Labs

Wireless routers

We test the wireless routers in a real-world situation by installing them in a real office. This is because, unless you live in a remote location, the reality today is that you're bound to find other wireless networks already there, spilling over from your neighbours or other businesses in the vicinity.

We could take the routers into the middle of a field, miles from civilisation to test the best-case scenario, but the results wouldn't be representative of what you're likely to see in a crowded Wi-Fi environment.

Thus, the 11 routers on test all have to contend with the challenges of interference from other networks, PCs, plus obstacles such as walls and furniture. We install each router in the same spot, knowing that neighbouring networks are operating, but we run tests several times to ensure results aren't compromised by heavy activity at the precise time of a single test.

To ensure the best performance, we update each router and corresponding PC Card or USB dongle with the latest firmware, and install the latest drivers.

Throughput and coverage

To test transfer speed, we hook up a Samsung X60 notebook via an Ethernet cable to each router. The matching PC Card or USB adapter is installed in a Fujitsu Siemens Lifebook S2110.

In our short-range test, we place the Lifebook 10m away from the router, but within the same room. We transfer a 100MB file to the Lifebook three times and average the times taken before calculating the average throughput in Mb/sec. We then stream a 720p high-definition trailer from the X60 to the Lifebook, giving the router a "pass" if the audio and video were smooth throughout, or a "fail" if it stuttered at all during playback.

To test long-range throughput, we take the Lifebook two rooms away (approximately 20m from the router). The signal has to pass through two partition walls along the way, and there are several metal filing cabinets against one of the walls. The same two tests are repeated in this position.

Finally, to test coverage, we use a signal monitor utility and continue to move away from each router until we can no longer obtain a connection. We factor these distances into the performance score. Most of the routers here failed to provide a strong enough signal at a distance of 30m and through three walls. Naturally, this is only a guide - coverage will vary, depending on your specific building and the amount of interference from other wireless networks.

Results

To give you an idea of what these results allow you to do, consider the bandwidth requirements for various multimedia data types. If you want to stream an MP3, that needs no more than 0.2Mb/sec. If you want to stream uncompressed CD audio (by sharing a CD drive under Windows), you need 1.4Mb/sec. Video is usually much more demanding, although Video CDs need just 2Mb/sec. Highly compressed DivX files require 4Mb/sec. If you have plans of streaming an MPEG2 DVD movie wirelessly, you'll need to have a constant 11Mb/sec. High-definition content with multichannel surround sound is out of the question for most of these routers at long range, although at short range it's now a realistic proposition, with the Belkin, Buffalo, D-Link, Linksys and Zyxel routers all stretching past the 25Mb/sec mark.

The final thing to remember is that firmware upgrades can yield big performance improvements so, while we've used the latest versions available for testing, future versions could improve routers that didn't perform so well here.

Ratings

At the end of each review, we give four star ratings. The Features & Design score is calculated by awarding points to each router's capabilities, such as security, firewall options and advanced features, and the warranty. The performance score is calculated using weighted results from our short-range, long-range and coverage test results. We then factor in the price including delivery (adding an ADSL modem for routers without one built-in) for the Value for Money score. Finally, the Overall score is a straight average of the other three ratings.