Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
×
The Internet

Growing Commercialization Threatens Net Security 199

Posted by timothy from the fewer-windows-to-jimmy dept.
dr3vil writes "The BBC is reporting that the concentration of the net's backbone in fewer hands has made it more vulnerable to attack. The report compares an attack to travel problems when traffic is disrupted at O'Hare. Hopefully someone in a position to act will pay attention."
This discussion has been archived. No new comments can be posted.

Growing Commercialization Threatens Net Security More

Growing Commercialization Threatens Net Security

Comments Filter:

  • In Soviet Russia ... (Score:5, Insightful)

    by dzym ( 544085 ) on Tuesday November 26, 2002 @06:43PM (#4763498) Homepage Journal
    Growing commercialization is threatened by Net Security.

    Surely you mean increased centralization, however.

    • Re:In Soviet Russia ... (Score:4, Insightful)

      by flewp ( 458359 ) on Tuesday November 26, 2002 @06:47PM (#4763541)
      Anytime you concentrate anything in one area, the risks are likely to be larger. If I put my money in various hiding spots, it'd be a lot safer than hiding it all under my mattress. Sure, someone may find one or two of the stash spots, but it outweighs the risk of losing it all if someone discovers it under my mattress. Okay, that analogy might have been a stretch, but I think it gets to the point of the article. I think it's only news because someone ran some tests.

      • Eggs in one basket... (Score:5, Insightful)

        by ryants ( 310088 ) on Tuesday November 26, 2002 @06:53PM (#4763590)
        Behold, the fool saith, `Put not all thine eggs in the one basket'--which is but a manner of saying, `Scatter your money and your attention'; but the wise man saith, `Put all your eggs in the one basket and--
        watch that basket!'
        -- Mark Twain (emphasis mine)
        • but the wise man saith, `Put all your eggs in the one basket and--watch that basket!'

          This is why the American parts of the Internet backbone should be administered and maintained by the Department of Homeland Security or a division of it. We must ensure that terrorists do not take down this vital information super-highway. Who better than Tom Ridge? Corporations? I don't think so, they're part of the conspiracy!

        • Re:Eggs in one basket... (Score:4, Insightful)

          by mat catastrophe ( 105256 ) on Tuesday November 26, 2002 @08:47PM (#4764329) Homepage

          Unfortunately, when there is only one basket, there are suddenly many, many thieves.

    • Centralization (Score:3, Insightful)

      by signine ( 92535 )
      I don't see why centralization would come up though, regardless of who owns the fiber, it's still in the same place. The routers are also still in the same place most likely, which basically means what's getting centralized are the servers, and we already know that. Imagine how many fewer webservers there would be if San Jose were to lose connectivity, or New York for that matter. It's also possible that with fewer providers we have fewer routers which means there are fewer places BGP is routing with. This decreases fault tolerance, of course, and to some degree performance. It's like how when you're in Iowa you see most of your traffic going through Kansas City, even if it's going to Chicago.

      *shrug*

      • Re:Centralization (Score:2, Insightful)

        by questionlp ( 58365 )
        One way that I think having less, but larger companies owning more of the backbone could be potentially be a hazard is that large companies are harder to steer, cutting costs and overhead is always on their minds (which means the could be cutting down the quality of the engineers and staff, which could create weak spots in portions of the company), and the fact that the larger the targets are... the more prone they will be to dissidents and people who just plain don't like them (i.e.: Microsoft or AOL/TW).

        Just a kludge of thoughts that crossed my mind.
    • Surely you mean increased centralization, however.

      Two Journalists Conversing.

      Journo 1:Story needs more punch. A snappy headline will fix that right up!
      Journo 2:Howabout something like a total collapse of the internet.
      Journo 1:Still not quite enough, needs a fashionable target. Let's see. Americans...no, that's overplayed. I know! Commercialization!
      Journo 2The story doesn't have anything to do with commercialization!
      Journo 1:Is that 'journalistic integrity' of yours flaring up again Bob?
      Journo 2:Yeah, I'm thinking about getting that checked
    • Before we jump onto some kind of legislative solution, I think all efforts of everyone in a position to make a difference (and that is everyone) should spread the word about meshnetworks [meshnetworks.com].

      Assuming we can de-regulate sufficient spectrum, wireless ad-hoc networks will completely solve the problem of network vunerability, centralization and commercialization. Meshnetworks have the potention to dentralize benadwith distribution in the same p2p decentralized content distribution.

      Planet P [planetp.cc] - Liberation with Technology.

  • We can only hope (Score:5, Funny)

    by Rick the Red ( 307103 ) <Rick.The.Red@ g m a il.com> on Tuesday November 26, 2002 @06:43PM (#4763501) Journal
    The report compares an attack to travel problems when traffic is disrupted at O'Hare. Hopefully someone in a position to act will pay attention.
    Damn straight! Chicago needs at least two more airports, one south and the other southwest of the city.

    Oh, you were using O'Hare as an example? Nevermind.

    • I take it there will be plenty of collisions?
    • Are they going to search your packets for knives and knitting needles before you log on?

    • Re:We can only hope (Score:4, Flamebait)

      by rnturn ( 11092 ) on Tuesday November 26, 2002 @07:16PM (#4763796)

      I'm puzzled why this was rated as ``off-topic''. Guess there wasn't enough anti-Microsoft content.

      I think the analogy with the airlines' penchant for these hub airports is right on target. (Though I think O'Hare gets an unfair level of criticism; problems in Denver -- especially in the winter -- and Dallas cause similar levels of disruption.) The airlines do it because it cuts costs. No need for as many mechanics and all the other ground personnel if you concentrate your operations in fewer sites. Same thing with data centers. C-level execs just love it when they can consolidate data centers because they can cut their leased office space costs, operations staff, etc. (Though, somehow, they never seem to catch on about the problem this causes with disaster recovery and then bawk at how much it costs to keep a second site available.) So why would we be surprised that the bean-counter mentality is found to exist within the companies that are providing the basic internet connectivity? After all they (the bean counters) are doing their job and if others in the company can't do their job of making sure the networks are available... well that's the other guy's problem. Too bad maximizing shareholder return was allowed to override the job of maintaining an available network.

