Stories
Slash Boxes
Comments

News for nerds, stuff that matters

Slashdot Log In

Log In

Create Account  |  Retrieve Password

ARPANet Co-Founder Predicts An Internet Crisis

Posted by Zonk on Thu Oct 25, 2007 12:11 PM
from the everyone-is-looking-for-their-cut dept.
The Insultant writes "Dr. Larry Roberts, co-founder of the ARPANET and inventor of packet switching, predicts the Internet is headed for a major crisis in an article published on the Internet Evolution web site today. Internet traffic is now growing much more quickly than the rate at which router cost is decreasing, Roberts says. At current growth levels, the cost of deploying Internet capacity to handle new services like social networking, gaming, video, VOIP, and digital entertainment will double every three years, he predicts, creating an economic crisis. Of course, Roberts has an agenda. He's now CEO of Anagran Inc., which makes a technology called flow-based routing that, Roberts claims, will solve all of the world's routing problems in one go."
+ -
story

Related Stories

This discussion has been archived. No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More
Loading... please wait.
  • by jollyreaper (513215) on Thursday October 25 2007, @12:15PM (#21115963)

    Of course, Roberts has an agenda. He's now CEO of Anagran Inc., which makes a technology called flow-based routing that, Roberts claims, will solve all of the world's routing problems in one go."
    So why is this making the front page again? Attention, ladies! My seed cures diseases. Can we get that on the front page, too? My agenda is no less shallow than his.
  • News Just In (Score:5, Insightful)

    by JamesRose (1062530) on Thursday October 25 2007, @12:15PM (#21115965)
    Man with Solution says the is a Problem.

    Yeah, not buyin it. A similar thing happened with electricity, when everyone bought TVs everyone bought computers etc. suddenly of course power usage sky rocketed, and lots of people said, well this is going to be the rate of growth now. Of course, with that, as it is with this, everyone go their TVs and then the demand levelled out, with this, everyone will start downloading videos, and the bandwith usage will level out. Yes, soon we'll need some new routers, but the problem isn't permanent, and it isn't something that we should trust a salesman to deal with.
    • I agree. The people who say this sort of thing are the people that were predicting we would run out of oil by 2020, etc,etc. Sure bandwidth is going to become more expensive at the backbone level over the next few years, but that will simply mean better returns for those who invest in it, and the problem will self correct. Economics interprets rising prices as damage and routes around it.
        • Re: (Score:2, Insightful)

          No, we aren't. Oil's just going to get more expensive until alternative technologies are economical, and the research is poured into them to make them usable. That's all there is to it. We will never really "run out" of oil, we'll just supplant a more expensive form of energy with a cheaper one.
          • Hey! (Score:5, Interesting)

            by SatanicPuppy (611928) * <Satanicpuppy@gmai3.14159l.com minus pi> on Thursday October 25 2007, @01:55PM (#21117469) Journal
            You and your knowledge of "economics" can go! We're predicting disaster here!

            People have been predicting that we'd run out of item X by time Y for hundreds of years. The reason we don't is because (as you said) when supply dwindles, there is incentive to find news supplies and substitutes.

            During WWII, it was thought that we'd completely run out of rubber, and this would kill our war effort, due to lack of tires, hoses, gaskets, etc. Along comes synthetic rubber, and magically we don't run out. These days most rubber is synthetic.

            This stuff happens all the time. When oil becomes expensive enough, alternative fuel use will become so desirable that an efficient solution will present itself. Hell, that's why we switched to cars in the first place, because our previous transportation (horses) produced untold...uhh..."pollution". A little Co2 seemed like heaven compared to mountains of horse crap, and it didn't take long before cars needed less maintenance than horses.

            There was a time, however, when the car was a choice only the rich could afford, one less reliable and less efficient than a good horse. Economics rarely gets the solution ahead of the problem, which is why it's an uphill battle to force people to switch to alternatives when the alternatives aren't as efficient as what they're already using.

