We used PerformanceTest 6.1 by PassMark which you can find out more information about here. It has a handy Advanced Networking section which is perfect for our testing.
Doing our best to emulate a real-world performance scenario, I setup a server running Windows 2008 (x64) Server, (2x Xeon Quad Core 3.0 GHz 1333MHz FSB, 4GB DDR2 FB-DIMMS, 2x 146GB SAS drives in RAID 1) and the client PC was an ASUS G51-VX With built-in Intel 802.11n (Intel WiFi Link 5100 AGN) adapter. The results were gathered by sending data from the ASUS notebook to the server at different distances with the built in adapter and a TPLink TL-WN821N. Average transmission speeds were recorded for each.
- Connection Speed
As with most N Spec routers, there is still a problem when you have a mixed mode network. This happens when you still have B or G spec products in your environment.
The RT-N13U has this as well. If you need to leave B and G enabled your connection speed tends to limit itself to around half the speed or about 75Mbps. This is very annoying for anyone with an N spec adapter. But again, it is a flaw with almost all routers these days, but I do hope they can fix it soon.
As usual, to test the speed of the RT-N13U I chose three common working points inside my house. One was in the lab within 10 feet of the routers; the next was in the bedroom roughly 35 feet away and requiring the signal to travel through a wall containing the main house electrical panel and a "wet wall". The last was outside on the back porch, roughly 45 feet from the router. Both the second and third positions were out of the direct line of sight of the router.
*signal travelling through wet wall and main house electrical panel
The RT-N13U does not do too badly. I do think that the lack of an external antenna (or grouping of antennas) does hurt it when dealing with long ranges and internal wireless cards (which also have small antennas).
Gaming latency is a problem for many. This is even truer as background traffic in the form of YouTube streaming, Netflix and much more clutters up our connection to gaming servers. For the RT-N13U, we wanted to test it with typical background traffic in both open and with EZ-QoS set to prioritize gaming traffic. We connected to the same game server for Counter Strike three times and took the middle "ping" time from each connection type (wired with and without EZ-QoS and Wireless with and without EZ-QoS).
Wired - EZ QoS
Wired - Without EZ-QoS
Wireless - EZ QoS
Wireless - Without EZ-QoS
Now look at that; it looks like the EZ-QoS setting for gaming while connected directly did almost nothing. I was surprised (very surprised) to see this happen. I checked using the D-Link DI-655 and received similar results. This makes me think that either the EZ-QoS profiles are not working properly over the wired connection or due to the lack of server load there was no difference noted between the two.