Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
The Internet Science Technology

Honeybees Might Prompt Faster Internet Server Technology 131

Posted by Zonk from the heard-the-buzz-about-it dept.
coondoggie writes "The Georgia Institute of Technology is working on the theory that honeybees can give us hints about how to improve the speed and efficiency of Internet servers. Honeybees somehow manage to efficiently collect a lot of nectar with limited resources and no central command. Such swarm intelligence of these amazingly organized bees can also be used to improve the efficiency of Internet servers faced with similar challenges." This has some similarities to the rules of the swarm discussion we had last week.
This discussion has been archived. No new comments can be posted.

Honeybees Might Prompt Faster Internet Server Technology More

Honeybees Might Prompt Faster Internet Server Technology

Comments Filter:

  • Oblig (Score:4, Funny)

    by crowbarsarefornerdyg ( 1021537 ) on Monday November 19, 2007 @03:22AM (#21404867)
    I for one welcome our new swarming server overlords!

  • clusters ? (Score:5, Interesting)

    by jacquesm ( 154384 ) <j@wwAUDEN.com minus poet> on Monday November 19, 2007 @03:28AM (#21404915) Homepage
    I think bees (or ants) should get the all-time patent rights to clustering a number of not so intelligent nodes into something that exhibits a higher degree of intelligence.

    It's still quite hard to come up with stuff that is not in some way already present in nature. If you are prepared to accept a certain level of metaphor.

    • Microsoft probably already tried; the text of the patent is simply 1 Microsoft Way, Redmond, WA.

    • See: MUTE (Score:5, Interesting)

      by trawg ( 308495 ) on Monday November 19, 2007 @04:19AM (#21405183) Homepage
      MUTE is a privacy-protecting p2p application: MUTE's routing mechanism is inspired by ant behavior. [sourceforge.net]

    • I think bees (or ants) should get the all-time patent rights to clustering a number of not so intelligent nodes into something that exhibits a higher degree of intelligence.
      The human brain is by far the best example of that.

       
      • yes, but bees predate homo pseudo-sapiens pseudo-sapiens, Probably by 200 million years.
        • Sorry, that should read 'at least 80 million years' (oldest bee-fossil found: Cretotrigona prisca)

      • clustering a number of not so intelligent nodes into something that exhibits a higher degree of intelligence.
        The human brain is by far the best example of that.
        So you've clustered a number of unintelligent human brains to build something that exhibits a higher degree of intelligence?

        I, for one, welcome our fungus from Yuggoth overlord.

        • So you've clustered a number of unintelligent human brains to build something that exhibits a higher degree of intelligence?
          I thought he was talking about Wikipedia.

          • So you've clustered a number of unintelligent human brains to build something that exhibits a higher degree of intelligence?
            I thought he was talking about Wikipedia.
            No, that would be using a number of human brains that exhibit a higher degree of intelligence to build something that exhibits.

            (Probably, in soviet russia)

        • Re: (Score:2)

          by mstahl ( 701501 )

          He said the human brain, which is comprised of millions and millions of neurons. Neurons themselves are pretty useless but put enough of them in one place, connected and pathed out over years and years of learning and practice, and they're a lot smarter than a hive of bees.

      • The human brain is by far the best example of that.

        Well, the human brain does what every other brain does, even bees -- a bunch of small parts build up into something more complicated.

        In this case, we're talking about discrete organisms which collectively are smarter than any of the individuals could ever be. That's gotta count for something. :-P

        Cheers

        • In this case, we're talking about discrete organisms which collectively are smarter than any of the individuals could ever be. That's gotta count for something. :-P
          Ok, yeah. However, the principle is the same. The intelligence is in the network, not the nodes.
           
    • I used an ACO algorithm in a system to direct cow corpse recovery trucks. http://en.wikipedia.org/wiki/Ant_colony_optimization [wikipedia.org]

      I wonder if the people at the The Georgia Institute of Technology (git?) has nightmares with bees running through a series of tubes as I had about giant cow-corpse-eating zombie ants.
    • "I think bees (or ants) should get the all-time patent rights to clustering a number of not so intelligent nodes into something that exhibits a higher degree of intelligence."

      Actually I think their "higher intelligence" isn't actually higher, I think it is combined by sheer "raw ability" of each individual bee to optimally find the correct path along a geometry. In my mind it's actually a function of little minds, navigating a geometric space optimally.
    • So, if I spray some pheromones onto my servers, will they work together much better?