      • Re:We can only hope (Score:3, Insightful)

        by daeley ( 126313 )
        I'm puzzled why this was rated as ``off-topic''. Guess there wasn't enough anti-Microsoft content.

        He should have linked to this picture [windowscrash.com]. :)
      • ... and then bawk at how much it costs to keep a second site available.

        ``Bawk'' is the noise a chicken makes. Very appropriate here. ``To balk at'' means to struggle against, or complain about, or so. That would have been the more usual thing to see in such a sentence.

      • I don't know how O'Hare is for passenger service, but speaking as one who occasionally has to ship live freight by air cargo -- of any major hub in the U.S., O'Hare is *the* airport most likely to lose, delay, misroute, or otherwise mung up a shipment that has to switch flights in Chicago, particularly if (ghod forbid) it has to switch airlines too. It's been this way since waaay before the computerized era, so I'd guess it's something inherent to the airport. (Tho when United first went computerized, its freight operations there got vastly worse.)

        Denver's 2nd attempt at a cargo port wasn't so hot either (bad weather or not, tho that doesn't help). If your freight has a choice, go thru Salt Lake or Minneapolis -- both are super-efficient.. and have been since (all together now) waaay before the computerized era.

        Point being (I think I had a point :) no amount of internet-security or data-management efficiency is going to fix an inherently-lousy system, nor its lack cripple an inherently-efficient system.

  • Commercialization of the Net is a Bad Thing? STOP THE PRESSES! Next thing you'll tell me is that Linux is still hard to install, and *BSD is dying!
  • and no not Internet2, that's just faster stuff.

    MIT got a grant for those DHT (distributed hash table) thingamajiggers, remember?

    Project homepage here [mit.edu]

  • From the article: (Score:3, Insightful)

    by RomikQ ( 575227 ) <romikq@mail.ru> on Tuesday November 26, 2002 @06:46PM (#4763525) Homepage
    The 11 September attack knocked out net hubs

    Can someone please explain WTF does that have to do with anything? Do they just throw that kind of stuff in as an onbligatery 9/11 reference?
    • Well in this post Sept 11th world, if we don't reference Sept 11th, the terrorists have already won!
      • By that logic, if we stop referencing 9/11 then the terrorists have already won, and if they've already won then they'll stop attacking us, right?

        • Re:From the article: (Score:1, Offtopic)

          by flewp ( 458359 )
          But if we let them win by discontinuing references to Sept 11, then what reason will the government have to strip us of our rights and extend a fist over the world?
    • A single attack was able to take out a large amount of net routing software. A similar attack, targeted at one of the net's chokepoints could be disasterous.

      It's not just a silly reference. It's a demonstration of the fact that an attack like that could have dire consequenses to the net, and at this point, there's not much we can do about it.

      Now, if they'd said "Sept 11 caused people to run around screaming, tripping over datacenter cables and unplugging the net", then I would see your point, but as it stands, it's a valid example.

      • A single attack was able to take out a large amount of net routing software. A similar attack, targeted at one of the net's chokepoints could be disasterous. It's not just a silly reference. It's a demonstration of the fact that an attack like that could have dire consequenses to the net, and at this point, there's not much we can do about it.

        We could hold insurance companies and private companies accountable for future terrorist attacks. By promising immunity to future attacks, our government has effectively taken away any incentives for insurance companies and corporations to lower their exposure to terrorist threats.

        If you were the CEO of a large american company, why should you decentralize anything if the government was going to bail you out. There is no business case for it. If you underwrote and insured the future World Trade Center, are you going to be as cautious as you need to be if the government was going to bail you out. I don't think so.

    • The WTC attack caused major damage to several Verizon central offices and a bunch of cell sites. This resulted in extensive disruption to voice and data services in lower Manhattan. According to Verizon, 3.6 million data circuits were damaged or disrupted.
  • Why hasn't the US government taken up some of the challenge? Surely they have the ability to set up infrastructure in a decentralized manner?

    When replying to this post, keep in mind that I am not addressing this issue from a free speech/privacy of individuals point of view. This is simply a question about why the government isn't interested in taking up this challenge.
    • This isn't the government's job. Surely you wouldn't recommend that the government start dabbling in long-distance voice networks, as well, would you?

      Besides, the internet isn't a "US-only" thing. While you can improve things on your home soil, the companies that operate the backbones extend beyond just one country; there's only so much the US government could do.
      • The U.S. Government has been "dabbling" in long-distance voice networks for decades. The government has a strong interest in having a secure and survivable long-distance voice network. AT&T worked closely with the federal government in hardening the long-distance network against natural disasters and military attack.
    • They will have to take it up; the huge database of personal information to be created under the terms of the homeland security bill will require it.

      This database will be vulnerable not only to direct attack but to attacks against the internet (on which it feeds).

      On the flipside, however it has often seemed to me that governments around the world, particularly in democratic nations (so-called, more accurately 'media-cracies') governments have been steadily giving up control of critical infrastructure resources to multinational companies, almost as if they are trying to phase themselves out.

      • Re:Why not get US in on this? (Score:5, Funny)

        by rnturn ( 11092 ) on Tuesday November 26, 2002 @07:22PM (#4763846)
        ``...that governments around the world, particularly in democratic nations (so-called, more accurately 'media-cracies')''

        Shouldn't that be: mediocracies

        :-)

        • yes I'd thought of that one too, rhymes with mediocrity.

          Unfortunately, the political power wielded by mediacracies is far from mediocre.

          (Off topic I know but) Thing is, in a democracy, what you really have to look at is how do people decide how to cast their vote.

          Then consider whether or not advertising works for *anything*

          Then consider who runs advertising and media coverage of the antics of politicians.

          Who has the power?

    • The government already has that in place and they don't want to share.
      Why do you think they can do videoconferences from Afghanistan?
      Much of that runs on commercial backbones and is just as vulnerable.

      Decentralize more of the net, spread out the backbones and remove the bottlenecks.
      Instead of having 5 or 10 backbone providers running through the same fiber bundles waiting to get cut by a "cable seeking backhoe".
      Run more of the infrastructure through residences, which are all interconnected. Instead of having a spider web type of infrastructure there is more of a fishing net infrastructure.

