Slashdot Mirror

← Back to Stories (view on slashdot.org)

Historic IEEE 802 Group Looks Back and Forward

Posted by samzenpus on from the state-of-things dept.

An anonymous reader writes "The IEEE MAN/LAN Standards Committee — better known as the people who brought us Ethernet, Wi-Fi, and Bluetooth — is celebrating its 30th anniversary next week. This article has interviews with the original committee chairman and other veteran members, and reveals some of the inside situation. It also looks at some of the upcoming 802.x standards including one that sends data by modulating visible light."

45 comments

  1. IEEE Bluetooth? by Anonymous Coward · · Score: 3, Insightful

    IEEE did not develop the Bluetooth standard

    1. Re:IEEE Bluetooth? by east+coast · · Score: 4, Insightful

      The IEEE didn't develop any of the mentioned technologies. They just standardized them.

      --
      Dedicated Cthulhu Cultist since 4523 BC.
  2. Upcoming? by silverdr · · Score: 2, Insightful

    Don't we use visible light in optic fibre for some time now? ;-)

    --
    Now, mod me down freely. My karma can't get any worse...
    1. Re:Upcoming? by fan777 · · Score: 3, Funny

      I think the summary is talking about lighthouses.

    2. Re:Upcoming? by zwei2stein · · Score: 1

      Indeed.

      And even in wireless data transmitions: http://ronja.twibright.com/about.php

      --
      -- Technology for the sake of technology is as pathetic as eschewing technology because it's technology.
    3. Re:Upcoming? by peragrin · · Score: 2, Funny

      true but this is for wireless light transmission. You see IEEE has been watching old sci fi series and thought they needed more blinking lights in the real future. So each one of those blinky lights is in reality an active data connection for some device.

      --
      i thought once I was found, but it was only a dream.
    4. Re:Upcoming? by silverdr · · Score: 1

      Ah - wireless... that means they probably read this article?

      --
      Now, mod me down freely. My karma can't get any worse...
  3. MAN/LAN by zlel · · Score: 2, Interesting

    This is really nostalgic I almost forgot what MAN meant!

    1. Re:MAN/LAN by Anonymous Coward · · Score: 0

      This is really nostalgic I almost forgot what MAN meant!

      hehe yes it is (posted via www.proxyglype.com)

    2. Re:MAN/LAN by FenwayFrank · · Score: 0, Offtopic

      You can always "man man", man

    3. Re:MAN/LAN by Hurricane78 · · Score: 0

      Since this is Slashdot...

      This is really nostalgic I almost forgot what WOMAN meant!

      There, fixed that for ya. ;)

      --
      Any sufficiently advanced intelligence is indistinguishable from stupidity.
    4. Re:MAN/LAN by aqk · · Score: 0

      Since this is Slashdot...

      This is really nostalgic I almost forgot what WOMAN meant!

      There, fixed that for ya. ;)

      this is Slashdot...
      I ... forgot what WOMAN meant!

      There, fixed that for ya. ;)

      --

      .
      - aqk
      F U /. illiterates.
  4. Ethernet was fine by TheLink · · Score: 1

    Ethernet was fine, but they sure screwed up with WiFi. Broken crypto etc.

    It's still messed up - you can't have easy encrypted anonymous WiFi the way you can have easy encrypted https connections.

    --
    1. Re:Ethernet was fine by TheLink · · Score: 3, Interesting

      In case people don't get it, with the current WiFi standards you cannot have an easy way for a Cafe/Hotel/Conference to provide encrypted wireless connections to guests in a way where they cannot snoop on each other's connections. if you use preshared key users can decrypt each other's traffic. If you use username and password, it's far more inconvenient for the user and the service provider.

      --
    2. Re:Ethernet was fine by Bearhouse · · Score: 2

      Agree. Wireless security was really broken from the start, and has got only slightly better, slowly...
      Interoperability between devices from different vendors is not so bad now.
      On the other hand, if you travel around the world you can actually enjoy the convenience of wirelessly connecting to the internet almost anywhere.
      (The other day I did a free conf call, via Wifi in some airport terminal, with three people in three different continents, using my bluetooth headset on a cheap laptop...could you imagine that 20 years ago?)
      So, better an imperfect solution that works, than no solution. Well done IEEE!

    3. Re:Ethernet was fine by Anonymous Coward · · Score: 0

      I don't have much experience with networking, so I'm mostly talking out my ass, but I think that you're expecting WiFi encryption to solve the wrong problem.

      WiFi's just a wireless data link layer protocol and so it should only address the problems that other data link layer protocols address. This isn't the layer where security is typically addressed.

