Airgo Networks Inc.s latest chip set provides unparalleled wireless throughput performance, far surpassing anything else on the market today and topping out at more than 100M bps. Although the chip set is targeted at consumer markets, IT administrators may find niche uses for the technology. However, they should be well-aware of the deleterious effect that products based on the technology can have on existing wireless LAN deployments.
Airgos True MIMO (multiple input, multiple output) Generation 3 chip set introduces ACE (Adaptive Channel Expansion), which leverages additional radio channels when necessary to push large amounts of data. For instance, a Gen 3 access point configured for Channel 1 will temporarily use Channel 5 when large amounts of bandwidth are needed.
Like previous versions of the chip set, Gen 3 uses two radios with multiple receive and transmit antennas to provide spatial multiplexing for increased performance.
To test the Gen 3 chip set, eWEEK Labs brought in access points and client adapters from both Linksys, a division of Cisco Systems Inc., and NetGear Inc. Linksys provided its new Wireless-G Broadband Router with SRX400, priced at $150, and its $100 Wireless-G Notebook Adapter with SRX400. NetGear submitted its $180 RangeMax 240 Wireless Router and $129 RangeMax 240 Wireless Notebook Adapter.
For each test, we wirelessly connected to each access point with a single True MIMO Gen 3 wireless adapter. We installed the adapter in a Dell Inc. Latitude D600 running Windows XP with Service Pack 2. Neither the Linksys nor the NetGear access points include Gigabit Ethernet switch ports, so we configured a pair of endpoints on the wired side of the network.
To measure throughput, we used Ixia Inc.s IxChariot 6 with Service Pack 1, using the High Performance test script. Each value in the graph at right represents the average throughput achieved over three successive 30-second test runs.
After examining the RF (radio frequency) landscape around our test location, we configured each access point to transmit primarily on Channel 1. Since the Gen 3 chip set uses ACE, the overflow channel was automatically set for Channel 5. On each access point, we used the default settings that balanced range with throughput.
During tests, the NetGear equipment provided slightly better throughput and range performance than the Linksys gear, but products from both vendors generally interoperated well.
We ran into some problems installing the software for the Linksys client adapter, but Linksys officials assured us that the problem has been fixed in retail units. In addition, when using the NetGear client adapter with the Linksys router, performance lagged significantly at 60 feet in our suite of tests. This shortcoming turned out to be a matter of antenna placement in association with some nearby walls; moving the client about 2 feet to the side spiked performance up to 60M bps, similar to that of the other access point/client combinations.
It is unlikely that users will run a Gen 3-enabled router only with Gen 3 clients, so we also looked at the performance of standard 802.11b gear with a Gen 3 access point, as well as the performance degradation that comes when introducing legacy clients to the Gen 3 network.
Using only a single Intel Corp. Centrino-based 802.11b client connected to a Gen 3 router, we were able to reach 4.9M-bps throughput. We then ran tests using a Gen 3 client simultaneously with an 802.11b client connected to the same Gen 3 router. The MIMO performance dropped to 74.9M bps, while the 802.11b performance slowed to 1.9M bps.
Disturbingly, we found that Gen 3 equipment can have damaging effects on neighboring wireless networks.
Using an off-the-shelf 802.11g access point and client adapter, we configured a separate 802.11g network on Channel 6. Under normal circumstances, we were able to push 23M bps through this network.
We then tested both networks simultaneously. While the Gen 3 network was configured on Channel 1, ACE automatically used Channel 5 during tests. With both networks under test, the Gen 3 network reached 93.7M bps, but, due to heavy co-channel interference, the 802.11g network lagged considerably—down to 3.5M bps.
Administrators therefore should be warned that they must be aware of the new technology, even if they have no intention of implementing it—a Gen 3 access point deployed in the cafe in a company lobby could very well bring an enterprise WLAN (wireless LAN) deployment to a crawl without careful evaluation and oversight.
Unfortunately, we have not yet seen wireless management tools up to the task of helping identify and troubleshoot MIMO-related RF problems. During tests, we used AirMagnet Inc.s latest edition of Laptop Analyzer, Version 6.0, to track environmental conditions, but the tool could register only a sharp increase in co-channel interference, without being more specific about the actual cause.
Technical Analyst Andrew Garcia can be reached at firstname.lastname@example.org.