Throughput vs. Path Loss - Product Comparison
Moving on to some product-to-product performance comparison, Figures 28 and 29 pit the Extreme against the three next-best performing products that we've tested: The Netgear WNR854T (Marvell-based), Buffalo WZR-AG300NH dual-band (Marvell), and D-Link DIR-655 (Atheros). Note that all have gigabit WAN and LAN ports.
Since the Extreme doesn't support 40 MHz bandwidth mode in the 2.4 GHz band, those plots show only 20 MHz bandwidth mode performance. The Extreme leads the pack for downlink, followed by the D-Link DIR-655, which is also based on Atheros XSPAN silicon.
Figure 28: Throughput vs. Path Loss product comparison - 2.4 GHz downlink
The uplink comparison is less clear, with only a clear laggard, the Marvell-based Netgear WNR854T, which falls off faster than the others.
Figure 29: Throughput vs. Path Loss product comparison - 2.4 GHz uplink
For comparison in the 5 GHz band (Figures 30 and 31), there is only the Buffalo Nfiniti dual-band, which does better than the Extreme in downlink and 40 MHz bandwidth mode.
Figure 30: Throughput vs. Path Loss product comparison - 5 GHz downlink
The last plot (Figure 31), which shows 5 GHz uplink is pretty interesting. The Buffalo quits pretty early on and doesn't even hit the "waterfall" portion of its curve. I suspect that this is due to a link rate adjustment problem that caused the Buffalo client to disassociate, ending the test.
Figure 31: Throughput vs. Path Loss product comparison - 5 GHz downlink
The Extreme's relatively poor showing, compared to the tests made with the MacBook, is probably due to the mix of Atheros and Marvell chipsets. While manufacturers have made advancements in interoperability, the results show that you're still better off using "matched" products for maximum performance.
If you don't like these comparisons, you can run your own using the Wireless Charts tools!