      As I see it, WiFi's encryption is just for access control purposes as a stand-in for the physical access control that you have with wired networks. Just let it stay that way and never assume that a cleartext message has any protection.

    4. Re:Ethernet was fine by TheRaven64 · · Score: 4, Informative

      In case people don't get it, with the current wired Ethernet standards, you cannot have an easy way for a Cafe/Hotel/Conference to provide encrypted wired connections to guests in a way where they cannot snoop on each other's connections either. If you want security, you should be using end-to-end encryption.

      --
      I am TheRaven on Soylent News
    5. Re:Ethernet was fine by TheLink · · Score: 1

      Users are responsible for their security. But as a provider, you may wish to do your best for them within your ability to provide a decent amount of privacy and security.

      You can do that with wired ethernet. You can't with WiFi.

      It is far easier to sniff someone else's WiFi traffic than it is to sniff someone else's traffic from an ethernet port.

      AND it is still much easier to make it even harder for guest users to snoop on other wired ethernet connections with various switch vendors (e.g. cisco, huawei) port security features. The use of these features do not require the guests or users to do anything except just plug in as normal.

      WiFi on the other hand is crap. It's either easy but unencrypted and completely insecure, or difficult for the users- they have to go get some credentials from the provider (cafe, conference organizer, hotel etc).

      And if the credentials are a preshared key (WPA Personal or WPA2 Personal), it's not really secure against the other users. Because WPA2 Personal is broken in that if the PSK is known, you can decrypt other people's stuff as long as you can see the 4 way handshake, which is sniffable.

      http://www.wi-fiplanet.com/tutorials/article.php/3667586

      So use WPA2 Enterprise? That involves users and passwords. How do you get those to the guest user's _machine_ (remember you can get it to the guest user, but they often fail at getting it to their machine ;) ).

      Now if Microsoft, OSX, Desktop Linux, provide an easy option to use "anonymous:anonymous" in this scenario and somehow verify the base station's credentials, then things will be much better.

      Lastly, regarding your suggestion "you should use end-to-end encryption". I have this to say: you should visit the real world more often and actually pay better attention. Only recently gmail started using encryption for everything. As for google searches, if I try https I get redirected to http.

      You ever do google searches when connected on a WiFi connection? As I pointed out only the "user must jump through many hoops" wifi is secure, the best you get at most places is WPA2 Personal which is not secure against attackers, so they can _easily_ pwn you when you do a google search.

      To pwn you when you're on a wired ethernet connection would involve them having to do more work - like social engineering, physically breaking stuff. They can't just sip a cafe latte nearby or even just plonk a "box" somewhere and walk away.

      --
    6. Re:Ethernet was fine by TheLink · · Score: 1

      See: http://slashdot.org/comments.pl?sid=1578784&cid=31437480

      WEP despite being "Wired Equivalent Privacy" was nowhere near that. And if you check, WPA2 Personal is nowhere near that either when it comes to guest usage (which I claim makes up a significant proportion of WiFi usage) .

      The fact that an attacker can be safely somewhere else makes it a whole different ballgame, if the IEEE bunch didn't realize that then that is just more evidence that they screwed up.

      --
  5. So let me just get this straight by Tomsk70 · · Score: 1, Interesting

    They 'standardised' the following -

    Ethernet (which you still have to set to 1000/Full because Auto-negotiate doesn't work properly)
    Wi-Fi (how many years has it taken for N to become standard? I've been through three pre-N routers....)
    Bluetooth (which is infamous for not working between devices by different manufacturers, to the point that no-one bothers with it. Oh and you get spammed).

    After decades of having to deal with this nonsense, yes - I'd have a few questions for them. Right after setting them on fire.

    1. Re:So let me just get this straight by Anonymous Coward · · Score: 1, Insightful

      Just a little pompous? The miracle is that equipment from hundreds of different suppliers can inter-operate at all. As imperfect as it is, I have a lot of admiration for the IEEE standardisation process, for what it has achieved. Those who work with IEEE standards (distinct from those who use the final product) will grumble about shortcomings, but still get excited when you can plug "A" into "B" and it works.

    2. Re:So let me just get this straight by aXis100 · · Score: 1

      Man, how ungratefull can you get. Why dont you go out and develop you own "standard" wireless protocol and see how long it takes you!

      The only hardware I've ever had an auto-negotiate issue with is Cisco switches, on many occasions with completely different clients over many years. Everyone else seems to play nice, but Cisco was well known for implementing their own "standard" early.

    3. Re:So let me just get this straight by fuzzyfuzzyfungus · · Score: 1

      Dell put out some switches that really didn't autonegotiate well(not even with the onboard NICs of Dell computers, just for giggles). Their switches always sucked, though, so that wasn't a giant surprise.

