Slashdot Mirror


Fast Wi-Fi's Slow Road To Standardization

CWmike contributes this excerpt from Computerworld: "For a technology that's all about being fast, 802.11n Wi-Fi sure took its sweet time to become a standard, writes Steven J. Vaughan Nichols. In fact, until September 2009, it wasn't, officially, even a standard. But that didn't stop vendors from implementing it for several years beforehand, causing confusion and upset when networking gear that used draft standards from different suppliers wouldn't always work at the fastest possible speed when connected. It wasn't supposed to be that way. But, for years, the Wi-Fi hardware big dogs fought over the 802.11n protocol like it was a chew toy. The result: it took five drama-packed years for the standard to come to fruition. The delay was never over the technology. In fact, the technical tricks that give 802.11n its steady connection speeds of 100Mbps to 140Mbps have been well-known for years."

7 of 140 comments (clear)

  1. Not the first time by causality · · Score: 4, Insightful

    Betamax vs. VHS, HD-DVD vs. Blu-Ray, now Wi-Fi draft N versus finalized standard draft N.

    Open standards are a good thing. They avoid these kinds of problems. They promote interoperability. They also force vendors to compete on the merits of their implementations of those standards instead of competing on the basis of who is better at customer lock-in. It also lessens but does not remove the competition of who is the best at marketing.

    If you care about assigning blame, it lies squarely on the people who purchased draft-N hardware. Whether they realized it or not, they were using their wallets to vote for this behavior. Those purchasing decisions reward this kind of behavior and make it profitable. Give companies the choice of agreeing on a standard or making no sales and they will agree on a standard every time.

    --
    It is a miracle that curiosity survives formal education. - Einstein
    1. Re:Not the first time by daveime · · Score: 5, Insightful

      So I guess you'll not be using any of that "not-yet-finalized" html5 stuff, or any beta software from Google ?

      After all, no one should invent anything until it's been discussed in committee for a minimum of 10 years, until the technology it is attempting to standardize has already been superseded by something better !

      Thank [deity-of-your-choice] they didnt invent the wheel using open standards. It probably would have had 6 sides, none of which are equal in length, a 100 page operating manual, a concession to Pantone that it should only be made in RGB color 255,147,97, and an alternative implementation involving Microsoft's .innerHTML

      Anything that takes longer to describe than it does to make is probably better not describing. Just use the bloody thing and be done with it.

    2. Re:Not the first time by causality · · Score: 5, Insightful

      So I guess you'll not be using any of that "not-yet-finalized" html5 stuff, or any beta software from Google?

      Terrible examples. I really don't think you appreciate the difference between open standards and proprietary "standards". That, or you understand it perfectly well but find it inconvenient for your argument, PR-style.

      HTML5 is intended to be an open standard, so in this case you're making my point for me. There was a draft standard of HTML5 released January 2008. This too was produced openly. A vendor who produces something based on this draft standard is using the same specifications that are available to all other vendors. The same will be the case with the finalized standard.

      That has not been the case with the proprietary draft-N implementations. Each vendor has their own version of draft-N. It's very similar to Microsoft's practice of embrace-and-extend. Interoperability with another vendor's implementation is not guaranteed. If you can't get Vendor X's equipment to operate with Vendor Y's equipment, or suffer reduced performance, neither vendor will file that as a bug and fix it. Instead, both will tell you "we recommend you use our products for all your networking needs". You think this is just like HTML5, that you're really comparing an apple to an apple here?

      Most of the beta software that Google has released for download has been open source (Chromium, for example). Open source is no good if you want to implement a proprietary standard. It's great when you want the world to see precisely how something was done so they can interoperate with your software or port it to other platforms. Google obviously understands the value of this. That again serves to reinforce my point.

      This is just another example of a phony debate tactic. If there's not a term for this, there should be. The procedure goes like this:

      1. Ignore any points that the other person made. This is important. If anything the other guy said contradicts your position, just pretend that you didn't notice. Best foot forward, even at the expense of intellectual honesty. Besides, this way you don't have to waste your time with refutation and can get right down to expressing your predetermined conclusion.
      2. Proceed to find anything the other person said that is generally true, and does apply for the specific examples that person gave. Then take the general truth to an absurd extreme.
      3. Pretend like this says something about the validity of the general truth. Whatever you do, don't acknowledge that it says anything about your ability to interpret the general truth within a reasonable perspective.
      4. Declare that the general truth is inherently absurd. State outright or imply strongly that it must be false in all cases. It was false when you took it to an absurd extreme well beyond its intended scope, so it must be totally useless in all cases. Right?
      5. Congratulate yourself for your ability to handle argumentation. For extra points, assume that the other guy was a total idiot, that your trivial objections never occurred to him, and that the existence of such trivial objections could not possibly have indicated that you missed his point.
      --
      It is a miracle that curiosity survives formal education. - Einstein
  2. Re:Drama...? by RichMan · · Score: 5, Interesting

    I went through two small companies working on pre-standard N devices. Both went under as a little company you can't pre-run a standard to market. We were ready for production 7 years ago.