      You break one leg of a spider web a section collapses (Think City)
      You break one leg of a fishing net there is one broken segment (Think Street)

      I fear a "cable seeking backhoe" more than I do any terrorist attack or a router failure, I have seen more regional digital havoc reigned from backhoes than directed attacks.

      FEAR the "cable seeking backhoe".

    • Re:Why not get US in on this? (Score:5, Insightful)

      by Uruk ( 4907 ) on Tuesday November 26, 2002 @07:19PM (#4763824)
      The government is the absolute antithesis of decentralization. Look at the heirarchy - if there's anything that public servants and the government structure as a whole is known for, it's a pecking order. Government doesn't understand decentralization, because ultimately that tends to make things harder to control and administer, and governments are all about controlling and administering. That's their core goal.

      The government's primary self-chosen mission in most countries of the world today is to promote economic growth, which often is interpreted as doing whatever the industrialists ask of them. And guess where the industrialists stand on the commercialization of the internet....

      • Re:Why not get US in on this? (Score:4, Interesting)

        by sakeneko ( 447402 ) on Tuesday November 26, 2002 @07:35PM (#4763925) Homepage Journal
        The government is the absolute antithesis of decentralization. Look at the heirarchy - if there's anything that public servants and the government structure as a whole is known for, it's a pecking order. Government doesn't understand decentralization....

        The Internet was developed under the watchful guidance, and using the money, of none other than Uncle Sam -- the U.S. government. Way back in the early days of the ARPAnet, it was deliberately made decentralized, and designed to treat any blockage to the free flow of information as damage, to survive a nuclear attack.

        Perhaps the government won't be willing to pay the bills to keep today's Internet from becoming overly centralized, but it knows how.

        • What I was replying to is a request to get the government involved. Yeah, DARPA was a part of the government, it was a sub-agency of a sub-agency of a sub-agency of the defense department (which is a sub-agency of a major branch, which is a sub-agency of the government) Jeez, did you notice any heirarchy in there?

          Oh, and nobody petitions DARPA to do things for them. You petition the government - your congressman, where all of my comments still apply. Using the DARPA argument that government knows about decentralization is like saying that because a 1,000,000 person company employs a contracting company of 500 that have a clue, that the company therefore has a clue.

          Which just ain't so.

      • You're absolutely right! Can you even imagine some part [arpa.mil] of the government trying to think about decentralization? Ha! Their brains would probably explode [isep.ipp.pt]! Stupid governments.
      • The problem of centralized control is different from the problem of centralized points of failure.

        Sure, if the government decides to break the infrastructure, it only has to make that choice once. That is the problem of centralized control.

        They are perfectly capable of putting connex between every police station in the nation, though, and providing incredibly decentralized points of failure. In fact, that's what they've done. There was some federal bill for emergency communications centers, so now many new police buildings take federal money. The feds pay for the whole building in exchange for using the basement as a communications center.

        The question is, are you more worried about a backhoe taking out an essential backbone, or are you more worried about our government turning into communist China. I'd say the backhoe is more likely, just because it already happened.

        Of course, the reason you're opposed to this isn't because the government can't do it properly. It's because you think the government would spend too much money doing it. And of course, you're right. Don't mean to bait, but when you start acting like you have some other set of reasons... you sound like a liar.
        • The question is, are you more worried about a backhoe taking out an essential backbone, or are you more worried about our government turning into communist China. I'd say the backhoe is more likely, just because it already happened.

          You haven't actually read the so called Patriot Act or the Homeland Security bill, have you?

          Technically, we're not communist, nor are we called "China", but every piece is now in place to make Orwell's vision a reality.

          Please start paying attention before it's too late if it isn't already.

  • I dunno... It's just the same as anything; If you put all of your eggs in one basket, and somthing happens to the basket...well, you're screwed...
  • From the Article:
    "In its early days the net was as decentralised, as possible with multiple links between many of the nodes forming it. If one node disappeared, traffic could easily flow to other links and route traffic to all parts.

    However, said the researchers, the increasing commercialisation of the net has seen the emergence of large hubs that act as key distribution points for some parts of the web."

    As a result, the net has become much more vulnerable to attack.
    Finally, someone other than a corporate Paki is commenting on the health of the internet. It is no longer an internet, but rather interconnected proprietary WAN's.

    • Re:A Simple Internet Model (Score:5, Funny)

      by Em Emalb ( 452530 ) <ememalb@gm a i l . com> on Tuesday November 26, 2002 @06:55PM (#4763615) Homepage Journal
      In what was considered a shocking move today, members of the Mouse Movement known as You moved my Cheese, you Rat Bastard, or YMMC,YRB for short, have declared war on the ever popular internet.

      Speaking from his private "nest" in the foothills of Santa Barbara, General Carlissimo P Rodentia had this to say:

      "You have bombarded my people for years with your unwanted peecees and aol ceedees. No longer. Your precious internet cannot stand the assault of 100 billion of my brother's and sister's teeth. Consider yourselves warned."

      A truly ominous sign of the times.

      Signing off, this is Reginald Rattus, reporting.

  • Resilience to Attack (Score:4, Insightful)