            The biggest issue right now is that the government is mucking with the damn problem by subsidizing industries to artificially make petroleum/cars seem more efficient than they actually are. For a bunch of "free market economists" they sure love to give away money to un-free the market. They're also dropping the ball by shouldering the pollution costs created by the fossil fuel industry, instead of passing it back to the industry in the form of taxes and fees. Take away the subsidies and fairly apply the costs to the industry that created them, and you'd see a much broader adoption of alternatives as the prices rose to reflect the "real" costs.

            China is a good example of this right now...They're polluting like mad, and passing the costs of that on to their citizens so that they can be super-competitive in the global market against people who have to actually pay some of those costs. It's going to catch up with them in a big way...It's like their propping up our currency. The more the dollar deflates, the more money China dumps down the drain trying to keep the dollar high, at the same time trying to keep their own money as depressed as possible.

            As the dollar deflates past a certain point, American goods will become "cheaper" and Chinese goods more expensive, leading to a local manufacturing resurgence, yadda yadda, whereas when the Chinese lose hold of their own money, they're going to have this explosion of costs internally, as well as having to watch their goods become much less competitive globally.
          • Re: (Score:3, Insightful)

            Energy Independence? Does that mean the same thing as invading the Middle East? Cos, y'know, I'm having some problem with that picture.

            No. It means becoming more efficient and drilling in places like ANWR.

            Now, stick with the topic.

            As demand for bandwidth grows, so will the supply. It's that whole supply and demand argument we learned about in Eco101.

              • Fiber cable is cheap, so when they lay 100 miles of backbone, they put plenty of fibers in the cable. Dark fiber is all over the place. However, what the article is about is router bandwidth. Adding router ports on the ends of those dark fibers is not cheap. If people were just sending text, there would be no bandwidth problem, but all those idiots linking to video and audio files is a problem.

                By the way, have you all seen the "Cat wake-up call" [youtube.com] animation on youtube?
    • I wonder (Score:2, Interesting)

      I wonder if everyone on Krypton walked around saying, "That Jor-El! Trying to sell us a rocket to escape an 'impending planetary disaster' to some backwater planet where we'll all have 'superpowers'. Yeah right. What a loser."
  • by BUL2294 (1081735) on Thursday October 25 2007, @12:16PM (#21115979)
    Robert Metcalfe predicted this in 1995 [wikipedia.org]. He literally ate his words (a printed InfoWorld article mixed with liquid in a blender) in front of an audience in 1997.
    • And John Dvorak thinks it's a good idea for AMD and Intel to merge. Sometimes, it's better to ignore asshats with stupid opinions.
      • Uh, yeah, but Dvorak doesn't have any credentials backing him up. Who is John C. Dvorak? Some stupid know-nothing tech journalist. Who is Bob Metcalfe [slashdot.org]? He's the co-inventor of Ethernet, he founded 3COM, and invented Metcalfe's Law.

        Some people at least thought he knew what he was talking about and, well, they had good reason to. I will say that I thought his comment was wrong-headed and stupid, but, then again, what do I know? I'm just some random guy on Slashdot.
        • Scarcity (Score:4, Insightful)

          by SatanicPuppy (611928) * <Satanicpuppy@gmai3.14159l.com minus pi> on Thursday October 25 2007, @02:20PM (#21117779) Journal
          People who freak out about scarcity don't understand economics. Economic pressures drive alternatives and expanded production; we've been seeing this with food since Malthus confidently predicted that food generation could never keep up with current population growth...in 1798.

          As the demand rises, people leap to fill it. When Metcalf decided we were going to run out of switching capacity, he was looking at current manufacturing capacity, and a projected increase in demand, and he was sure that capacity could never keep up with demand.

          What he didn't see is a horde of people looking for ways to make money, who were looking at the same numbers and thinking, "Holy crap! If I make switches I'll be RICH!" Demand drives supply, not the other way around.
      • Wouldn't malware cease to be much of a problem *if* we had secureable hosts?

        I mean, at some point, if the malware didn't succeed, it wouldn't be written.

