Inter VLAN Traffic (ACL blocking ARP)
Hi everyone,
I'm experiencing an issue with ACLs on Omada-managed switches. I have three rules configured:
Allow all traffic from GatewayIP to IoT
Allow all traffic from IoT to GatewayIP
Allow all traffic from 192.168.30.9 to 192.168.30.4
Allow all traffic from 192.168.30.4 to 192.168.30.9
Deny all traffic within the IoT VLAN (source: IoT → destination: IoT)
As soon as I enable the deny rule, communication between .9 and .4 becomes unstable. Pings from .9 to .4 occasionally show very high latency (hundreds of ms), followed by a series of "Destination Host Unreachable" messages. This happens even though the specific allow rules for those two IPs are placed above the deny rule.
After some testing, I found that ARP is the problem:
When I clear the ARP cache on .9, it sends a broadcast ARP request for .4, but doesn't receive a response.
During that time, ping fails.
Eventually, it gets a response and communication resumes.
Static ARP entries temporarily resolve the issue.
So it appears that the switch-based ACL is blocking ARP replies from .4 to .9, even though I have IP-based allow rules in place. Since ARP is a Layer 2 protocol, it seems to be affected by the broad "deny IoT → IoT" rule.
Unfortunately, I cannot use Gateway ACLs – only Switch ACLs are available in my setup.
My questions:
Is there a way to allow ARP replies between two specific hosts while still denying all other intra-VLAN communication?
Does Omada offer protocol-level control in Switch ACLs (e.g. allow ARP or filter by EtherType)?
Are there plans to support stateful ACLs or more granular protocol control?
Any suggestions or workarounds would be appreciated!
Thanks.