    4. Re:So let me just get this straight by Tomsk70 · · Score: 3, Informative

      Man, how ungratefull can you get. Why dont you go out and develop you own "standard" wireless protocol and see how long it takes you!

      Err...no, that's what they were supposed to be doing. Or do you think an eight-year lead time is acceptable? And your answer is stupid anyway - you don't say 'well write your own OS' when someone complains about Windows.

      The only hardware I've ever had an auto-negotiate issue with is Cisco switches, on many occasions with completely different clients over many years. Everyone else seems to play nice, but Cisco was well known for implementing their own "standard" early.

      Which tells me exactly how much networking hardware you've actually worked with, so let me fill you in - ISCSI not working? Set all adaptors to 1000/Full. Backup Exec not working? Set all adaptors to 1000/Full. Network generally slow? Set all adaptors to 1000/Full. I could list around 20 more off the top of my head...and then there's stuff such as - Vista network auto-tuning buggering up your system? That's because there's *no standard*.

      So yes, after wasting my time for *years* with non-compatible Wireless, Bluetooth and Ethernet 'standards', I think I've earnt the right to be ungrateful, thank you very much :-)

    5. Re:So let me just get this straight by BubbaDave · · Score: 1

      Some, I believe, were made by Asante.

      Dave

    6. Re:So let me just get this straight by BubbaDave · · Score: 1

      They 'standardised' the following -

      Ethernet (which you still have to set to 1000/Full because Auto-negotiate doesn't work properly)
      Wi-Fi (how many years has it taken for N to become standard? I've been through three pre-N routers....)
      Bluetooth (which is infamous for not working between devices by different manufacturers, to the point that no-one bothers with it. Oh and you get spammed).

      After decades of having to deal with this nonsense, yes - I'd have a few questions for them. Right after setting them on fire.

      Heh, I was in the room at an IEEE 802 meeting when someone actually apologized for having insisted gig-e even have a half-duplex mode.

      Dave

    7. Re:So let me just get this straight by OldeTimeGeek · · Score: 1
      The IEEE is a standards body, not an enforcement one. If a manufacturer wishes to deviate from the standard, there's nothing that the IEEE can do about it. In the time I've been doing this I've learned this about vendor equipment: There's a standard and then there's the way that a vendor chooses to implement the standard.

      Every manufacturer I've worked with has added "features" that make them not-completely-compatible with equipment from other manufactuers. Sure, they'll work, sometimes completely, but not consistently. And your autonegotiation example is an excellent example - for us, it has worked about half the time and has caused more problems than it's worth - so we don't use it any more. We also do rigorous testing to make sure that new equipment plays well with the existing equipment before we place it into production.

    8. Re:So let me just get this straight by tlhIngan · · Score: 1

      Wi-Fi (how many years has it taken for N to become standard? I've been through three pre-N routers....)
      Bluetooth (which is infamous for not working between devices by different manufacturers, to the point that no-one bothers with it. Oh and you get spammed).

      Well, it's your fault for buying pre-N equipment in the N case. After all, they couldn't get consensus on how N should work, which is why it took so long. Every manufacturer implements their own idea on how to do it, especially if there was no standard. It's just like the pre-v.90 days of modems - everyone had their own idea how to do 56k. With no standard, you end up with that. Now, N took so long (and you got burned) because there were different ways to accomplish the same thing, and there were arguments on which method should be chosen (primarily politics - after all, what method gets chosen gets patent revenue). In the meantime, you decided to go ahead and buy whatever nonstandard equipment was available because it was, well available. Manufacturers are always doing that - releasing unstandardized stuff in order to carve a niche in the market, or how technology progresses.

      Bluetooth, let's not go there. It's a horrible standard that's overly complex. And this is the core spec. For the most part, this part does interoperate quite well. It's the high-level stuff that doesn't and leads to all sorts of wierd interop problems. Profiles, especially (HID, SSDP, A2DP, Headset, FTP, OBEX, Dial-up, LAN, etc), and the incompatibility happens when various stacks implement different profiles and different subsets.

    9. Re:So let me just get this straight by RzUpAnmsCwrds · · Score: 1

      Which tells me exactly how much networking hardware you've actually worked with, so let me fill you in - ISCSI not working? Set all adaptors to 1000/Full. Backup Exec not working? Set all adaptors to 1000/Full. Network generally slow? Set all adaptors to 1000/Full.

      Let me fill you in - if auto-negotiation is failing, there's something wrong with your hardware or cabling. Forcing your adapters to a specific setting just makes the problem less visible - you still have the same shitty defective adapters/switches or bad cabling.

      It's amazing that millions of devices - most of which are based on cheapshit hardware - can interoperate at all.

