|
This is quite a can of
worms.
We just completed a WLAN
client test in our labs at Syracuse University, the results of which will be
published in Network Computing next month. I can't broadcast all of the detailed
results prior to publication, but I can tell you what we did and what we
discovered in general terms. I can provide more details to you privately if time
is critical.
We started this project
because we were curious about how commoditized notebook computers with embedded
wireless were. More specifically, we wanted to answer three
questions:
1. Does the choice of
radio module (Intel, Broadcom, Atheros) make a significant difference in
performance?
2. How much of
performance variation is attributable to the radio module or the system design
(including integrated antenna).
3. Does battery life vary
depending on radio module.
We focused our analysis
on multi-mode, dual-band integrated mini-PCI radio modules because we feel
strongly that all enterprises should be supporting abg for capacity
reasons.
We asked the four leading
notebook vendors (Dell, HP. Lenovo/IBM, and Toshiba) to send us 2 systems, one
Centrino and the other whatever other radio module they supported. Dell and HP
sent us Intel and Broadcom, Lenovo sent is Intel and Atheros, and Toshiba just
sent us Intel (Hello, Toshiba. Please follow directions).
We tested the radio
modules on an Azimuth test system that allowed us to attenuate signals is 1 dB
increments and measure the performance, across 11g and 11a. We discovered that
most systems had similar performance profiles with 11g but there were
substantial variations with 11a.
Because Azimuth testing
bypasses the antenna, we also did field testing in our building at 5 different
locations, between midnight and 6 am with our production network turned off. We
used a rotating turntable system and tested performance for each combination.
The bottom line is that system design does make a difference, though not as much
as some vendors may want you to believe.
Finally, we used BAPCO's
MobileMark to evaluate battery life, comparing identical systems with different
radio modules. Again, there were differences, though they weren't major.
Interestingly, we discovered that battery life is significantly shorter with
Ethernet than with wireless.
We also looked at client
utilities and formed some impressions about what we liked and didn't like. We
tried to test roaming but ran into some technical problems. Since we don't feel
roaming is a critical issue with notebook computers, we don't consider this to
be a huge issue but we'll continue to investigate it. We also did a sidebar on
Cisco CCX.
dm ********** Participation and subscription information for this EDUCAUSE Constituent Group discussion list can be found at http://www.educause.edu/groups/. |
- wireless NIC evaluation David Boyer
- RE: [WIRELESS-LAN] wireless NIC evaluation Dave Molta
