Feature Request: Immediate DHCP Renewal for IP Passthrough Mode

Krillsson

Hello Zyxel Team and Community,

I’d like to submit a feature request regarding the behavior of the IP passthrough mode on Zyxel routers, specifically the NR5101.

Current Behavior: I have a setup where a Unifi UDM Pro is connected to the Zyxel router's LAN1/WAN port, with the Zyxel configured in IP passthrough mode. The issue arises when the mobile broadband connection reconnects, causing the ISP to assign a new public IP address. In this case, the Zyxel router obtains the new IP, but the UDM Pro does not pick up the change immediately and continues to use the old IP address until it renews the DHCP lease (which can take up to 10 minutes). This results in an internet connection downtime, which could be avoided.

From my understanding, the Zyxel router, being aware of the new public IP, does not relay this change to the downstream device immediately. Instead, it waits for the UDM Pro to request a DHCP renewal on its own. This behavior feels counterintuitive because the Zyxel router is already aware of the updated IP, and there’s a significant delay before the downstream device realizes it needs to renew.

Feature Request: Would it be possible to implement a feature (either as a default behavior or an optional setting) where the Zyxel router proactively relays the new IP assignment to the downstream device as soon as the mobile broadband reconnects? This could minimize the connection downtime that occurs when the ISP renews the public IP, as the downstream router would be immediately notified.

Reasoning: Given that cellular networks are inherently prone to occasional disruptions, I believe this behavior would benefit a significant portion of your users who rely on IP passthrough mode for their network configurations. Reducing the downtime during IP renewals would improve the overall user experience.

I’d appreciate feedback from the Zyxel team and the community on this suggestion. Thank you!

PeterUK

I just got me a FWA510 and wonder If I will run in to this problem. I'm not sure how IP passthrough works but it seems like you say the unit does DHCP client then is a DHCP server and the two lease time mean you have downtime one fix is to set the DHCP lease low like 2/3 minutes. A better way is to have the DHCP forwarded when your end does a DHCP request and receive the lease time the unit gets and forwards that.

Edit: have posted a case for this issue

PeterUK

Here is a Wireshark of what goes on

Note time drift by PC vs VPN300

at 00:13:20 was the last packet on the old IP from EE
from nebula I ping to my VM to see new IP which the FWA510 has the new IP its side
at 00:51:55 I start see traffic from new IP due to DHCP renew which the FWA510 sets at 1 hour meaning a long down time wait which is why it needs to be less
at 00:57:52 is when DDNS updates the new IP to allow the tunnel