    by otisaardvark ( 587437 ) on Tuesday November 26, 2002 @06:49PM (#4763559)
    Internet access and bandwidth are very vulnerable, but remember there are lots of copies of the DNS server records, and the actual content is extremely widespread and can easily be put online again given some time - in a genuine emergency situation internet access would only be a priority to those on the periphery anyway. Fine, we need more hubs and greater decentralisation, but lets not get carried away.
      • A long, long, time ago, the net was very decentralized, but not very scalable.
      • Then they (Al Gore) helped invent the ARPAnet, which was a backbone that all the regional nets tied into, which made things much more centralized - if the Arpanet routing protocols freaked, which they did periodically, parts or all of the net fell off, and the Acceptable Use Policy censored business use of the net, though fortunately it took an increasingly lax attitude toward interpreting the policies for "official" use.
      • Gradually things like the Commercial Internet Exchange helped ISPs build around the ARPAnet's backbone, the Feds funded FIX West and FIX East to manage their traffic, and funded the various MAE and NAP complexes partly to keep connectivity between the commercial and government-funded worlds, partly because it was cool, and partly because the Feds wanted to keep some control over what the Internet was doing even though it was no longer theirs. (Among other things, there were active wiretap reflectors running on MAE West.)
      • But the MAEs were too overloaded and non-scalable for ISPs to run reliable traffic, so they started building private peering, which has gradually become quite extensive. Most major ISPs do 90-99% of their traffic on private connections (either peering or transit, depending on how you follow the money.)
      • Access costs, operational costs, and economies of scale catalyzed building of internet hosting centers by a variety of businesses. During The Bubble, there was a wide ecosystem of businesses providing services within these spaces. Some of them were carrier-operated, with interconnections provided by the carriers, while others were carrier-neutral (or at least had 2-3 primary carriers instead of just 1), so the interconnection topologies varied greatly.
      • Competitive businesses are increasingly building carrier-neutral facilities to increase interconnections; some of them centralize the locations where private peering happens, but to a large extent they're mainly displacing interconnectivity that occerred in telco pops.
      • Europe's Internet connectivity is somewhat more dependent on centralized exchanges, such as LINX and AMSIX, in part because of former telecom monopoly policies on in-country and between-country facilities pricing.
      • Telecom liberalization in Europe and Asia has greatly increased the variety of connectivity between countries, and Global Crossing and their competitors have radically increased the amount of cross-ocean bandwidth and physical diversity. (It may seem otherwise, since it's one bankrupt carrier's fiber doing the majority of trashing everybody else's business model through first-mover advantage, but believe me, trans-ocean bandwidth was *much* less diverse a decade ago, and the internet fraction of that was even less diverse, even for the then-current definitions of "high bandwidth".)
      • The following 5-10 years will be dominated by chaos and anarchy, with major players appearing and going bankrupt, but unlike the software business, where a bankrupt company usually vanishes, a bankrupt fiber carrier sells off its access for pennies on the Euro to some new carrier who then proceeds to undercut the fragile stability of prices the other carriers are briefly enjoying. Some people like to predict that we'll end up with about 3 carriers before there's anything resembling stability, or at least before the price of the fiber bandwidth accross the ocean becomes cheaper and less interesting than the price of the bandwith in the last 100 meters or the uninterruptible power supply system at the destination.

      • DNS was originally Jon Postel's hosts.txt list, plus everybody else's hosts.txt, which was quite decentralized, and occasionally coordinated with the UUCP decentralized naming.
      • Then it was Jon's IANA, which was well-behaved, but alas, quite centralized.
      • There were a few competitive-root proposals, like Kashpureff's, but they never really caught on.
      • There were also country-code TLDs, which were decentralized governmental control (by a bunch of generally monopolist telecom authorities, but at least they weren't cooperating with each other, and weren't interested in .GOV being the US in charge of the World's governments.
      • The IETF's Ad-Hoc committee tried to broaden the DNS, but the Powers That Be squelched that.
      • ICANN emerged, declaring itself to be in charge, with US government backing, and enough people believed them that it now appears to be true, regardless of real legitimacy. They're strongly in favor of centralized control of any decisions affecting intellectual property, prevention of privacy, and dispute resolution, and discouraging experimentation with policy and technology.
      • In the non-US jurisdictions, country-code TLDs have become a hot commodity, and some countries have been willing to sell off use of their initials to various commercial companies which take a more divergent view of policies and pricing. Unfortunately, their power grab has included declaring ownership of the IPv6 namespace and setting prices at a level to discourage use of it.
      • To give ICANN some credit, they've at least called for decentralized pricing and sales, though with a centralized database registry, which has made it easier for commercial activity to provide some variety in names available to the world's general public.
      • The Distributed Denial of Service vandals demonstrated the continued efficiency of centralized control of distributed resources.
      • Some of the quasi-centralized Root Server cabal are developing decentralized implementations of servers for the centralized DNS namespace, which should help the centralization problems at least operationally.

  • Commercialization ruins so many things. (Score:5, Insightful)

    by vectus ( 193351 ) on Tuesday November 26, 2002 @06:49PM (#4763561)
    The Internet really isn't alone. Ads have really taken over society. Everywhere you look, from people's clothing to the garbage on the ground, to blatently all over every layer of packaging on the goods you buy at the local Safeway.

    I've gotten so sick of it. The reason I switched to Linux (probably the dumbest reason in a lot of people's opinions) was to escape the fact that every program I installed had huge logos and ads plastered all over.

    I remember when you were mocked and considered weird if you sold out. Now, if you don't sell out, you're considered stupid for not making money while you can.

    I get the feeling this blatent lack of ethics will be part of the downfall of our economy. You can only have so many people leeching at one time before it runs out of blood.
    • And what are you going to do about it? Absolutely nothing! Here's the real problem with commercialization: despite the fact that everyone thinks it sucks, nobody wants to do anything about it.

      The commercial forces are "driving the economy", providing jobs, providing tax revenues to governments, and filling pockets all over the globe. Why on EARTH would anyone EVER want to turn away from that path?

      The grand success of commercialism is tying the interests of the rubes^H^H^H^H customers to the interests of the corporation. Trust me, as much as you bitch and moan about commercialization now, if it weren't there most people would be twice as pissed off at the loss of their wonderful privacy-invading, wallet-vaccuuming feature-creeping, RAM-sucking functionality that allows them to talk to hairy-backed 50 year old men posing as 14 year old school girls any time, day or night, from anywhere on the planet!

    • Sorry you feel that way, but if ads = free things, I'm all for them. As long as they're easy to ignore and don't require me to do anything, I'll take my free radio/sitcoms and cheap postal service over paying a lot for them.

      And every once in a while, I see something I might like to buy.
    • You have legitimate points, but you're not talking about commercialization bringing down our economy... you're talking about lack of ethics bringing down our economy. And while you're complaining about all the bad things commercialization does for us, don't forget completely about the good things it's brought us.