    10. Re:So let me just get this straight by Tomsk70 · · Score: 1

      And you learnt that where?

      You only have to google to see that the 1000/Full thing (and the Windows autotuning problem) is down to there being *no standards*. It can be caused by different firmware, mixed hardware, different NIC's, the list is quite long.

  6. Also coming up this year... by dtmos · · Score: 3, Informative

    ... is the 20th Anniversary of the 802.11 Working Group itself. The Working Group held its first meeting September 10-14, 1990, in Oshawa, Ontario, Canada.

    1. Re:Also coming up this year... by Mikkeles · · Score: 2, Funny

      So, in this case, one can actually Blame Canada, eh?

      --
      Great minds think alike; fools seldom differ.
  7. Maybe they can change IEEE's pro-swpat stance by H4x0r+Jim+Duggan · · Score: 2, Interesting

    In the Bilski case, IEEE filed a brief pushing *for* software patents. Maybe specific groups in IEEE, like the 802 group, should push for a change in this position. Having the whole wifi industry paying a tax to CSIRO for a wifi patent must make this group a little more clued in about the harm caused.

    --
    Please help publicise swpat.org - the software patents wiki
  8. IEEE802 did not bring us WIFI by Anonymous Coward · · Score: 0

    The CSIRO brought us WIFI. Get your facts straight.

  9. Get the configuration right by thoughtspace · · Score: 4, Interesting

    I am staggered how complicated it is to setup WiFi a lay-person. Far too much jargon (SSID, WPA, WPA2, WEP, TKIP, AES+TKIP, channels ...), and stupid ideas like multiple WEP keys. Let alone connecting via ethernet, change the subnet, browse to an IP address, etc etc etc just to get it going. What an awful decade of design.

    Look ... from day 1 we just wanted a secret password.

    Public networks are different and need to be publicly identified - don't shoe-horn it into the same user interface.

    Start thinking like a user and stop this engineers crap.

  10. Wonderful... by elFarto+the+2nd · · Score: 2, Insightful

    ...prehaps they could get around to increasing MTU from 1500.

  11. Ah, the ol' 802.3 gang... by BubbaDave · · Score: 3, Insightful

    Cool. I've worked with Paul Nikolich (when ADC broadband bought bought the CMTS company he was at), and have run into some of this cast of characters during the 802.3ah Ethernet in the First Mile meetings. Interesting folks.

    I think it was Geoff (I could be wrong, this was a while ago) that said we would not need high-speed uplink from the home because 'there just isn't that much relevant content out there'. That was a pretty good chuckler.

    I'm sure Michael Coden of Codenoll feels snubbed, he always claimed to me he was the co-inventor of ethernet.Never believed him.
    He did pioneer one interesting product- a distributed ethernet switch that would operate over a unidirectional fiber ring- worked pretty well after I fixed the gaping hole in his protocol.

    Dave

    1. Re:Ah, the ol' 802.3 gang... by aqk · · Score: 0

      Fuck Ethernet.

      Token-Ring is gonna make it obsolete pretty soon. You'd have to be a moron to buy that Ethernet crap!

      - - - told to me (and my boss) by an IBM sales-rep a quarter-century ago.
      (IIRC, as an IBMer, he probably didn't use such impolite expletives)
      Luckily neither of us paid much attention to him- we continued laying our thin-net and cat5.
      But there was much FUD directed at senior management and non-technical analysts by IBM.
      And believe me, there was a shitload of pointy-haired bosses in that place! Probably still is...

      --

      .
      - aqk
      F U /. illiterates.
  12. Small correction by dtmos · · Score: 2, Insightful

    The official name of 802 is the IEEE 802 "LAN/MAN Standards Committee," not the other way around.

  13. jumbo frames by Anonymous Coward · · Score: 0

    ...prehaps they could get around to increasing MTU from 1500.

    They did. They're called jumbo frames and go up to ~9000 bytes.

    At $WORK we use them on our "storage VLANs" for things like NFS and iSCSI.

    1. Re:jumbo frames by elFarto+the+2nd · · Score: 1

      Judging from this, it should be more like 96kB, not ~9kB.

  14. 802.1 AVB is awesome by TheSync · · Score: 1

    The new 802.1 "AVB" standards (IEEE 802.1AS:Timing and Synchronization for Time-Sensitive Applications, IEEE 802.1Qat: Stream Reservation Protocol (SRP), and IEEE 802.1Qav: Forwarding and Queuing for Time-Sensitive Streams) are awesome.

    This finally will allow for the reliable transmission of high-bandwidth data streams (such as uncompressed HD video at 1.5 Gbps) over Ethernet switched networks.