Wireless Features
The K3C defaults to using separate SSID's for each band and auto channel setting. It made a poor choice on 2.4 GHz, setting up on Channel 7 to avoid my house WLAN on Channel 11, but at least used 20 MHz bandwidth mode. I wish consumer router manufacturers would follow what Pros do and use only Channels 1, 6 and 11 in 2.4 GHz.
|
Time Switch in the Basic Wi-Fi settings shown below controls Wireless on/off scheduling. As near I can tell, the only wireless security mode supported is WPA2/AES PSK; WPA, WEP and WPA or WPA2 Enterprise (RADIUS) are not supported. Wi-Fi Protected Setup (WPS) seems an odd omission on a consumer router. But since "mesh" Wi-Fi systems have opted to not support it due to security risks, I suppose it's ok for the K3C, too.
K3C basic wireless settings
The Advanced settings don't include anything to control band steering (which Phicomm says is not supported) or airtime fairness, but do let you disable MU-MIMO and beamforming. The AP Isolation Help description says this restricts clients to internet access only. But when I enabled it for 2.4 GHz and connected a Windows 10 notebook, I could reach Ethernet-connected network shares just fine.
K3C advanced wireless settings
Guest Wi-Fi gets its own menu page but supports only 2.4 GHz. Transmit power control is found under the Signal Control menu and offers Low, Medium and High (default) settings, which I assume apply to both bands.
A setting I recommend you don't play with, at least until it's less buggy, is Wireless Extension. This does not set up wireless bridging/repeating on the LAN side of the router. Instead, it uses Wi-Fi as a WAN connection. This is handy if you rely on a wireless ISP for your internet or plan to use the K3C as a travel router. But I found once it was turned on, I couldn't turn it off due to the way the confirmation box blocked me from actually throwing the disable switch. I had to reset the router to default to clear this mode.
K3C advanced wireless settings
Storage Performance
Our standard router storage test procedure was used to measure file copy throughput for FAT32 and NTFS volumes connected via USB 3.0 only because there is no USB 2.0 connector. Media server and Samba encryption were first disabled and USB 3.0 mode enabled to ensure USB 3.0 operation.
The results below are not much to write home about. The FAT32 results landed the K3C dead last for that benchmark for write and read. NTFS results weren't much better, with the K3C second from last for write and fourth from the bottom for read. Note that best-in-class routers achieve over 100 MB/s in all these benchmarks.
Test Description | Phicomm K3C |
---|---|
FAT32 Write | 18.5 |
FAT32 Read | 34.3 |
NTFS Write | 30.5 |
NTFS Read | 60.2 |
Firmware Version | 33.1.33.197 |
Table 2: Storage performance (MB/s)
Routing Performance
The K3C was tested with the V10 router test process, loaded with 33.1.33.197 firmware. Although the simple iperf3 based WAN - LAN & LAN-WAN Throughput tests maxed out at 941 Mbps, the tougher HTTP based tests showed a relatively low score.
Test Description | Phicomm K3C |
---|---|
WAN - LAN Throughput (Mbps) | 941 |
LAN - WAN Throughput (Mbps) | 941 |
HTTP Score - WAN to LAN (%) | 35.1 |
HTTP Score - LAN to WAN (%) | 34.8 |
Bufferbloat Score- Down Avg. | 528 |
Bufferbloat Score- Down Max. | 420 |
Bufferbloat Score- Up Avg. | 453 |
Bufferbloat Score- Up Max. | 339 |
CTF Score (%) | 100 |
Firmware Version | 33.1.33.197 |
Table 3: Routing throughput
Scores on the HTTP Score benchmarks range from 68.1 to 0.3 WAN to LAN and 68.4 to 4.9 LAN to WAN. So the 34.1 and 34.8 land the K3C in the middle of the pack. The plots below show the breakdown for the four filesizes used for the top product for both benchmarks, ASUS' GT-AC5300 the K3C and two other AC1900 class routers, D-Link's DIR-878 and the venerable NETGEAR R7000 Nighthawk. Looks like the K3C could use some improvement in handling smaller file sizes.
HTTP Score comparison
Plot key file size: [A] 2 KB, [B] 10 KB, [C] 108 KB and [D] 759 KB file
Average Bufferbloat scores again place the K3C in the middle of the tested products. But I've found bufferbloat, at least using our benchmark method, shouldn't be much of a concern, since converting the results back to milliseconds of latency [(1/score) x 1000] yields averages of 1.9 ms downlink and 2.2 ms uplink.
The Cut Through Forwarding tests look for throughput reduction when various router features are used. The 100 score indicates no throughput reduction was found when individually engaging Firewall DoS, Parental Control, QoS or URL Filter features.