    • The "invisible hand" of market forces does not always outperform a regulated piece of social infrastructure. It's high time we started the dialogue between the lassais faire capitalist/libertarian crowd and the socialists.

      You see, a free market cannot exist without the social infrastructure of a legal system and a police state to enforce it, and the critical consensus to support good social infrastructure cannot exist without the freedom to violate the social norms and critically compare actual alternatives. We live in a mixed economy, both social infrastructure and free-market aspects are necessary. Some things should be given: free (peer-to-peer) telecommunications for all! Otherwise you have a "closed" free market with limited internal market forces to regulate it.

      We should socialize the Internet as a free (as in beer ALSO as in freedom) resource to STIMULATE the free market part of society by providing more pressures from everyone. You have to look at what is going on and ask yourself: "could it be better? Should it be more cooperative or competitive? Where is the balance? Why?"

  • They do have a point... (Score:3, Insightful)

    by IdleTime ( 561841 ) on Tuesday November 26, 2002 @06:50PM (#4763567) Journal
    They do have a point here.
    The fewer centralized points the traffic has to go through the higher the risk of failure. And with failure, the lack of service to millions of people.

    I can't validate the correctness of the story, but my impression has always been that the backbones are designed to failover if they hit a problem and that there are several routes between multiple backbones that is serving the same strecth of net. I may be wrong on this, but at least that was the goal back in the 80's when I first started using the net.

    The article needs to be taken serious, as more and more business depends on the net. If it fails one one or more backbone stretches, it will have enormous consequences for business, meaning your's and my paycheck may be endangered. Oh, and the answee is not to get rid of Microsoft in this case :-)

    • Re:They do have a point... (Score:4, Interesting)

      by 1984 ( 56406 ) on Tuesday November 26, 2002 @06:57PM (#4763639)
      Depends. A telco has a network, which carries IP traffic (perhaps other traffic, too). That may or may not have multiple routes within it connecting any two points. And it may peer with other networks at various points. But it's not necessarily a given that a) if a big network disappears that there'll be routes *besides* that network connecting everything that was connected to it, or that if such alternatives exist, that they'll have sufficient bandwidth to cope with the loss of that network.

      After all, it's notionally not economic to keep too much excess capacity around -- why bother? So it'd be a surprise if ever major route was 100% (or more) backed up by another major route.

      Also, physical separation and logical separation are different. A large logical separation may, alas, boil down to two pieces of fiber in the same conduit, two wavelengths on the same piece of fiber, that sort of thing.

      So yes, it *can* all be made to be redundant, but that's not neceesarily how it plays out. Other factors may act against redundancy.
      • ``Also, physical separation and logical separation are different. A large logical separation may, alas, boil down to two pieces of fiber in the same conduit,...''

        I recall pointing out something like that to a boss many years ago who was proud of the fact that they'd put in place redundant leased (analog and expensive) lines running to a remote (and I mean remote) facility. He looked a little pale when I mentioned that I only saw one set of phone poles leading up to the site. Until then it'd never dawned on him...

  • But life will go on.... (Score:4, Insightful)

    by SmoothOperator ( 300942 ) on Tuesday November 26, 2002 @06:51PM (#4763573) Homepage
    "If you destroyed a major internet hub, you would also destroy all the links that are connected to it," said Morton O'Kelly, Professor of Geography at Ohio State University.


    It would have ripple effects throughout the internet..."

    ... and the Montana rancher will still herd his cattle, and the wine-maker in Italy will still stomp his grapes, and the crossing-guard will still be out there at 7 AM... Life will go on, boys and girls, life will go on, like it has before the 'net...

    • OK, yes, in the aspect that you describe, life will go on. Those jobs you listed probably do not have a business need for the internet.

      Tell that to a bank, or a mdeical facility. Data communications are a very important everyday part of life. If you disrupt it, sure, the low-tech grape stompers won't see a thing. Any company with a web presence will though.
    • However, that Montana rancher may have one hell of a problem getting to his CitiBank account and that low-tech grape stomper might find a "CLOSED" sign on that farm whose ability to take orders from the companies they supply suddenly got unplugged.

      I'm amazed to see comments like yours on a tech forum. Civilization has put its eggs in the internet basket. Basically, because it's cheaper.

      Most data traffic having to do with operating the supply chain that gets those grapes to your grocery store in terms of wine and that cattle rancher's product to your store in terms of steak goes through the Internet. Even in the cases where this isn't so, you can bet that at least a few critical links in the supply chain are via Internet.

      Could workarounds be found? For the short term, maybe. However, perhaps you'd notice if the price of milk in your grocery store went up 50% or average prices at WalMart went up 100%.

      The only people who wouldn't notice the effects of a long-term loss of the Net are so remote from civilization that the international market economy doesn't touch them much, and that doesn't even describe most of the Third World. They might not know why they suddenly can't make a living or the price of anything imported doubled or worse, but they would notice.

    • ... and the Montana rancher will still herd his cattle, and the wine-maker in Italy will still stomp his grapes, and the crossing-guard will still be out there at 7 AM... Life will go on, boys and girls, life will go on, like it has before the 'net...

      And then Kevin Costner will show up and reestablish the postal service. It's cool, don't worry.

  • let's see, I'm an ISP, all I want is to get connected to the internet as cheaply as possible. Thus I don't want too many links to other ISPs raising my expenses. Okay now I need to choose who to peer with, there's big company A and another, smaller network B. Who am I going to choose? A of course, because it will be more reliable and have more direct links. Thus you have a few large companies connected to a lot of ISPs and if a couple of those go down, then ISPs start routing strictly through another one, which causes it to get significantly higher traffic. Conclusion: ouch. laaaag and problems with reliability and other routers going down

  • It's simple: Less Security = More Convenient (Score:4, Insightful)

    by dagg ( 153577 ) on Tuesday November 26, 2002 @06:53PM (#4763597) Journal
    The rule of "less security = more convenient" applies in nearly all situations... and it applies here as well. The only way to increase the security in this particular situation is to de-centralize the big hubs. But that will be very inconvenient to the big companies that own the hubs. There are few reasons to do inconvenient things.

    --
    Sex Gateway [tilegarden.com]

  • Hopefully someone in a position to act will pay attention.