        OTOH, if we had some way to enforce cleaning up hosts that spewed malware, that would get rid of it also.

        Comcast, you listening? How about filtering Storm and a few, just a few, of the well-known trojans and worms that infest the Internet? Starting with your own subscribers, residential AND commercial.

        If spam and malware are bandwidth threats, maybe THAT
  • by Seumas (6865) on Thursday October 25 2007, @12:17PM (#21115991)
    Create a problem. Solve that problem. Uniquely own that solution in the market. Make everyone need what you have to offer.

    Of course, the first step is that these guys need to really convince everyone that the internet is about to implode and that the companies who need the enormous bandwidth and services simply can't or won't make the hardware investment that is necessary.

    The real threat to the internet are the legislators and lobbyists who want to nerf the internet so that the only use for it is the commercial enterprises and everything should be nerfed down to a Disney-fied toddler's level. That's an actual legitimate threat.

    However, maybe he should peddle the "piracy and torrents are killing the internet and I can save you!" angle. Might work.
  • Wow (Score:4, Insightful)

    by EveryNickIsTaken (1054794) on Thursday October 25 2007, @12:18PM (#21116017)
    Someone submits a slashvertisement, acknowledges it in the summary, and it still gets put on the front page. Brilliant! Also, routing will be just fine. F-U-D.
  • by lbmouse (473316) on Thursday October 25 2007, @12:19PM (#21116029) Homepage
    1. Run around screaming that the sky is falling
    2. Develop and market a product that fixes the sky
    3. ?
    4. Profit!

    He must have read Chicken Little.
    • Re:Nice Formula (Score:5, Insightful)

      by Sunburnt (890890) on Thursday October 25 2007, @12:23PM (#21116121)

      1. Run around screaming that the sky is falling 2. Develop and market a product that fixes the sky 3. ? 4. Profit!

      This would make more sense if step 3 was actually a mystery. I thought step 3 was obvious: "Convince influential idiots with money that your product is the greatest and most urgently needed thing since free porn."

  • "Dark fiber"? (Score:5, Insightful)

    by dazedNconfuzed (154242) on Thursday October 25 2007, @12:20PM (#21116041)
    What happened to all that talk of "dark fiber"?
    And how much of the routing problems stem from backbone ISPs (Comcast, Verizon, etc.; see recent /.) wanting to fiddle with packets instead of simply routing them?
  • by Anonymous Coward on Thursday October 25 2007, @12:20PM (#21116047)
    When the ice caps melt, the tubes will get clogged with dead polar bears.
  • See "On Distributed Communications", published in 1964.
    • Re: (Score:2, Informative)

      A quick google search on "packet switching" reveals several people involved in the development [nytimes.com] of packet [cnn.com] switching [ciol.com], and Larry Roberts is not one of them. He, in fact, supports Leonard Kleinrock [slashdot.org]. That last article [ciol.com] on packet switching may actually be one of the more interesting ones, as it is written by someone that was involved in yet another application external to ARPANet.
  • by 140Mandak262Jamuna (970587) on Thursday October 25 2007, @12:21PM (#21116091) Journal
    If the cost increases, they will invest the money and upgrade the network. What is the problem? When MSFT thinks Facebook is worth 15 billion dollars, routers are chump change for them. What is the crisis here? Cost of something is going to go up? Big deal. Oil prices are shooting up. College tuition costs are shooting up. Y ! routing costs?
    • Re: (Score:3, Insightful)

      If the cost increases, they will invest the money and upgrade the network.

      Not quite. They will only invest in infrastructure if they think the return on that investment will at a bare minimum keep the same level of profit, and likely only if it will increase their profit.

      Companies don't increase their capacity because cost goes up, they increase capacity because by doing so they can increase or maintain profits.

