Our new Indie Games subforum is now open for business in G&T. Go and check it out, you might land a code for a free game. If you're developing an indie game and want to post about it, follow these directions. If you don't, he'll break your legs! Hahaha! Seriously though.
Our rules have been updated and given their own forum. Go and look at them! They are nice, and there may be new ones that you didn't know about! Hooray for rules! Hooray for The System! Hooray for Conforming!

Scrap the internet and start again?

2»

Posts

  • MKRMKR Registered User regular
    MKR wrote: »
    stilist wrote: »
    This is by far the stupidest thing I've seen in a while and unless those researcher's comments are being taken grossly out of context, my opinion of them is seriously degraded.
    :?

    You’re against people experimenting to see if they can improve stuff?
    I'm against the "we'll restart the internet!" concept. Because it is stupid, and patently unworkable. They could sell it like every other new internet technology as a new set of backbone protocols etc. and then they'd have a market because only a few major players are interested in getting the most out of their existing backbones so if they could get more bandwidth by only upgrading routers instead of laying more fiber then hey - they'll do it and from their evolutionary improvements in the speed and cost of hardware will move more people over to it.

    They can already get speed improvements by upgrading routers. Right now we're only using a tiny fraction of what fiber optic cabling is capable of, and it's only limited by how fast the routers can process incoming and outgoing data.
    QED really, though I suppose an argument for formats that are better for routers to process could be made. However that articles make a lacklustre case for substantially changing underlying and established protocols like TCP/IP as opposed to just wrapping them appropriately as needed.

    Is there someone here with a handle on this who could provide a decent critique?

    Even with a year and a half in to a networking specialist degree, I don't know enough to properly asses the merits of redoing the internet. :(

    Hopefully someone with more experience with it will show up.

  • electricitylikesmeelectricitylikesme Registered User regular
    Hm, what the issues CERN is encountering with the LHC seem like a decent metric for where we have performance issues with the internet. LHC is expected to generate something on the order of 1000 terabytes of data per year - all of which pretty much needs to be distributed throughout the EU, US and Japan/Oceania.

    A presentation I found pretty interesting was one which included the statistic that over a 10GBps fiber link, the loss of just one packet effectively reduced the bandwidth 25% for the next 5 hours - I don't know on what mechanism this metric is based.

    Dis' wrote: »
    Cancer is when cells stop letting the body mooch off their hard work - clearly a community of like-minded cells should isolate themselves and do the best job each can do, even if the rest of the body collapses!
  • RichyRichy Registered User regular
    Shoggoth wrote: »
    I don't understand enough about the actual workings of the technology but Isn't too much of the economy invested in the internet to just "start over"?
    That's part of the problem. We have a large and growing part of our economy based on what is essentially a prototype built from whatever people had available at the time and which got way out of hand. And the solution most people seem to favour is "bootstrap more stuff on it and it'll be fine".

    RichyFlag.gifsig.gif
  • MKRMKR Registered User regular
    Hm, what the issues CERN is encountering with the LHC seem like a decent metric for where we have performance issues with the internet. LHC is expected to generate something on the order of 1000 terabytes of data per year - all of which pretty much needs to be distributed throughout the EU, US and Japan/Oceania.

    A presentation I found pretty interesting was one which included the statistic that over a 10GBps fiber link, the loss of just one packet effectively reduced the bandwidth 25% for the next 5 hours - I don't know on what mechanism this metric is based.

    I definitely want to see that article. TCP/IP is built to be obscenely fault tolerant. I don't see how one packet could do that (again, limited experience :P).

  • electricitylikesmeelectricitylikesme Registered User regular
    Richy wrote: »
    Shoggoth wrote: »
    I don't understand enough about the actual workings of the technology but Isn't too much of the economy invested in the internet to just "start over"?
    That's part of the problem. We have a large and growing part of our economy based on what is essentially a prototype built from whatever people had available at the time and which got way out of hand. And the solution most people seem to favour is "bootstrap more stuff on it and it'll be fine".
    But that's just the point - the internet is only formalized as a set of protocols out the necessity of interfacing multiple disparent systems. It always has been designed towards robustness rather then throughput, and I'm not entirely convinced you can argue that literally the global network should be anything but designed as such.