    Yeah right! - They'll pay attention when it breaks.

  • The desktop has been in the control of very few hands for years now and aren't we all better off? ;)
  • If P2P networks hadn't been used for illegal purposes right from their creation, I wonder if maybe more the the Internet might be in a P2P form at this late stage. Certainly, you can't disagree that P2P didn't get pushed back in the technology development cycle because of all the political issues surrounding it. It wasn't "embraced"...

  • The Greater Danger of Centralization (Score:3, Insightful)

    by Bouncings ( 55215 ) <ken&kenkinder,com> on Tuesday November 26, 2002 @07:06PM (#4763703) Homepage
    Ok, security is harmed. Given. But to me there is a much larger problem with the centralization of the Internet: control.

    Think for a minute, what country has about the most centralized internet backbone? That would be China, or, The Great Firewall of China. Look at it this way, in order to Do Something Really Bad in China, they have to implement it on one set of backbones with one central authority.

    Now that the backbone is mostly owned by big business in the United States, it centralizes control of the Internet toward big businesses. Which yeah, could really pretty much suck.

  • from the story: Hopefully someone in a position to act will pay attention. hopefully someone in a position to act will not have to be told about potential dangers by the BBC
    • ``...hopefully someone in a position to act will not have to be told about potential dangers by the BBC''

      Yah but one wonders how many deaf ears in D.C. the results of the study by the professors at OSU fell upon. Funny that it didn't seem to escape the notice of the folks at the BBC. If the feds were even concerned about this problem, they were probably assured by the telecom lobbyists that it was nothing to worry about.

  • Missing Key Point (Score:4, Insightful)

    by Hamstaus ( 586402 ) on Tuesday November 26, 2002 @07:12PM (#4763752) Homepage
    In its early days the net was as decentralised, as possible with multiple links between many of the nodes forming it. If one node disappeared, traffic could easily flow to other links and route traffic to all parts.

    I would not give this article a lot of serious thought. It describes how simulated attacks show vulnerable spots in the internet, and seeks to lay blame for it. However, comparing the current state of the Internet to it's own beginnings is obviously going to show differences (DUH!). I mean, back in the pre-web days (you remember those, folks? ah, sweet gopher. R.I.P.), if you didn't know exactly where or what you were looking for... well... none of this fancy googlin' stuff, that's all I gotta say.

    If you consider the growth of the internet from that point, which was basically a loose, random interconnection of .edu's, .gov's and .mil's, there was no need for centralization. However, suddenly, one day everyone wanted to be on the net! And out of that chaos, logical central points developed.

    I like to explain the internet to non-techie people as something like the Interstate highways in the United States. And using that metpahor... if you take out a central location... well, it'll be a lot slower and harder to get to where you need to go, but it's not like you've isolated an entire region for all eternity.

    My point is, there are centralized locations because it was efficient to do so. Eventually, as more and more high speed wire is laid out across the world, these will slowly become less important. It's just that the growth has been too fast for the present time!
    • you remember those, folks? ah, sweet gopher. R.I.P

      Actually, there's a movement underway to bring the roots of the Internet back to the forefront. For instance, gopher isn't really dead, it's just residing here [quux.org], among other places, waiting patiently for the commercial Internet to implode so the good old days can return again.

  • You can get the PDF for the paper here: (Score:4, Informative)

    by Saint Aardvark ( 159009 ) on Tuesday November 26, 2002 @07:12PM (#4763753) Homepage Journal
    Click on this link:

    http://www.elsevier.com/locate/tele [elsevier.com]

    You'll see "View their sample issue." Click on that, then click on the link for Volume 20, Issue 1. Go there. Then you'll see "A geographic perspective on commercial Internet survivability", and you can download the PDF there.

    Looks like it's meant to give you only one chance at the free issue, so I think giving the direct link would be pretty useless. Whatever; you're only three clicks away from greatness. :-)

  • Eh. (Score:2)

    by Uruk ( 4907 )
    Why would anyone care? Isn't it extremely intentional that the control of the internet is being consolidated into fewer and fewer hands?

    Who is it exactly that would object or do something about it? Do you think the few companies who own major backbones are going to decide that it's not in everybody's best interest and sell their portion off to 10 other companies?

    Sure, this is a bad thing, but it's done in order to suit the interests of the people who are doing it. The idea that somebody would wake up, decide this is absurd, and correct the error of their ways is absolutely ridiculous.

    Of course, we could always hope that MS would realize their software licensing is not in the best interest of the consumer and turn it all around....but it's statistically safer to bet on being struck by lightning 12 times in succession...

    • "Of course, we could always hope that MS would realize their software licensing is not in the best interest of the consumer and turn it all around....but it's statistically safer to bet on being struck by lightning 12 times in succession..."

      Let's see, get struck by lightning or go through a Windows "upgrade". Tough call.

  • CIDR and the centralisation of routing is to blame (Score:3, Insightful)

    by ikekrull ( 59661 ) on Tuesday November 26, 2002 @07:13PM (#4763765) Homepage
    Obviously there were good reasons to introduce CIDR (Classless Inter Domain Routing) and concentrate the ability to route around problems to the 'core' of the internet, but this is the price you pay.

    The only way real redundancy and fault-tolerance will be restored is to introduce IPV6 - or some other means to widen the availablity of routable IPV4 space, and remove the barriers currently in place for people to partipate in the 'routable' internet.

    Of course with this comes lack of control for MPAA/RIAA/Governments, increased freedom for independent operators, and also increased complexity and route-table storage requirements for all.

    However, if the internet is to withstand prolonged and/or distributed attack, then the ability to route effectively will have to be extended further toward the edge of the net than it currently is.

  • Torrerists! (Score:3)

    by Dark Lord Seth ( 584963 ) on Tuesday November 26, 2002 @07:13PM (#4763767) Journal
    Simulated attacks on key internet hubs have shown how vulnerable the worldwide network is to disruption by disaster or terrorist action.

