN interface is connected to an interface in the switch with 802.1q encapsulation, allowing only the VLANs we want to extend between sites.

We opened a support case with Peplink however, they have not been able to come with a solution nor, willing to work with us over the phone to fix the problem and my customers has been impacted for two days now due this issue. That is why I decided to post the issue here. Ticket #789222

Note that the MPLS link has 50 MB of BW and we have already tested them with the service provider and no errors were found. We ran RFC2544 tests.

Any help to solve this issue will be appreciated.

Hi Israel

Looking though the ticket I can see that this has been worked on by three of the teams around the world and has involved at least 2 calls with your selves. The engineering teams have also looked over the diagnostics and the network design.

This case is proving tricky to troubleshoot from our side due to the fact that the peplink devices have no internet connections and I believe we are now waiting for some temporary connections to be fitted so that we can perform some additional on device diagnostic.

As soon as this is done then we will be able to continue troubleshooting.

Thanks
James

James, what IP address on internet the PEPLINK requires access to? I am enabling remote assistance now

Knowing that support team is working with you via Peplink Remote assistance service & Hangout chat.

Working with the support team we found that the SP was limiting the UDP traffic somehow, we ran some tests with different ports until we stop seeing packet drops on the SF tunnel.

:+1::+1::+1::+1:

Please followup closely with the support team :blush::blush::blush:

Interesting. Some months back we started seeing something similar with our UAE customers.

Turns out the state telco was semi-randomly dropping UDP packets in order to preserve their international long distance revenues by disrupting anything that might be a voip connection. I wonder if this is similar?

James,

It looks similar. Because we ended up changing SF data port to TCP4900 since with UDP ports and TCP 18XXX or higher we were having packet loss.

1 Like