    So yes, drama that personally affected me as I went through two collapsing companies.

  3. Re:Drama...? by Monkeedude1212 · · Score: 4, Funny

    I hear HBO has hired someone to write a season or two based on the whole ordeal. Rumours are it will be called "N" and the tagline will be "Wi the Fi is this taking so long?". It's not above any of the normal problems that HBO shows have. You know, the kind where there is a secret love plot between two characters that have no influence on the story whatsoever. Or the writers write in a love scene, and then it gets cut short for commercial breaks.

  4. History Repeats Itself by organgtool · · Score: 4, Interesting

    The same thing happened with 802.11g. I remember going through four 802.11g PCI cards before finding one that could communicate with my 802.11g router at a distance of more than three feet. I was not aware that the devices were pre-draft (they didn't state "pre-draft" on the packaging like they do now), so I did not realize that was causing my problem. Eventually the standard was ratified, and if my memory is correct, the manufacturers released firmware updates so that the devices complied with the ratified standard. I doubt that this practice will go away since the manufacturers want to release bleeding-edge technology to stay ahead of the competition, but at least their packaging now states "pre-draft" so that cautious consumers will know to avoid it.

  5. Re:Blueray of Wifi by jasonwc · · Score: 4, Informative

    I completely agree with the parent. Wireless is often more trouble than it's worth. It's great for casual internet access on laptops, but it's really unreliable for HD video streaming, system backups, or large file transfers. If you live in a single-family home, you probably don't have many interference issues to deal with, but in a multi-unit apartment building, there is often significant interference on the 2.4 Ghz spectrum not merely from other Wireless routers but from phones, microwaves, baby monitors and other devices. The end result is often dropouts, unreliable connectivity, and slow speeds.

    Wired Ethernet is a reliable and mature technology that pretty much always works. While the vast majority of wireless routers are crap, it's quite easy to purchase a quality Gigabit Switch, and pretty much any wireless router will provide stable wired connections. People seem to ignore the fact that wireless only provides shared speeds while wired ethernet provides dedicated bidirectional bandwidth per port. In addition, the stated maximums for wired ethernet provide a sense of real world speeds while wireless does not.

    For example, 802.11g provides real-world speeds of 20-25 Mbit/sec - not the 54 Mbit theoretical speed pasted on the box. 802.11n is advertised as 300 Mbit/sec but generally provides 100-130 Mbit/sec at best. However, these speeds can only be obtained with a line-of-sight connection at a short distance - a distance so short that you could easily connect over wired ethernet and obtain 10x the speed! At 100-BaseTX I am able to obtain 90 Mbit/sec after network (94 Mbit/sec with Jumbo frames). With Jumbo Frames Gigabit can achieve 950 Mbit/sec or higher speeds.

    So, 802.11n in a best-case scenario (little interference, 10 ft from the AP) provides speeds only slightly in excess of 100-BaseTX, a standard formed 14 years ago (1995)! If you actually connect from any significant distance, 100-BaseTX will provide better speeds. In addition, you get dedicated upstream and downstream per-port bandwdith on 100-BaseTX.

    Gigabit switches with Jumbo frame support can now be had for less than $30 - and in some cases $20. Nearly all laptops and desktops now come with Gigabit NICs and support Jumbo frames. Modern dual and quad-core CPUs can easily take the overhead of transferring at gigabit speeds with or without jumbo frames. Furthermore, modern OS's are more efficient at high-speed network transfers.

    Yet, I see 802.11n routers advertised for streaming HD video, system backups, and large file transfers. Would you really want to backup 100 GB over 802.11n? Sounds like fun watching your entire network come to a crawl for 3 hours. Gigabit ethernet over Cat 5e/6 is generally limited only by your hard drive setup, and can be used for all the aforementioned tasks without any impact on network performance.

    The following is my recent experience with wireless networking:

    I use a WRT54GL with the Tomato (Linux) firmware. This provides a rock-solid solution, yet I was getting dropouts on wireless at my current apartment. I know it's not the router as it worked fine in my cinderblock college dorm, and I have another WRT54GL running Tomato at my home in NJ - which usually is up for months at a time - until a power outage.

    At my latest apartment, I was getting constant dropouts on wireless. I ran a second AP to get a stronger connection. The connection was indeed stronger but I still got intermingle dropouts. I tried changing the wireless channel, antenna placement etc. but nothing worked. Finally, I just said fuck it, bought a 100' Cat 5e cable for $8 and ran it directly from the ADSL router to the router in my room. No dropouts since then - and I get more consistent and faster speeds.

    I purchased an 8 port Gigabit switch with Jumbo Frames for my internal network and now my network speeds are limited only by my hard drives. Transferring between a 1 TB eSATA and 1 TB SATA drive, I was able to transfer a 12 GB file in 2:15 at an average speed of 95 MB/sec - around 800 Mbit/sec. With