    • Re:clusters ? (Score:5, Funny)

      by poot_rootbeer ( 188613 ) on Monday November 19, 2007 @09:57AM (#21407045)
      I think bees (or ants) should get the all-time patent rights to clustering a number of not so intelligent nodes into something that exhibits a higher degree of intelligence.

      Which is not to say that there isn't any room for improvement. There's a lot to be learned from wolves, for example, where each member of the pack serves a unique and important role.

      It's quite likely that by combining aspects of many of these ecologies, we could create a system even more efficient than any individual one.

      Imagine a Bee-Wolf cluster...

      • Imagine a Bee-Wolf cluster...


        That is the single best pun-based abuse of a /. meme that I have yet seen.

        Bravo!

    • Re: (Score:3, Interesting)

      by daem0n1x ( 748565 )
      Very much the opposite of humans who are very intelligent but, as a crowd, behave in a very stupid way.

    • Re: (Score:2)

      by HardCase ( 14757 )
      I don't know...the evidence is pretty clear that their last attempt [beemovie.com] at using the legal system to their advantage didn't work out so well.

  • Compulsory Comcast comment (Score:5, Funny)

    by BestNicksRTaken ( 582194 ) on Monday November 19, 2007 @03:35AM (#21404957)
    Its not good making a new internet protocol, Comcast will only block it!
  • Combining them is now called a virtual dance floor. Either that or I don't get this article.

  • Nanny nanny boo boo. (Score:4, Funny)

    by rice_burners_suck ( 243660 ) on Monday November 19, 2007 @03:38AM (#21404983)
    heh heh. This so-called "swarm intelligence" will do nothing to teach us how to make efficient web servers. The hive and the swarm of bees operate efficiently but not because they have some sort of innate intelligence that allows them to do so. They operate in this manner because they are programmed to do so. The actions of each bee are based on something akin to a computer program. This program is designed in such a manner that when many units are executing it in parallel, with each unit operating on its own timer, so that statistically all parts of the program are being executed simultaneously across the bees in the swarm, the result is the efficient overall operation that we witness. However the point is that the individual program is designed so that the overall program will execute efficiently, regardless of where any particular instance of the individual program might be in its program code. Who did this programming? God. And the crazy thing is that beehives are only one tiny part of it. The overall program encompasses the entire universe. So ha ha ha... cuz you can study those bees all day long and it won't make you a better web programmer.

    • Re:Nanny nanny boo boo. (Score:5, Funny)

      by stranger_to_himself ( 1132241 ) on Monday November 19, 2007 @04:34AM (#21405237) Journal

      cuz you can study those bees all day long and it won't make you a better web programmer.

      No, but you'll be a web programmer who knows a lot about bees. Think of the possibilities!

    • Re: (Score:1)

      by ch0ad ( 1127549 )

      The hive and the swarm of bees operate efficiently but not because they have some sort of innate intelligence that allows them to do so. They operate in this manner because they are programmed to do so.
      is that not a complete contradiction?
    • You of course realize, people reverse engineer software all the time, right?

    • Re:Nanny nanny boo boo. (Score:5, Funny)

      by arevos ( 659374 ) on Monday November 19, 2007 @07:16AM (#21406001) Homepage

      Who did this programming? God. And the crazy thing is that beehives are only one tiny part of it. The overall program encompasses the entire universe. So ha ha ha... cuz you can study those bees all day long and it won't make you a better web programmer.
      I'll say. This God character has put together something pretty impressive in only a week, but it's all indecipherable spaghetti code. Where are the comments? The well-named functions? The bloody documentation? We're stuck with this system, and working out what the hell is does is pretty much a full time job for millions of experts. You think you've seen bad COBOL systems? Take a look at Universe 1.0; it's got so many quirks and undocumented features that it'll make your head spin just trying to understand what the hell it's doing half the time. I mean, sure, maybe quantum superposition made sense as an optimization feature at the time, but some, any, documentation on it might help!

      • Re: (Score:1)

        by egamma ( 572162 )
        Documentation? Haven't you heard of the Bible? It may not cover HOW Universe 1.0 was created, or how it works, but then again, neither does Science. The Bible does cover the WHO (God), What (Universe 1.0), and WHY (a God of love needs somebody to love) questions. We've been given a Manual, because we're the Users, not the programmers.

        Science can't tell you WHY H20 becomes less dense as it approaches 0 degrees Celcius, when H and 0 separately do become denser as they get colder. Science can't tell you WHY 1

        • Re: (Score:2)

          by arevos ( 659374 )

          Documentation? Haven't you heard of the Bible?

          Meh, that's just the user's manual, and it's not exactly a good example of clear and precise technical writing.

          We've been given a Manual, because we're the Users, not the programmers.