    In other news, Bin Laden has been sighted in Saudi-Arabia with 20 Al-Quada script kiddies. Latest findings of the CIA conclude that Bin Laden is trying to build a biochemical weapon throwing sludges of contaminated biomatter called a "GES BioRifle" and that Australia has mysteriously disappeared of the world map. Weapons experts disagree with these findings, claiming "Redeemeers" are much better, though the news about Australia was ethusiastically welcomed with cheers like "No more Steve Irwin or Kylie Minogue!".

    However, a recent investigation in some random MS Monopoly lawsuit indicated that Bill Gates does indeed cheat, playing with several copies of the authentic Broadway and Park Drive cards, as well as a recent donation of 20 Windows XP Pro packages to Palm Tree Nr 137 in Saudi-Arabia with a note reading "BOMB FINLAND" and enough funds to construct a backbone connection to Saudi-Arabia. US officials are skeptic about the current findings, saying "Haven't we blown up Saudi-Arabia yet? Oh, that was Australia?" Several high ranked military officials were unavailable for comment, but disapproved of Bill Gates cheating at Monopoly.

    Coalition forces have responded by pre-emptively bombing Iraq like they have done for the last decade. US fighter-bombers scrambled and succesfully bombed 3 hospitals, 2 schools and a Burger King in Washington DC. Brittish commandoes went in and simply cut the backbone connection with Margaret Thatcher's fake teeth. Bin Laden and the 20 script kiddies have escaped, leaving a videotaped message behind, calling for a holy war against the US and against Saudi-Arabia for disconnection power to Palm Tree Nr 137. Bin Laden was last seen hiding on the North Pole in a red suit, a sleigh, a bunch of biochemical reindeer and 20 script kiddie elves. US bombers are underway as this article is written.

    Film at 11.

  • Grassroots net (Score:3, Interesting)

    by etcshadow ( 579275 ) on Tuesday November 26, 2002 @07:14PM (#4763774)
    It would be interesting to see if more people started running alternate routes through friends houses and what not. A guy I work with has a p2p 802.11b link to another guy I work with's cable modem 5 miles away, despite having DSL himself. I know that when I pulled my (late) linksys router out of the box, I was surprised to see that it supported RIP.

    The truth is that it is really not that hard to run multiple routes out of your bedroom. If you use *nix for your router (like I do since I burned up my linksys), it's as easy as dropping in another NIC (wireless, or ethernet, or modem, or whatever) and configing the new interface.

    There's also the growing trend in community nets (particularly wireless community nets)... these could link themselves together fairly cheaply by setting up additional wireless links with directional antennae pointed at other peer community nets.

    Anyway, I'd be curious to see how many new routes start springing up between these 2nd-class (and no-class) networks. The beauty of Internet Protocol is that this really works.
  • with a 3 14.4 kbps dialup modems (arpanet, i mean?) i have one, i know another guy with one. if the internet was in *real* jeapordy, couldn't the universities, and induviduals just 'start fresh'...i mean the rfcs' appear to give a pretty much bleuprint method of how to move..right? what is really stopping us from building supercomputers...etc...? especially with modern tech, we could just buy a regular computer, get a whole load of modems...and go back to TTY ! why not?
    of course, i'm concerned about the internet as anyone, but i'm connecting currently through stolen bandwidth anyways - the 'net is too expensive for most people it seems to me...decentralization could probably help that, though...but keep in mind...no matter how bad it gets, we can always start anew, so long as we have those 3 14.4kbps...
    • It could be useful to point out once again that multiple interconnections and multiple routes was an important part of the original Arpanet that led to the Internet. It was (as the commercial people keep forgetting) a project funded 100% by the US Defense Department, and they wanted a network that would survive in battle conditions. Fact is, this is also a good design principle for design in a world where many of the components have a MTBF of days or weeks.

      Problem is, commercial folks invariably see reduncancy as a needless expense. Their natural tendency is to reduce everything to the bare minimum (while selling the maximum, of course). Then when anything breaks, big chunks of the system are down.

      The World Trade Center attack is an excellent example that woke up a lot of people. There was far too much infrastructure passing under those buildings, and as a result, a lot of the communication systems in Manhattan collapsed along with the buildings. This stupidity was pointed out by people before the attacks, but the commercial interests in charge of the comm lines saw no profit in decentralizing. Even now, they're resisting the idea and merely rebuilding a lot of the destroyed capacity, because a better system would be more expensive.

      Governments have stepped in and forced things like the phone, electricity and highway systems to have alternate routes that can be used in disasters and emergencies. The Net is becoming an important part of the world's infrastructure, and eventually those evil old governments are going to step in and force the commercial crowd to supply redundancy in the same way.

      --
  • So the internet is more vulnerable now than In The Olden Days, when the backbone was carried entirely by MCI? It must be that new math.

  • Monoculture Considered Harmful. Film at 11. (Score:5, Insightful)

    by mudshark ( 19714 ) on Tuesday November 26, 2002 @07:35PM (#4763923)
    Yes, Virginia, the health of the Internet *does* depend on decentralized technologies such as multiple backbones, gegraphically distributed root name servers, and standards committees not answerable to any single political entity or product vendor.

    It's no different from a business monopoly, (or cartel, or oligopoly) which tends to create artificially high prices, poor quality of goods and services, and in the case of computing and networks a fertile breeding ground for viruses, worms and other nasty exploits.

    And the analogue these worlds share with real live ecosystems is uncanny: Plant an entire state in one strain of corn for a few seasons in a row and watch the fun.

    Didn't we already learn this crap? Why do the FCC, FTC, SEC and other god-forsaken, nutless bend-over wastes of acronyms keep rubber-stamping all the mergers?

  • bad title (Score:4, Interesting)

    by asv108 ( 141455 ) <asvNO@SPAMivoss.com> on Tuesday November 26, 2002 @07:44PM (#4763974) Homepage Journal
    What does commercialization have to do with the Internet backbone being in fewer hands, shouldn't the title be "Growing Backbone Consolidation threatens Net Security. The last thing we need is G.W. thinking that their are comunists on slashdot. We will all be branded as terrorists.

