Testing

Finally, we are down to the testing. What I will be testing is throughput or Mbps (Megabits per second), not to be confused with MBps (Megabytes per second). To determine what kinds of speeds I was getting from my setup, I utilized 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 called Endpoint in the background. It enables you to test from either direction on any two machines that it is installed on, (i.e. A to B, or B to A). All throughput testing was done with a 1,000 KB setting in Qcheck. The scores are an average of 20 runs, 10 runs from the server to wireless connection and another from the wireless connection to server.

For the laptop testing, the AP was set in the center of the house on a workbench. The laptop was then tested while I was sitting on the couch. This is a distance of about 25 ft with one wall standing between the notebook and the router. I would consider this to be an average amount of interference and I think it will provide a good approximation of the average signal quality. During testing, signal strength varied between 75-80% signal strength. The desktop was tested with the wireless access point sitting on top of it. This is essentially at 0' and should provide for the fastest possible transmission rate.

Since Qcheck requires two ends to test, I will be using a new Dell Poweredge SC420 Server and my test system that contains the A7N8X-E Deluxe motherboard with onboard LAN.

For my first test, I will be testing the ability of the router to provide the maximum signal. To do this, I placed the router on top of the test system, essentially at 0 ft. I tested both the "a" channel on all settings and the "g" channel on all settings. Each one was tested with the other radio signal disabled. The Throughput testing was performed with WPA-PSK (no server) enabled using the AES encryption. Preliminary testing showed no discernible difference with varying security in place (WEP vs WPA). Therefore the numbers being reported are for what I would consider an optimal setup (i.e. not open).

PCI Adapter Throughput at Various Settings
(Show All Graphs) (Collapse Graphs)

802.11a

802.11g

108 Mbps Only

Auto 108 Mbps

802.11a Only

10/100 Wired

41.722

23.053

23.039

94.118

0

(Mbps)

125

 

 

802.11a

802.11g

108 Mbps Only

Auto 108 Mbps

802.11g Only

802.11g and 802.11b

802.11b Only

10/100 Wired

30.091

10.534

12.784

10.087

5.175

94.118

0

(Mbps)

125

 

 

Speed vs. Router Settings

All speeds are listed in Mbps (Megabits per second). As you can see from the graph, how you set your router up will determine your speeds. What it all boils down to is backward-compatibility as well as multi-vendor support. Enabling the 108 Mbps only gave the highest reading for both the PCI card and the PCMCIA card, no surprise there. Unfortunately, if you are using other vendor's products, the 108 only setting may be incompatible. In that case you will need to implement the g only setting that will knock down your throughput by about 50%. Going further back down the wireless evolutionary tree, enabling support for the 802.11b protocol only will reduce your throughput by a factor of 10. Basically, unless you are using Belkin's products only, buying speed-boosted wireless is a waste of money. Of course if speed is what you desire, just use the built-in 4 port switch with measured speeds of 92 Mbps!

It has been my experience in the past that once a speed-boosted product is implemented, all other non-conforming products have a terrible time connecting. Imagine my surprise to find that the D-Link laptop card I previously reviewed connected to the 108 only setting of the Belkin router. Since both cards use the Atheros Super G chip to get the higher speeds, they speak the same language. On average I received a 40 Mbps connection speed on the laptop.

802.11a vs 802.11g

Since this router has both an "a" and a "g" radio, you have the choice of connecting to either. Since "a" is an older standard, you might ask yourself why they have included it on this product. The short answer is that "a" operates on the 5Ghz band that is far less crowded than the 2.4 GHz that "b" and "g" both operate on. The major drawback to "a" becomes very apparent as you begin to move further away from the access point. The "a" signal will drop very quickly compared to the "g". A Netstumbler scan of both signals shows just how weak the "a" is when compared to the "g". Even with the decreased range, I was still able to get a good "a" wireless signal through most of my house. This is with the AP in the basement. A secondary advantage to using "a" rather than "g" is that the signal will not be picked up unless you use a card that supports "a". Most laptops and network cards sold today only include the "b" and "g" bands.


PCI Adapter Test Scan:


PCMCIA Notebook Adapter Test Scan: