A home user installed a MikroTik L009UiGS-RM router to replace an ISP-supplied device, configured PPPoE over VLAN 35, enabled Wi-Fi using CAPsMAN, and applied FQ-CoDel queuing to reduce latency.

The user wanted greater control over traffic, improved performance, and enhanced security after experiencing issues with the ISP router. The setup relied on wired connections to maximize speed between devices, used the router's PoE capability to power a separate access point, and took advantage of its compact form factor for wall mounting.

First, the user powered on the router, connected a laptop to a LAN port, and opened WinBox. After changing the default admin password, the management address was set and the default wireless bridge was removed. An Ethernet cable from the optical network terminal was plugged into port 1, the WAN port, which was then cloned to match the ISP router's MAC address to satisfy authentication requirements.

The ISP tags PPPoE traffic with VLAN 35. A VLAN interface named internet-vlan-35 was added on top of ether1. A PPPoE client was created on the VLAN interface, set to obtain the default route and use the ISP's DNS servers, and then credentials were entered. The default route appeared with pppoe-out1 as the gateway, confirming Internet connectivity.

Wi-Fi was configured through CAPsMAN. A security profile was defined with WPA2-PSK and WPA3-PSK, a passphrase was set, and channel templates for 2.4 GHz and 5 GHz were created. Two configuration entries, one for each band, were added to the bridge, and the CAPsMAN controller was enabled. On the access point, CAP mode was enabled and the radios were set to join the manager. Dynamic radio entries appeared in the router's interface list, indicating successful adoption.

During an evening test, the user observed increased latency while playing Counter-Strike. The spike coincided with an iOS update and a photo backup, which filled the router's packet queues. Bufferbloat, defined as excessive latency caused by overfilled queues, was measured with the Cloudflare Speed Test, which reports loaded and unloaded latency.

The user enabled FQ-CoDel queuing. A queue type was added with parameters targeting a 5 ms delay, a 100 ms interval, and a limit set to about 90 % of the measured upload and download speeds. The simple queue was applied to the PPPoE interface, and the fasttrack rule in the firewall was disabled to allow the queue to process all traffic. After the change, the Speed Test showed improved bufferbloat grade under upload load.

The configuration provides a functional home network with wired and wireless connectivity, and the queuing adjustment mitigates latency spikes caused by bulk traffic. The user noted that DS-Lite service may require a paid public IPv4 add-on for inbound access.