802.11n Draft 2.0 gear still a neighbor killer

802.11n Draft 2.0 gear still a neighbor killer

Summary: As we get to the final draft stages of the 802.11n standard, it was thought that the neighbor interference issues have been put to rest by the Draft 2.

SHARE:
TOPICS: Wi-Fi, Intel
19

As we get to the final draft stages of the 802.11n standard, it was thought that the neighbor interference issues have been put to rest by the Draft 2.0.  The new standard supposedly has a strict ban on dual-channel operation whenever any legacy devices are operating.  The product is suppose to back off from 40 MHz wide bands to 20 MHz so that it isn't hogging most of the 2.4 GHz spectrum.  The only problem is that it doesn't work according to Tim Higgins' tests.  Without this functionality working, 802.11n products are simply a legal way to jam your neighbor's 802.11b/g Access Points.

The problem here is that consumers don't know that this causes their neighbor problems and even if they did, it "isn't their problem" since they're causing their neighbors grief.  Furthermore this is an underhanded way of driving adoption of new 802.11n gear through an arms race.  The intent of Draft 2.0 was to fix these serious problem from Draft 1.0 with strict operation rules but it seems that it simply doesn't work as advertised.  Intel on the other hand seems to be doing the right thing with their Centrino certification rules that ban all 40 MHz operation in the 2.4 GHz band.  The IEEE 802.11 standards body should be considering the exact same ban.

Topics: Wi-Fi, Intel

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

19 comments
Log in or register to join the discussion
  • That may be true for D-Link firmware updates but . . .

    they probably weren't tested by the Wi-Fi alliance because the alliance just started testing yesterday for chipsets. I wonder if meeting Draft 2.0 specifications means extra hardware inside the box.

    However thanks for pointing out the review. Is D-Link just creating extra hype by saying you can achieve Draft 2.0 compliance without waiting until August or is it really the fault of the IEEE?
    n.stockwell@...
  • Wish there was a WU for 802.11N gear!

    Too bad we can't FORCE the manufacturers to operate a "WiFi-Gear Update" site for all their WiFi, WiMax Wi-Whatever gear so that all of the Part 15 junk being built in Asia and sold here could be brought into compliance automatically. The FCC would never grow 2 big brass ones and tell the manufacturers to do it or else, then we wouldn't have so many interference possibilities.
    Xwindowsjunkie
    • Actually many of the major players for the ....

      ... routers do offer this in their management interfaces. Belkin for instance has an autoupdate you can set on the router. Unfortunately I do not know of the same functionality for the cards.
      ShadeTree
  • Is this a draft 2.0 or an Atheros/D-link ....

    .... implimentation problem? There are working examples of this functioning with some Draft 1.0 products. It sounds like D-Link may have rushed this out and may still have some tweaking to do. I wouldn't however assume that there is anything wrong with the Draft 2.0 specification.
    ShadeTree
    • Should be 2.0 draft

      It's suppose to be draft 2.0. I'm hoping this isn't a trend.
      georgeou
      • I still think this is a problem with the unit ....

        ... tested and not the draft. As I said, I have seen this work with Draft 1.0 products.
        ShadeTree
  • Nothing new, 11g jams itself anyways

    Bah, the 11g standard is pretty fragile itself. The channels overlap, so you effectively have only 2 channels. Also, a lot of cordless phones use those frequencies. 11g practically jams itself if you have too many people in a neighborhood using wireless.

    Frankly, it's not really worth bothering with a wireless setup. I'd rather spend some money on cables and wire everything. Plus, wired is faster anyways, especially with gigabit becoming more common.

    Leave wireless to hotspots in town for the travelers. For inside my own house, I'm sticking to wired.
    CobraA1
    • I hope you're joking . . .

      because what you are talking about is bit different than what George and everyone else is talking about. Kind of an apples and oranges thing.

      What your talking about is common to all Wi-Fi in the 2.4 GHz spectrum. A channel takes up 22 MHz and overlaps with the next 5 consecutive channels. Therefore, when you do the math you wind up with 3 (not 2) non-overlapping channels. This is not unique to 802.11g because 802.11b is the exact same way. Vendors set most AP and wireless routers usually to channel 6 and most consumers not knowing any better leave it there.

      What we're talking about is a specific feature in the 802.11n standard, called Multiple-input multiple-output (MIMO), which takes up 44 MHz or two channels instead of just one to increase both range and throughput. The downside is that older equipment has less space to work with; only two non-overlapping channels in the 2.4 GHz frequency.

