Besides measuring the speed of the router’s network interfaces, we will also compare it with the TEW-633GR, another router from Trendnet. We have mentioned this model repeatedly in this article and we are also interested to see what particular advantages the TEW-672GR has, besides the support for two frequency bands, in comparison with the TEW-633GR if they cost the same money. So far, we haven’t found anything special in it.
Here is the equipment and software we used for this test session, besides the TEW-672GR:
- Trendnet TEW-633GR router
- Two Category 5e Ethernet cables
- Intel Centrino notebook
- PC (Core 2 Duo X6800, 4GB RAM, Realtek RTL8168, Ubuntu 8.10)
- Linksys WPC300N v.2 PC-card (Atheros)
- Edimax EW7708Pn PC-card (Ralink)
- IxChariot (High_Performance_Throughput and Throughtput scenarios)
We measure the maximum data-transfer rate of the router’s wireless interface by sending data between two of its LAN ports.
The speed is okay if you don’t mind the occasional slumps – we haven’t seen them when we tested other routers.
Next we measured the bandwidth of the router’s WAN-LAN segment. We tested a direct connection as well as a VPN tunnel. As you know, a VPN connection is almost always accompanied with a considerable reduction in the data-transfer rate. We tested the direct connection by specifying the WAN port address manually and exchanging data between the two endpoints in both directions. When testing a VPN connection, we chose the PPPoE protocol. We established a PPPoE server on the Linux machine (see the list of test equipment above) and created an endpoint. Then we connected the tested rooter to that server and attached a notebook with IxChariot to the router’s LAN. Here are the results for both routers.
So, the TEW-672GR doesn’t show anything exceptional here. The LAN ports of this router are quite up to the Gigabit Ethernet standard in speed, but the speed of the WAN port is almost as low as Fast Ethernet, which is especially obvious with the VPN tunnel. The results of this test depend on the router’s processor, and we know that the processor of the TEW-672GR is based on a cheap SoC-controller.
The TEW-672GR has quite a lot of system memory, which may have an effect on the number of simultaneous connections the router can support. To perform this test we created a pair in IxChariot using the Throughput scenario in which we changed the file_size parameter from 100,000 to 1,000,000. Then we began to increase the number of pairs by replicating them until there were errors during the test. This test was performed for both direct and VPN (PPPoE) connections. We stopped at 700 direct connections and didn’t go any further due to the limitations of the test program. With PPPoE, the TEW-672GR reached 190 simultaneous connections whereas the TEW-633GR maintained 700 simultaneous connections.
The wireless connection in draft 802.11n mode was tested using WPA2 PSK encryption with the AES algorithm. These security settings are the default ones in the draft version of the new standard and are likely to remain such in the final version. The TEW-672GR is based on a Ralink chipset, so we tested it with a Wi-Fi card based on the same chipset (Edimax) as well as with a Wi-Fi card based on an Atheros chipset (Linksys) in order to compare the performance with the TEW-633GR.
And again we have to note that the TEW-672GR is inferior to its opponent. The Gigabit Ethernet router with Draft N support has no advantage when it comes to wireless connection speed.
Finally we tested the router’s coverage at different distances and with different obstacles. We measured the signal level in five points:
Point 1: Near the router
Point 2: At a distance of 4 meters without obstacles
Point 3: At a distance of 5 meters + two thin gypsum wallboards
Point 4: At a distance of 6 meters + one brick wall, about 30cm thick
Point 5: At a distance of 17 meters + one thin gypsum wallboard and two 30cm brick walls
Here are the results:
The TEW-633GR wins the last test, too. The TEW-672GR had a nearly zero speed at the last point, which is bad. It means the TEW-672GR is inferior to the TEW-633GR in every performance-related test. We guess this is all due to the relatively cheap hardware solutions as well as to imperfect firmware. We doubt the developer will be able to increase the router’s performance dramatically but perhaps its firmware can be improved so that the router was up to the high-end class it is positioned into.