The router was flashed to the latest firmware, SPI firewall disabled and the machine on the LAN side is placed in the DMZ.
For the testing of the router's routing throughput, we used two machines, one connected to the WAN port of the router and the other to the LAN port as shown in Figure 1.1. For both WAN to LAN and LAN to WAN tests, we ran two benchmark software, PassMark Performance Test and IPerf. Each test was repeated three times and the average is taken as the final result. Figure 1.2 shows the following parameters used on the IPerf.
For testing of VPN throughput, we came up a scenario where the VPN client is on a different subnet from the router WAN. We accomplished this by creating a virtualized environment. RT_A, RT_B and GW are running Vyatta. Figure 1.4 shows the network topology for this setup.
For wireless throughput tests, one machine is connected to the LAN port and the other by wireless. A 200MB file is then transfered from the LAN machine over to the wireless machine and vice versa. This is to test the upstream and downstream quality of the wireless link. The wireless connection is encrypted with WPA2-AES. The test is performed on both the 40MHz channel and the 80MHz channel (5 GHz only). The wireless client is then moved to different rooms (Figure 1.5) and the tests are repeated again. This is to compare the transfer quality when the signal line of sight is being obstructed.
These are the hardware specs of the machines used in this testing.
OS: Microsoft Windows 7 Sp1 64-bit
CPU: I7 920 @ 3.4GHz
RAM: 8GB DDR3
NIC: Intel Pro/1000 ET Dual
OS: Microsoft Windows 7 Sp1 64-bit
CPU: AMD Turion X2 RM-74 2.2GHz
RAM: 3GB DDR2
NIC: Realtek 8111B
WiFi: Asus RT-AC66U in Bridge mode
|Direct Connection||945 Mbps|
|Wan to Lan||938 Mbps|
|Lan to Wan||932 Mbps|
The advertised specs was up to 300,000 simultaneous connections but we maxed out the connection on the 4 pairs which was a total of 182,636 connections, which admittingly is still more than enough or needed for the usual p2p applications.
|USB Storage Throughput|
|SAMBA||160.8 Mbps||108 Mbps|
|FTP||176 Mbps||104 Mbps|
For this test we attached a 2.5" portable hard drive to the USB port. The drive was formatted to NTFS. We then transfered a 200MB data payload to the drive. From the results, it looks like writing data to the drive using FTP yield slightly higher speeds as compared to SAMBA. However, reading data using SAMBA gives higher speeds as compared to FTP.
|128bit MPPE Encryption|
|Vpn to Lan||74 Mbps|
|Lan to Vpn||45 Mbps|
The VPN server doesn't work when DMZ is enabled, which is an issue that we've highlighted during the review of the RT-N66U.
|2.4 GHz Wireless Throughput|
|2.4 GHz 40Mhz Downstream||67.5 Mbps||62.3 Mbps||33.6 Mbps||64.4 Mbps||65.0Mbps|
|2.4 GHz 40Mhz Upstream||73.7 Mbps||70.0 Mbps||41.8 Mbps||69.1 Mbps||68.3 Mbps|
Not much improvement on the 2.4GHz band as compared to the RT-N66U.
|5 GHz Wireless Throughput|
|5 GHz 80Mhz Downstream||165.6 Mbps||160.4 Mbps||78.2 Mbps||161.6 Mbps||157.1 Mbps|
|5 GHz 80Mhz Upstream||210.1 Mbps||198.9 Mbps||91.5 Mbps||200.4 Mbps||171.9 Mbps|
|5 GHz 40Mhz Downstream||120.8 Mbps||108.3 Mbps||56.0 Mbps||108.6 Mbps||111.7 Mbps|
|5 GHz 40Mhz Upstream||175.2 Mbps||169.6 Mbps||62.7 Mbps||165.9 Mbps||167.2 Mbps|
On the 5GHz band there is a slight performance improvement on the 40MHz as compared to the RT-N66U. It seems like there is an overall 20% improvement on the 80MHz. At location C which has the poorest coverage with the 2.4GHz band, does have a better throughput performance with the 802.11ac.