          Clearly you don't subscribe to the ideals of the free software movement, sir! Besides which, have you seen the number of bugs [wikipedia.org] in the system? Shocking!

          And don't let me get started on how many errors crop up in Homo Sapiens that we regularly have to patch [wikipedia.org]. It would be a lot easier if the programmer would just provide some docs.

          Science can't even tell you WHY different numbers of protons create liquids, gases, and solids--it just says that "at X temp, Y element is in state Z."

          I suspect we'll find a comment to explain that:

          /* Horrible hack to get

      • Re: (Score:2, Insightful)

        by McD ( 209994 )

        This God character has put together something pretty impressive in only a week, but it's all indecipherable spaghetti code.

        In Lisp or Perl? [xkcd.com]

        I have a theory: As time goes on, the odds of any slashdot thread becoming an XKCD comic, or vice-versa, approaches one.

      • "I'll say. This God character has put together something pretty impressive in only a week, but it's all indecipherable spaghetti code."

        Actually, if (hypothetically) a god exists, we all know it's not the christian god. If the universe is spaghetti code, (indistinguishable, not distinct, not equal to anything) then it can't exist. Even if the code is "bad" you cannot judge if the code is bad if you are inside the simulation because you don't have access to the code.
        • we all know it's not the christian god. We do? Speak for yourself.
          • Actually yes, the Christian god makes a claim he cannot lie, if he falsifies that claim to have that attribute, then he doesn't exist. An unintelligible communication means an unintelligable definition, and anything that cannot be defined distinctly cannot exist. Therefore the existence of the "Christian god" is not intelligible.

            Either way, a god that gets his facts wrong (i.e. creation in 7 days vs billions of years), cannot be a god, since a god would KNOW that the universe is old. His godhood is condi
        • you say that we all know that a god couldn't be the Christian idea of a god. Why?

      • spaghetti code (Score:4, Funny)

        by jefu ( 53450 ) on Monday November 19, 2007 @09:57AM (#21407037) Homepage Journal

        Obviously if the universe is mostly spaghetti code, it is a clear indication that the Creator must have been somehow involved in, well, spaghetti. Like say the Flying Spaghetti Monster. Talk about Occam's Razor - there is no simpler hypothesis available. Pasta -> Pasta. QED.

    • Natural selection programmed bees.

    • Who did this programming? God. And the crazy thing is that beehives are only one tiny part of it. The overall program encompasses the entire universe. So ha ha ha... cuz you can study those bees all day long and it won't make you a better web programmer.

      It is the glory of God to conceal a matter, But the glory of kings is to search out a matter. [bible.cc] So clearly, God *intended* for us to learn from His creation, contrary to the parent (which could be either a sarcastic dig against a straw man, or a troll). In fact, one of the criticisms offered by Intelligent Design against Materialism is that science is sometimes hindered when it is assumed that some aspect of the world has no design or function, when it turns out later that there is a well hidden design/fu

  • no central command ? (Score:4, Interesting)

    by permaculture ( 567540 ) on Monday November 19, 2007 @03:41AM (#21404997) Homepage Journal
    Don't tell that to the queen.

    • Re: (Score:3, Insightful)

      by Anonymous Coward
      I know that's supposed to be funny, but lest someone take it seriously, the queen isn't supposed to (and she doesn't) command anyone or anything - she exists for reproduction purposes only. The workers seem to "serve" her because of her unique function on the colony, that is necessary for the colony continued existence. After all, the workers are just doing their part for the survival of the colony, as much as the queen is. There's no such thing as an hierarchy on a colony, everyone works for everyone.

    • Don't tell that to the queen.

      Not really. I don't think the queen actually gives specific directions about what needs to be done by the workers. The workers feed her, and help her breed, and when she's getting old and failing they look at making a new queen. The determination that they need more food, or have found a new source, or there's an incursion from another species, or what have you is really handled by the workers.

      She's not the master controller, she's just sort of the focal point of everything.

  • Bees (Score:2, Funny)

    by archeopterix ( 594938 )
    Honeybee method? Now that's a good buzzword.

    • Re: (Score:1)

      by pwilli ( 1102893 )
      We need a new buzzword for those heavy, google page rank increasing, spam link invested websites. I vote for "honeypot"

  • After all... (Score:5, Funny)

    by Burpmaster ( 598437 ) on Monday November 19, 2007 @03:50AM (#21405037)
    The Internet is basically a series of bees.