    • oops (Score:2)

      by asv108 ( 141455 )
      Accidental submit, Why does the submit button need to be next to the preview button? It should be after the drop down list or have forced preview.
    • And it's not an Internet-wide phenomenon. And the claim that the Internet was much better in the old days because nowadays, smaller towns can be drop off the Internet in the case of a major disaster is a bit strange. In the old days, organizations in these smaller towns were connected to the Internet via modem lines, if they were lucky.

      I guess that those organizations which had network access in the early nineties often still have quite a bit of redundancy (despite the backbone consolidation) because they care about their Internet connection, it's often an integral part of their work. The newcomers don't care that much, can afford outages of days in a row, look extremely closely at the price tags etc.

      Or another strange claim:

      "If you destroyed a major internet hub, you would also destroy all the links that are connected to it," said Morton O'Kelly, Professor of Geography at Ohio State University.

      The links are not destroyed, they are still there and could be reconnected in most cases. Of course there would be a major outage, but you still wouldn't have to reconnected the country from scratch.

      I hope the actual paper is a bit better. Despite all concentration, I don't see that physical interference with network components is a major threat to the network. It just doesn't scale too well.

  • Not only organizations, also USA centricity too. (Score:3, Interesting)

    by Samir Gupta ( 623651 ) on Tuesday November 26, 2002 @07:47PM (#4763991) Homepage
    Much of global Internet traffic on the intercontinental level is routed through the USA, even though the origin or destination may be totally outside the USA. For instance, traffic between Asia-Europe, or South America-Australia will almost always pass through the US, because most of those "hubs" are, as the article mentions, in the USA.

    I believe more work should also be done on interconteninental links that do not go through the USA as well.

    I have nothing againt the USA, but the Internet is critical to more than just the USA now, and were the unthinkable to happen again in the USA, there should be redundancy. Also, it would be much more efficent in terms of latency (eg, Europe-Asia instead of Europe-USA-Asia).

  • More or less reliable? (Score:3, Interesting)

    by jcam2 ( 248062 ) on Tuesday November 26, 2002 @07:53PM (#4764024) Homepage
    It seems to me that the commercialization of the
    Internet has brought so much new capacity online
    that it is more reliable than the old days, due
    to the existance of competing long-haul cables
    operated by different companies.

    For example, back in the early 90's Australia was
    served by a single 10mbps trans-pacific Internet
    connection. If it went down (as frequently
    happened), the whole continent was cut off!
    Today there are several links to the rest of the
    world, and outages of that kind are unknown.
    Guess who paid for those links? That's right,
    for-profit commerical corporations.
  • Notice that the article says "hopefully someone in a position to act will pay attention."

    I totally agree that fewer backbone operators == greater "single points of failure".

    However, there is no doubt in my mind that the "people in a position to act" are probably not hanging out at /.

    But then again, they would already be aware of this too, if only for business reasons.

    Unfortunately *very* few people are influential or wealthy enough to influence backbone operation -- does this make these people another "single point of failure"? (honest question, not flamebait)

  • One of the biggest problems in the backbone is that attempting to support arbitrary routing policies driven by a myriad of different customers overconstrains the problem of global internet routing. This leads to configurations in which either many solutions exist or no solutions exist to the routing problem and causes routing instability. Couple this with the fact that router configuration is a black art that is extremely error-prone and you get WorldCom-like outages. Such problems will actually IMPROVE with more consolidation. If you're interested, check this paper [att.com] out.

  • Back in the good old days (Score:4, Interesting)

    by per unit analyzer ( 240753 ) <EngineerZ@NOSpam.gmail.com> on Tuesday November 26, 2002 @09:46PM (#4764639)
    The author of the article is waxing nostalgic about a day that never existed. Back in the NSFNET days (not the earliest days of the Internet but precommercialed nonetheless), if the NSS your regional network was connected to had problems, you would have had certainly felt it. Regional networks connected large swaths (several states) of the US to the Internet much like the author describes what is going on today. Eventually some regionals became multi-homed, but even then many were not designed to properly handle all traffic failing over to a single link to the backbone. I didn't start using the ARPAnet until it's final days, but even then I suspect the loss of a core site would isolate a number of leaf nodes.

    The design of TCP/IP allows for redundancy and survivability, however most if not all of the research backbones that evolved into the commercialized Internet never had a great deal of redundancy. Granted, later incarnations like the NSFNET T3 network were better, but most had single points of failure which could be felt across large parts of the Internet when those points had problems...

    --zawada

  • Bell System (Score:4, Informative)

    by Detritus ( 11846 ) on Tuesday November 26, 2002 @11:57PM (#4765140) Homepage
    If you look at network diagrams of the Bell System, when AT&T still ran everything, you will see a system that was designed to cope with disasters and excessive loads. It provided a great deal of flexibility in how calls were routed through the network. Each central office had multiple links to peer central offices and parent central offices. A call could be routed in many different ways. If a link to a peer central office was out, the call could be "kicked upstairs" to a parent central office, which would route it over a different link to the destination central office. The only single points of failure were the local central office and the wires in between the local central office and the subscriber.

  • This is twaddle (Score:4, Insightful)

    by ethaz ( 413842 ) on Wednesday November 27, 2002 @01:58AM (#4765555)
    With multiple commercial carriers, all operating their own backbones from multiple POPs the likelihood of the destruction of a building, or for that matter, an entire city having an impact on Internet connectivity overall is nonsense. The largest backbone providers, AT&T, UUNet, Sprint, Qwest, Level 3 all operate with SONET rings at the physical layer plus BGP4 routing. And all of them operate from separate physical facilities (UUNet and Sprint don't normally share a building, for example). Further, since the MAEs, the NAPs and other public peering points are, for the most part, irrelevant to the major backbones (their private peering arrangements are separate from these places), their connectivity to each other would survive. Sure, it might need to be shifted from SF to, say, Chicago, in the case of an emergency, but that could be done in a day or so, if not in hours. If anyone of them lost a major node, they continue to operate. The only effected connections would be those directly connected to that disable node.


    This is far better than the pre-1993 days when there was a single backbone, operating on non-redundant private lines.

    I guess this guy wanted some publicity. He got it.

Slashdot Top Deals

"I've finally learned what `upward compatible' means. It means we get to keep all our old mistakes." -- Dennie van Tassel

Close