      The notion that increased revenue increases capacity only works when the markets are free enou

  • I Doubt It (Score:3, Insightful)

    by penguinboy (35085) on Thursday October 25 2007, @12:22PM (#21116101)
    People have been claiming "new technology $foo is going to overwhelm the Internet!" for ages. Yet somehow the Internet keeps up. I'm not worried - especially since this guy just so happens to be offering to sell us a solution.
    • People have been claiming "new technology $foo is going to overwhelm the Internet!" for ages. Yet somehow the Internet keeps up.

      Agreed. Yet it's important to keep in mind that part of "the Internet keeping up" is that the users modify their usage according to what technology allows. Now that it is possible to download video relatively quickly, people are doing it. But trying to stream high-def wouldn't work (either you'd have to wait a really long time to buffer or the video would stutter), so people basic

  • But can we unpublish this article? I mean it doesn't really belong here. The Internet is not going to be overwhelmed by video, VOIP or anything else. It also will not cause a problem with the economy that more data needs to move around. In fact it will HELP the economy.
  • If it's too expensive to deploy the services, then perhaps people will do without the services?

    The traffic will only increase dramatically if people continue to use the services that demand the traffic, and pay for the bandwidth they need to do it.
  • I think this is a dup. This is the virtual circuit guy again, isn't it?

  • If we don't get this problem under control, it could mean the END OF THE WORLD!!!!!!

    uh. Of Warcraft. [wikipedia.org]
  • The picture on this page [anagran.com] says it all.

    This is not at all about circuit switching, or routing more efficiently. This about tracking connections through the router so that they can apply policy based on a simple lookup, rather than examining each packet. If they didn't intend to muck with the packets, a "dumb" router is perfectly fine.
  • by hackus (159037) on Thursday October 25 2007, @12:51PM (#21116543) Homepage
    Leave my internet protocols ALONE thank you very much.

    We will do quite OK without you meddling with our open standards.

    We only need linux, an open TCP stack, and anything that happens I am sure we can handle it with JUST those tools.

    Well, that and an army of a million penguin volunteers.

    We will do fine, really.

    Please peddle your proprietary CRAP OLA somewhere else.

    Thank you.

    -Hack
  • He's now CEO of Anagran Inc., which makes a technology called flow-based routing that, Roberts claims, will solve all of the world's routing problems in one go.

    (rolls eyes)

    What did Anagran pay Slashdot for this posting?

    An anagram (barely) of Anagran is "A nag ran"
  • Not again!

    The imminent death of the internet has been predicted too many times now, but it hasn't happened yet.

    The real killer will be the one we don't see.

    And anyway - the bandwidth limit will always limit the services available at any time. If a service uses too much noone will use it.

  • I can assure you that where there is demand, there will be businesses rushing to provide the supply. Its appalling to see people complaining about excessive demand.
  • Zonk you fucking moron. You already posted this earlier this month right here [slashdot.org]. Different website, but same guy and same company, of course. Same message, same bullshit!

    You have officially crossed into the JonKatz zone. Not only do you post duplicates, but you post slanted slashvertisment duplicates! Your articles are worthless.

    It's too bad all I can do is ignore you, but it's about time I finally did. I recommend everyone else do the same, so we can finally hit home that bullshit editors will not be tolerated.
    • Re: (Score:2, Interesting)

      There's no explicit circuit setup or teardown: the flows are detected by the router rather than being established by the endpoints.
        • Re: (Score:3, Insightful)

          From the looks of Anagran's description of their 'flow' based routing, routers analyze individual TCP flows very closely and rate them according to behavior in order to predict future needs of each flow, and then make adjustments on the fly by various means (closely timed discard of TCP segments to force endpoint TCP adjustments, QOS-like throttling of flows that look like they may be coming from slow endpoints, etc.)

          All of this looks to be enhancements and accelerations to QOS. It could be really cool,
    • What's the difference between this flow routing and circuit switching?

      Flow-based routing attempts to identify flows of packets - TCP connections, related streams of UDP packets, etc. - and cache information about them. Then when future packets of the flow arrive and are successfully identified they can be handled using the cached information, rather than performing a full lookup of routing, QoS labeling, permission checking, etc.

      It may also attempt to identify more things about it - such as what kind of tr