    • Re: (Score:3, Funny)

      by FinchWorld ( 845331 )
      Forced through tubes? So basically too many bee's cause the tubes to jam up? So bascally what we are trying to avoid is bees in swarms entirely? Because as far as I can tell I've avoided contact with swarms of bees due to a healthy regard for my own well being...

      • Re: (Score:2)

        by RuBLed ( 995686 )
        So now we know why site admins fear botnet attacks...

      • Forced through tubes? So basically too many bee's cause the tubes to jam up?
        Actually, I think the correct terminology is 'honey up'. Maybe the internet is a bunch of honeypots and we're all like Winnie the Pooh trying to get up at them with balloons and.. uhh.. ohh look.. it's past my bed time.

  • Almost historical concept ... (Score:3, Interesting)

    by foobsr ( 693224 ) on Monday November 19, 2007 @03:57AM (#21405073) Homepage Journal
    Quote [wikipedia.org] (Lem, The Invincible, paraphrased):
    "A powerful military space ship a "second-class cruiser" called Invincible, lands on the planet Regis III to investigate the loss of sister ship, Condor. During the investigation, the crew finds evidence of a new form of life, born through evolution of autonomous, self-replicating machines. The evolution was controlled by "robot wars", and the only form that survived were swarms of minuscule, insect-like machines. Individually, or in small groups, they are quite harmless to humans and capable of only very simple behavior. However, when bothered, they can assemble into huge swarms displaying complex behavior arising from self-organization, and are able to defeat an intruder by--what could have been called today--a powerful surge of EMI. Some members of the spacecraft crew suffered a complete memory wipe-out as consequence. The angered crew attempts to fight the enemy, but eventually recognizes the meaninglessness of their efforts in the most direct sense of the word." (emphasis mine)

    Hint for a scientific career; Revive old stuff!

    CC.

  • Not going to work outside of individual systems (Score:5, Insightful)

    by vux984 ( 928602 ) on Monday November 19, 2007 @04:03AM (#21405093)
    Honeybees, and swarm intelligence in general assumes that the other members are working towards the good of the swarm. That is the polar opposite of what we need for a robust internet.

    Rogue nodes would be able to disrupt the swarm in the same way that scientists are able to wreak havoc on hives, ants, and other 'swarms' by deliberately injecting fake disruptive markers/signals etc.

    This technology sounds about as bright as cooperative multitasking. Suitable for a closed system (e.g. a single application) but an utter disaster if applied in an environment where some threads are just defective, or worse, hostile.
    • the individuals don't necessarily need to work toward the good of the swarm. In fact they usually just act selfish, but the resulting emergent behavior is good for the whole swarm. For example, when huge locust swarms run out of food, they begin to eat each other. The effect is that each bird tries to fly away from its followers and eat the one flying before him, which results in the whole swarm moving to a new area where they can find more food.

      • Re: (Score:2)

        by vux984 ( 928602 )
        the individuals don't necessarily need to work toward the good of the swarm. In fact they usually just act selfish, but the resulting emergent behavior is good for the whole swarm.

        The point is that emergent behaviour results because the members of the swarm are all behaving a certain way.

        Imagine a botnet of locusts that DIDN'T behave like other locusts, and instead were maliciously intent on disrupting the swarm. Perhaps they'd NEVER eat, and ONLY chase other locusts around in order to get the swarm to move
    • It may be for the good of the swarm however there is still several behaviors to observe.. such as when we observe ants, generally each ant is involved with a specific "job", and they are required to do for the good of the hive. They drop pheromones along the path they choose and eventually the paths that contain the most amount of pheromones become the most desirable and taken, eventually finding the most optimal path (this happens because pheromones disappear after time, so the longer path gradually becom

      • Re: (Score:2)

        by vux984 ( 928602 )
        It may be for the good of the swarm however there is still several behaviors to observe.. such as when we observe ants, generally each ant is involved with a specific "job", and they are required to do for the good of the hive...

        Have you ever heard of ant slavery? Where one species of ants enslaves another colony to feed its own queen. They do this by copying and disrupting the enslaved ant colonies pheromone markers to 'trick' the ants into working for them.

        This would not the a good foundation to build the
    • All you need is giant Bee Overseers to police the workers, and kill off any rouge nodes. The actual implementation of the killings is left as an exercise for the reader.
    • I totally agree with this- I think Innarwebs is not the place for this theory to be tested. Believe it or not, I've had a similar idea for some time now, but I am not in academia, so it'd be tough for me to work it out. I think they could some up with something if they moved the idea elsewhere in computing.

  • Obligatory (Score:3, Funny)

    by sw155kn1f3 ( 600118 ) on Monday November 19, 2007 @04:21AM (#21405193)
    So it's a lot like beowulf cluster of bees, right?
  • That is the question

    • Surely it must be either one bee or zero bees? Good god, I've just invented Beenary!