    I mean, one of the flaws of IPv6 is simply that it specifies that EVERY device will have an address at the global level on the internet, which isn't exactly conducive to good security.

    Dis' wrote: »
    Cancer is when cells stop letting the body mooch off their hard work - clearly a community of like-minded cells should isolate themselves and do the best job each can do, even if the rest of the body collapses!
  • ShoggothShoggoth Registered User
    Richy wrote: »
    Shoggoth wrote: »
    I don't understand enough about the actual workings of the technology but Isn't too much of the economy invested in the internet to just "start over"?
    That's part of the problem. We have a large and growing part of our economy based on what is essentially a prototype built from whatever people had available at the time and which got way out of hand. And the solution most people seem to favour is "bootstrap more stuff on it and it'll be fine".

    But is any alternative actually viable? It doesn't seem so. I'm confused and hopefully some one can clarify. What exactly is the reason for "starting over"? Beyond concerns about security what is being demanded that requires a complete overhaul?

    I mean...

    It works for me. And with a few more "bootstrapings" the internet will do everything I could want it to.

    11tu0w1.jpg
  • YarYar Registered User regular
    Feral wrote: »
    Yar - you can blame NAT for that.
    I guess I could "blame" NAT, or I could credit it with being the single most important security and convenience feature in internal networking. But yeah, if it weren't for NAT, we would have to have gone to IPv6 a while ago. But I'll take IPv4 + NAT over IPv6 without it.

    Anyway, the main issue is that every person, every business, everywhere, will have to upgrade everything. The cost would be almost incaculable. Too many people will claim that their Internet works fine and they don't want to buy new network equipment and download new software, and then their will be two Internets, and so forth.

  • SentrySentry Registered User regular
    Yeah, I have to say, the moment it even looks like they will be close to accomplishing anything, corporations will be there to make sure everyone gets fucked right and proper.

    [SIGPIC][/SIGPIC]
    wrote:
    When I was a little kid, I always pretended I was the hero,' Skip said.
    'Fuck yeah, me too. What little kid ever pretended to be part of the lynch-mob?'
  • YarYar Registered User regular
    You guys who keep claiming "idiot corporations" or "boneheaded people in charge" need to grow up. Unless you've had first-hand experience with how executives run corporations, you don't have any clue what you're talking about.

  • CaswynbenCaswynben Registered User
    MKR wrote: »
    Hm, what the issues CERN is encountering with the LHC seem like a decent metric for where we have performance issues with the internet. LHC is expected to generate something on the order of 1000 terabytes of data per year - all of which pretty much needs to be distributed throughout the EU, US and Japan/Oceania.

    A presentation I found pretty interesting was one which included the statistic that over a 10GBps fiber link, the loss of just one packet effectively reduced the bandwidth 25% for the next 5 hours - I don't know on what mechanism this metric is based.

    I definitely want to see that article. TCP/IP is built to be obscenely fault tolerant. I don't see how one packet could do that (again, limited experience :P).
    With TCP/IP, you have a certain number of packets that you can have out at any given time. This is called the window, and it basically prevents me from flooding the network with packets. I have more and more pakcets out there that I haven't gotten acknowledgements from the person I am connected to. The general metric for having 'too many packets' out there is if one is lost. The way TCP/IP works in this regard is that it starts out with a small Window size, which gradually gets bigger (by a factor of 1 or 1/WindowSize, depending on some factors). If even one packet is lost, you drastically reduce the size of your window and start over again.
    What this does is keep the network from being too congested so more people can use it. However, individual connections are severely hampered by it. In the CERN case, you probably have a window that is absolutely gigantic, which gets cut to basically nothing after the lost packet, and then it takes some time to ramp it back up.

  • electricitylikesmeelectricitylikesme Registered User regular
    That sounds like the gist of what I was reading.

    Dis' wrote: »
    Cancer is when cells stop letting the body mooch off their hard work - clearly a community of like-minded cells should isolate themselves and do the best job each can do, even if the rest of the body collapses!
  • ObsObs __BANNED USERS regular
    I don't know shit about the internets working besides 7 layer OSI

    but exactly how the hell is someone supposed to even shut down the entire internet? Stop all the ISPs?

    spacer.png
    spacer.png
    Obs.gif
  • ObsObs __BANNED USERS regular
    shitty looking faces.

