Test Gear: Cisco Aironet 1142 And Aruba AP125
Ruckus' main competitor is the Cisco Aironet AIR-LAP1142N-A-K9 sitting alongside a Cisco 4402 controller running firmware 6.0.182.0. As mentioned before, this is our one possible candidate for chip-based beamforming to compare against’s Ruckus’s antenna-based beamforming (BeamFlex). The curious thing is that Cisco ships the 1140 with beamforming disabled.
Like all dual-band access points, Cisco uses two radios here, one for 2.4 GHz and the other for 5 GHz. The 1142 attaches three antennas to each radio, utilizing a 2x3 transmit/receive array. Cisco’s beamforming uses two transmit antennas, so it follows that you can’t have beamforming and spatial multiplexing operating concurrently.
So, it’s curious that Cisco opts to prefer spatial multiplexing out of the box rather than its shiny new on-chip technology. Perhaps this is because the feature is so new that the company wants to ease into adoption. But could it have something to do with performance? We’ll see.
Comparing the architecture of the 1142 against the Ruckus 7962 is intriguing. I couldn’t dissect the Cisco unit, but I did find these photos in some online patent documentation.
We also tested with an Aruba AP125 access point and Aruba 3200 controller with firmware 3.4.0.1 build 21611. There’s not much to report here. This is a standard issue, enterprise-class, dual-band access point with three antennas you’re never quite sure how to orient. The AP125 is pretty representative of “standard 802.11n,” and I view it here as a performance baseline for comparison against Cisco and Ruckus.
Rounding out the test platform, I used a Dell 620 notebook with the Broadcom NetXtreme 57XX Gigabit Ethernet controller as the server and a Lenovo X61 with an Intel 4965AGN adapter (driver version 12.4.0.21) as the target client. The switch tying everything together was a 3Com 3CDSG10PWR OfficeConnect.
- beamforming ,
- wifi ,
- ruckus
first post?!?!?!
No, REAL first, dipshit.
I don't think it's a problem that this is really only enterprise-class hardware. The very fact that there's an tenna sensitivity that can cripple the entire system shows that for Joe Apefist this is too much trouble for its own worth.
But, the tech shows amazing potential and given some tweaking time, I'm sure it will become more robust and more economical and will rapidly see adoption at home.
Personally I can't wait!
As a crude guide if you want 10dB gain over omnidirectional (10x the power in some direction) then you need to have 10 antenae to cover all directions. It works but with an obvious price in money and size, and a more subtle one in intereference for/from other transmitters unlucky enough to be in the chosen direction.
Personally I'd prefer multiple omni basestations and just focus on minimising distance. Inverse square law is your friend.
Personally I'd prefer multiple omni basestations and just focus on minimising distance. Inverse square law is your friend.
:-)
Yeh totally why punch through 4 walls when you can punch through 2. Plus you can site access points/ repeaters in free space away from mwave reflective objects.
Also note that much more important to enterprise wireless LANs is NOT the raw AP to single client thoughput that so many of these gearhead tests do. We are constantly faced with offering stable and usable wifi for dozens to hundreds of concurrent users in crowded areas (conference centers, auditoriums....) Like any shared medium, Wifi suffers from co-channel interference and overly RF loud clients.
One BIG advantage that you will see enterprise vendors work towards is NOT how much speed to any one client you can get, but how much Reduced interference beamforming will allow to neighboring wireless APs in the same ESS. The net result is that all users see benefit of solid and stable wireless connectivity.