Cisco Support Community
Showing results for 
Search instead for 
Did you mean: 

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. And see here for current known issues.

New Member

Traffic flow stops after 30 mins when proxy arp is disabled

We configuring a cisco 2811 router for internet access. One ethernet interface of this router connects to the lan and the other to the ISP. As part of hardening we disabled proxy arp on both ethernet interfaces.

We noticed that the traffic flow through the router stops after 30 minutes.

When the traffic is not flowing , users on the lan can ping the LAN ethernet interface , but not the Internet Ethernet Interface.

A simple ping to the Internet from the router works fine. However , an extended ping sourced from the LAN interface fails.

The same router configuration with proxy-arp enabled on both ethernet interfaces , the traffic flows smoothly even after 30 minutes.

Can disabling proxy arp cause such a behaviour ?


Re: Traffic flow stops after 30 mins when proxy arp is disabled


One possible reason for this could be that your ISP has configured a subnet mask on its interface facing you to be different to what you have configured on your interface to the ISP. For example, say the ISP gave you the network and you split it into and Also assume that the ISP used the mask on its own interface to you. If you then assigned the first network to your WAN link and the second to your internet link, the following will happen: when the ISP wishes to get to an address in the second subnet, it will think it's on the same network as its interface to your router. It will then send out an ARP for it. WHen proxy-ARP is enabled, your router will respond with its own MAC address. If proxy-ARP is not enabled, this will fail.

Does this match your scenario ?

Hope that helps - pls rate the post if it does.



New Member

Re: Traffic flow stops after 30 mins when proxy arp is disabled

Thank you for te suggestions. The scenario appears to be similar. Need to check if this is indeed causing the problem.