Testing

To determine what kinds of speeds I was getting from my setup, I used a program called Qcheck that is distributed by IXIA. This free utility allows for the determination of response time (latency) and throughput. To do this, the program has to be installed on all machines used in the test. The program runs as a service in the background. It does enable you to test from either direction on any two machines that it is installed on, (i.e. A to B, or B to A).

Each test ran with the same parameters, and I ran it 5 times to get a good average. The final results are the average from 3 different machines all located on the same switch. This should compensate for differences in hardware and networking components between each machine. I tested the USB device attached to my desktop system, while moving the NeverWire 14 device to different outlets in my home.

Response Time Testing

To determine the latency (time from point A to point B), I employed the Qcheck from my source machine to 3 other machines to give a good average read. For the response time runs, I set the program to do 10 iterations with a packet size of 32 kBytes. This is the largest size that can be tested with this software. The test system where the USB is located is in the basement of my home. Outlets labeled as close are a crossed the room on the opposite wall. The Isolated Outlet Upstairs is a separate 20 Amp circuit used for a window AC unit and is thus isolated from the rest of the circuits in the house. The other upstairs outlet is approximately 1 ft to left of the isolated one. The Close (Surge Supressor) is just that, the unit was plugged into the same outlet, except that it was routed through a surge suppressor. This type of hookup is expressly warned against in the manual, and you can see why: the 50% increase in response time. The 100 Mb is the stock NIC in my machine which was disabled during testing.

Response Time Testing (lower is better)

Close Same Room

Close (Surge Protector)

Upstairs Outlet Opposite End

Upstairs Isolated Opposite End

Hallway

Bedroom 1 Outlet # 1

Bedroom 1 Outlet # 2

Bedroom 2 Outlet # 1

Bedroom 3 Outlet # 1

100MB

51.07

125.60

64.07

54.67

63.40

55.60

55.23

51.93

53.20

4.53

88.11

0

Response Time (ms)

150

 

Average speed of 3 machines

 

For the most part, the speeds of the NeverWire USB stayed about the same. There was some variance between outlets, but nothing compared to plugging it into a surge suppressor. Still, it can not compete with a Fast Ethernet connection (100Mb).

Throughput Testing

For the determination of throughput, I used a data size setting of 1 MByte. This should give a good indication of the speed if you were attempting to transfer large emails or files. It also just happens to be the maximum the software will test at too, so 1 MByte it is.

Again, the results appear to be similar to the response time testing. As expected, the surge suppressor kills approximately 50% of the throughput. Another interesting thing to note is that I appear to have a bottleneck with machine 3. This only shows up in the 100 Mb Fast Ethernet connection test. Even though I have managed to drop my throughput to about 28 Mbps, this does not affect the NeverWire since its average throughput sits at 5 Mbps.