      If Draft 2.0 works correctly there will be no problem with jamming. Assuming that the person who sets up the network knows what their doing, of coruse.
      n.stockwell@...
      • No, I hope you are joking...

        because MIMO has nothing to do with using wide channels. You can use MIMO on regular channel too, and it's still faster than 11g. No, the wide channel feature is called channel bonding.
        GW Mahoney
        • If we agree on when channel bonding is . . .

          different than MIMO. There logically the same in so far that two radios can be used for one connection, but the similarities end there.

          MIMO is a radio technology, channel bonding is a computer networking technology. When MIMO is used two radios are being used which means that more of the 2.4 GHz spectrum is being used.

          The speed increase when MIMO's disabled is due to using a different encoding scheme called Alamouti coding. Since you aren't using the other radio, but keep on using the same encoding scheme, your throughput goes down by half.
          n.stockwell@...
        • Your right . . .

          I'm sorry for confusing terms. MIMO is

          "the use of more than one antenna to send and receive two or more unique data streams over the same channel simultaneously in wireless devices, resulting in networks with long ranges and high throughputs."
          n.stockwell@...
    • No, this jams up to 3 channels

      When something takes 40 MHz (44+ MHz with leakage) and it's centered on channel 6, it kills everything from channel 2 to channel 10. That has a huge negative impact on channel 1 and channel 11 because of a huge overlap. Normal single-channel 20 MHz 11g only takes up 1 channel at a time and it has more dead time within that channel. 11n MIMO has far less dead time within the channel so it murders its neighbors within that channel and it can effectively take up to all three channels if it's centered on channel 6. If it?s centered on channel 3 or channel 9 that?s only slightly better but it?s taking 2 of 3 channels.
      georgeou
  • You like Intel?

    Aren't they a charter member of the EWC? The group that pushed for including wide channel support in the 802.11n standard? I suppose it served their purposes at the time. For more hipocrisy watch them be very quiet about disabling this feature to customers. They'll advertise the 5GHz 11n rates only.

    By the way, there's another company that has decided not to use 40MHz channels on thier routers and laptops in the 11g band, but it's not a popular company on this blog, so it shall go unnamed ;-)
    GW Mahoney
  • D-Link now says they have first Alliance Draft 2.0 cert device

    The router they list is the DIR-655 which was the model that Tim Higgins had problems with backwards compatibility. However, they do not say in the press release if new firmware is needed for existing models or if current owners need to replace the DIR-655.

    Here's the press release:
    http://www.prnewswire.com/cgi-bin/stories.pl?ACCT=109&STORY=/www/story/06-29-2007/0004617938&EDATE=

    More confusion left to come . . .
    n.stockwell@...
    • More mud . . .

      Tim Higgins latest article explains that the DIR-655 didn't violate the Wi-Fi alliance's certification. Draft 2.0 specifies that the Out Of the Box (OOB) settings for devices in the 2.4 GHz band have the setting 20 MHz mode enabled. However, since the the IEEE is still debating the implementation details of 2.4 GHz 20 / 40 MHz channel mode the Alliance has decided not to test how devices operate in the 40 MHz channel mode.

      Thus, the manufactures can be bad neighbors and good neighbors and still receive certification. Here's the article:

      http://www.smallnetbuilder.com/component/option,com_myblog/show,the-gotcha-in-draft-11n-wi-fi-certification.html/Itemid,199/
      n.stockwell@...
      • I spoke with Tim on this

        I spoke with Tim on this and it's clear that the Wi-Fi Alliance has no teeth on enforcing good neighbor behavior.
        georgeou
  • it's not the standard, but the hardware...

    ok, maybe George has trouble writing clearly, but to make this clear:

    In this case, the problem is not with the standard, but with a particular hardware setup.

    But a casual read of the Ou's blog, makes it look like the trouble is the standard and not the hardware that was tested... gee whiz, imagine this, a hardware manufacturer has trouble implementing a newer version of standard.... nothing new there.. the only ugly part is because it fails to correctly implement a part of the standard, the equipment can jam others...

    P.S. lots of hardware out there do not correctly implement a portion of the standards they claim to follow (the worst being graphic cards under OpenGL and D3D).
    kRogue
    • The standard has no teeth to enforce good neighbor behavior

      The standard has no teeth to enforce good neighbor behavior and that's the root of the problem.
      georgeou
  • RE: 802.11n Draft 2.0 gear still a neighbor killer

    We need several freq ranges for more types of devices. Need 8x16 real channels over a 4GHz range of frequencies. That should reduce interference.
    tygrus