2011/1/22 Daniel Halperin <dhalperi@xxxxxxxxxxxxxxxxx>: > 2011/1/21 Gábor Stefanik <netrolller.3d@xxxxxxxxx>: >> 2011/1/22 Gábor Stefanik <netrolller.3d@xxxxxxxxx>: >>>> about the market, how many 3-stream APs can you buy today? >>> >>> Well, there is the Airport Extreme, the D-Link DIR-665 (with a >>> Marvell-chipped mini-PCI card inside), and also an offering from >>> Senao > > Wow, seriously, thanks for the list! I had no idea Apple added 3 > streams to the Airport Extreme (the first gen and the first dual-band > one didn't have it); the only place i could find it was in the WFA > certificate. I also didn't know that D-Link made a plausibly useful > dual-band AP. That's awesome. > > Still, that's 3 out of 325 of the 802.11n APs on the WFA certified > list. Okay, let's say there's 10 more that neither of us mentioned, > that's still maximum 3%, or say 5%. And they all cost around $200. > For one randomly-selected (well, selected from their "Value-priced" > category) example, D-link's DIR-515 costs $38. > > A possibly silly, possibly illuminating question here: is there > anything *WRONG* with the 6300? Is there a reason you're looking for > a new iteration? What might be funny is that Wi-Fi, unlike CPU, > doesn't really get that much better over the years (as long as you > stick to the standard). And the newest, most exciting stuff -- e.g., > Wi-Di -- works with the 6300. I'm not really referring to the 6300 - I'm talking about what the 6350 would have been, true 3-stream 802.11n + WiMax. The 5350 had only partial WiMax implementation, and even that is gone now that the 5350 has been discontinued. Also, if they could pull off 3-stream+WiMax, and they are doing 2-stream+Bluetooth 3.0 now, I'm pretty sure 3-stream+BT3.0 would be doable. > > Why might you want a new version? Lower cost? -- the chips already > cost like $30, that's pretty cheap, say Intel could bring it down to > $25 would that increase volume enough to make it worth while? I doubt > it. To save power? -- well, 6300 is a pretty feature-rich chip, it > probably goes in devices like fancy laptops where it isn't a > relatively large power drain, and pretty efficient use will come from > Wi-Fi sleep modes. Power efficiency matters more when Wi-Fi is a > relatively large contributor to power draw. Oh wait, that's in the > small devices that only want 1x1 or 1x2, so maybe it IS worth Intel's > time/money to keep revving those chips every time they can use a > smaller process. Well, power efficiency does make sense for higher-end devices too - expect that instead of lowering consumption, it should go into higher TX power at the same consumption level. AFAIK both the 5xxx and 6xxx series are capped @ 15dBm, while regulatory allows at least 20dBm, and even more than that in some places. Also, by creating lots of different non-3-stream devices, Intel is indirectly discouraging OEMs from using 3-stream technology (because there is only one 3-stream offering, the 6300, while in the 1- and 2-stream category, there are lots of different designs with different bells and whistles each). While they did innovate in the high-end range with the 5300 and the 6300, they went *backwards* in the mainstream - remember how the 4965 was 3x3:2x2 (antennas/streams), while its mainstream successor, the 5100, was 1x2:1x2 - 50% upstream performance loss, plus losing all the benefits of extra antenna diversity. Even the 6200 is only 2x2:2x2, still weaker than the 4965. > > I don't know anything about Intel's internal plans, and this is all my > own speculation, but ... I'd love to hear what it is you'd like the > new chip to have that the current ones don't. Certainly, no one's > asking for them to rev the 3945, and I think for good reason. ;). > > Dan > -- Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-) -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
