Wi-Fi Alliance plans Super G tests

I mentioned last time that The Tolly Group plans to soon release a report demonstrating that Netgear wireless LAN access points based on Atheros 108M bit/sec Super G chips degrade the throughput of nearby 802.11g devices by as much as 95%.  This is a complex situation, and I’d ask that you reread that article to grasp the salient technical points and caveats.

Since that article was written, apparently the Wi-Fi Alliance has decided to conduct its own tests. Reportedly, the alliance, which certifies IEEE-standard WLAN products for interoperability, will test “real-world” implementations of how Super G might affect a network in a nearby office suite, apartment, or house. No word at this writing of when we can expect the results.

Meanwhile, Tim Higgins, an independent tester of small-network products, has also tested the impact of Super G-based Netgear access points. One round of tests involved nearby Broadcom 802.11g-compliant access points; the other involved GlobespanVirata 802.11g-compliant access points as Netgear “neighbors.”

Higgins posts his findings at Tom’s Networking, a Web site for network enthusiasts (see “Related links” at the end of this article for the URLs linking to his results). His testing of Atheros-based Netgear equipment in the presence of Broadcom-based devices came out very much the same as The Tolly Group’s. However, he found less interference when the Netgear Super G device was used in the presence of the other manufacturer’s device.

The Super G issue began at Comdex last November, when Atheros competitor Broadcom demonstrated the “bad neighbor” issue and began talking it up. So an attendee asked about the issue at a panel session I led. One of my panelists, from Cisco, pounded his fist on the table and loudly advised those in the audience to, in effect, “just say no” to Super G. “Use standards!” he commanded.

Interestingly, many folks claiming to be “standards advocates” have plenty of their own proprietary technology and extensions to 802.11 standards. We’ll take a look at that issue next time.