      This will go nicely with my upcoming suite of organic IT products that already includes the bean array (which can double as an alternate fuel source, but which is not well suited to use in cubicles...)

      I have no idea where all that came from. Sorry. Please move on to the next comment.

  • This is all just a clever trick to gain publicity for Buzz [wikipedia.org] and possibly even to provide more work for the mascot during the off season!
  • Better hope the RIAA/MPAA don't hear about it, though, or in four years we'll all be dead....

    "Silly bunt."
  • Given the current problems ( mentioned on /. somewhere previously ) with bee colonies mysteriously disappearing I'm not sure its a good idea to base something as serious as web servers on their behaviour. It's all well and good whilst they behave themselves and work away as they should but what happens when they decide to mysteriously vanish ? What then ?

    I say that nature and technology do not mix and only disaster awaits for mankinds foolish attempts to dally in that which it cannot understand.

  • Company Intelligence (Score:5, Insightful)

    by zaydana ( 729943 ) on Monday November 19, 2007 @05:10AM (#21405369)

    This sounds like the opposite to today's corporate culture, where a whole lot of smart people are part of a swarm, and the end product is utter stupidity...

    "None of us is as stupid as all of us".

  • Sometimes swarm behavior is inefficient (Score:3, Interesting)

    by MichaelCrawford ( 610140 ) on Monday November 19, 2007 @05:28AM (#21405457) Homepage Journal
    Several times I've seen flocks of birds flying in circles. One time I watched this for several minutes. The birds were flying really fast but going nowhere.

    • perhaps they were herding insects or gaining altitude on thermals?

    • Re: (Score:2, Interesting)

      by TheGoodSteven ( 1178459 )
      Those birds you speak of are doing something called "updrafting". Basically, they find a spot where warm air is rising from the ground and glide around in circles in order to attain a higher altitude using much less energy. You might see this over highways quite often, since the black pavement sometimes causes warm air drafts. I think the best demonstration of inefficient swarm behavior is when it arises in humans.

    • Re: (Score:1, Funny)

      by Anonymous Coward
      This is a common scheme you'll often see played out in subways and other heavily trafficated areas.
      You'll have a group of birds (the 'flock') flying over you in circles, trying to get your attention. Meanwhile, while you're not paying attention, an accomplice on the ground (the 'duck') will swipe your wallet.
      This happened to me while visiting London, and while it sucked to be stranded in the city without any cash, it is a very good example of how so-called 'gang behaviour' can show considerable street smart

  • Honeybees somehow manage to efficiently collect a lot of nectar with limited resources and no central command.
    This sounds good also in applications other than internet and datacenters. Like politcs and economy.
    Maybe someday we'll get ruled by bees or ants.

    • Maybe someday we'll get ruled by bees or ants.

      Hmmmm ... might this be the one time where I can say "I for one welcome our new ant overlords" and actually be on topic? :-P

      Cheers
  • Could this process be accelerated by allowing the bees to shoot from the mouths of barking dogs?
  • Isn't this kind of what bittorrent does for us? Instead of 1 strained server we have one person or thousands of people that we download the information from.
  • Then again, honeybees unexpectedly disappeared/died this year in large quantities, an event that I would not like to see our servers duplicate.
  • I mean really, can you imagine how many people would be killed when you have billions of honeybees employed to carry data packets around. Also, I dont think it would be faster either. Just imagine the lag time when you connect to a server 1000 miles away. You would keep saying, "Oh man, hurry up you shitty bees!!"
  • And there's a picture of a bee!
  • Honeybees? About time really, we have had honeypots for ages, and they don't fill themselves on their own you know.

    Now all we need is to figure out some tech based on Winnie the pooh, so we can get to the honey. Mmmmm, honey.
  • ...that the school whose mascot is Buzz [wikipedia.org] would think bees might be the solution for, well, anything...

    -Proud Georgia Tech alum
  • ...that tell you why bee swarm-sourcing works and human swarm-sourcing never will:

    Bees Don't Lie.

  • My first reaction to the summary was that there's been lots written on this topic, and it's all about routing. So it doesn't apply to web servers at all, because they're not the modules doing the routing. (At least not in any system with a minimally intelligent design. ;-) Ant- and bee-inspired algorithms are useful, and have been used by routing code, of course, but why would one be talking about web servers?

    So I took the radical step of actually reading the article. And I found it remarkably incoherent

Slashdot Top Deals

news: gotcha

Close