 |
|
|
|||
| |||||
Interoperability
What wireless protocols does the device support? Can it be managed centrally to reduce admin work?
Futureproofing
Does the device include recent security standards? Will it grow with your needs?
ROI
Does the enhanced level of security the device provides justify the price?
Service
What is the warranty? What service and maintenance contracts are available?
We constructed a series of tests to determine the effectiveness of each package's ability to secure a wireless connection, and if there was any reduction in performance as a result of enabling this security.
Test 1: netstumbler
We ran Netstumbler, a wireless network discovery, mapping, and logging tool on a generic clone notebook fitted with a dual-mode network card to see what information was being broadcast from the access points. NetStumbler logs information such as MAC addresses, SSIDs, and possible encryption such as WEP. All the access points in this review could hide their SSIDs effectively.
Test 2: performance
We developed an in-house test procedure, using a local FTP server connected directly to the LAN port of the access point with a crossover RJ45 network cable. This was kept totally separate from all other wired networks to ensure no wired LAN interference.
This test shows the maximum physical distance achieved between the client and the access point. We ran the test both with WPA security disabled and enabled, to see if there was a noticeable difference in performance as a result of using the secur-ity features.
First we tested each vendor's access point using an Acer Travelmate 8000 notebook with an Intel Pro Wireless 2200BG integrated network interface with no wireless security set up. This test provided a baseline result of a vendor non-specific connection between the notebook and the vendor's access point.
The second test used the same configuration with WPA-PSK encryption enabled on the wireless network. Interestingly some of the devices performed better with WPA-PSK encryption enabled while others took a performance hit. This goes to show it is worthwhile running performance tests against a baseline as we have done before deploying equipment to see if there are any overheads with various configurations or vendors.