    Too many characters.

    spacer.png
    spacer.png
    Obs.gif
  • MKRMKR Registered User regular
    Caswynben wrote: »
    MKR wrote: »
    Hm, what the issues CERN is encountering with the LHC seem like a decent metric for where we have performance issues with the internet. LHC is expected to generate something on the order of 1000 terabytes of data per year - all of which pretty much needs to be distributed throughout the EU, US and Japan/Oceania.

    A presentation I found pretty interesting was one which included the statistic that over a 10GBps fiber link, the loss of just one packet effectively reduced the bandwidth 25% for the next 5 hours - I don't know on what mechanism this metric is based.

    I definitely want to see that article. TCP/IP is built to be obscenely fault tolerant. I don't see how one packet could do that (again, limited experience :P).
    With TCP/IP, you have a certain number of packets that you can have out at any given time. This is called the window, and it basically prevents me from flooding the network with packets. I have more and more pakcets out there that I haven't gotten acknowledgements from the person I am connected to. The general metric for having 'too many packets' out there is if one is lost. The way TCP/IP works in this regard is that it starts out with a small Window size, which gradually gets bigger (by a factor of 1 or 1/WindowSize, depending on some factors). If even one packet is lost, you drastically reduce the size of your window and start over again.
    What this does is keep the network from being too congested so more people can use it. However, individual connections are severely hampered by it. In the CERN case, you probably have a window that is absolutely gigantic, which gets cut to basically nothing after the lost packet, and then it takes some time to ramp it back up.

    Oh, I know what all that is.

    I was wondering how one dropped packet could cause a 25% drop in bandwidth for 5 hours. Since No specifics were provided, I assumed he referred to a normal packet. I wanted to see this study so I could read it myself (I like internet related studies). Even with one massive packet, it's hard to believe one packet could cause that kind of drop in bandwidth for so long, unless they were testing some kind of DoS attack with a malformed packet.

  • JohannenJohannen Registered User
    What would the costs be to do something like this? Would it make one countries net system incompatible with anothers? Is it even plausible? (By that I mean redoing the whole system physically and financially)

  • MKRMKR Registered User regular
    Johannen wrote: »
    What would the costs be to do something like this? Would it make one countries net system incompatible with anothers? Is it even plausible? (By that I mean redoing the whole system physically and financially)

    #1: You need to get everyone to agree on new standards
    #2: ISPs have to update their routers to use the new technology
    #3: Users have to update routers and their operating systems
    #4: Developers have to update their programs. Legacy programs that aren't supported will need to be provided a way to access the new internet, similar to how old programs with no security can be tunneled through SSH.

    1 and 2 are relatively painless. Three takes the longest because most users don't care, or you will have trouble convincing them to take the time and spend the money to do this. Four isn't too hard, but developers can be very stubborn, especially ones in corporations.

    For the most part, existing infrastructure is sufficient for a massive change. The hardest part is getting people to go along with it.

    Edit: Missed some steps:
    First you need to get funding to develop the standards, then you need to develop them. Then you can move to the first step in the list.

  • ShurakaiShurakai Registered User
    I am wary of a redesign because any change at this point will result in less freedom and flexibility for the end user. Also, I don't see the need for much faster speeds than we have now at the highest tier. Unless we suddenly have 100 terabyte something-other-than-a-hardrive(s), I don't really need the capability of downloading a gigabyte in less than 5-10 minutes. Of course in the far future, who knows, we might invent a new form of media (holographics? VR? Ultra-HD?) that is 50 gigs a pop but not any time soon (besides a Blu-ray or HD-DVD image I guess).

    That reminds me, bittorrent better damb well work in this "new internet" if they ever get around to doing it. Otherwise there will truly be parallel Networks -- Bittorrent and Everything Else.

    9d930f05c30d1172.png
  • MKRMKR Registered User regular
    Shurakai wrote: »
    I am wary of a redesign because any change at this point will result in less freedom and flexibility for the end user. Also, I don't see the need for much faster speeds than we have now at the highest tier. Unless we suddenly have 100 terabyte something-other-than-a-hardrive(s), I don't really need the capability of downloading a gigabyte in less than 5-10 minutes. Of course in the far future, who knows, we might invent a new form of media (holographics? VR? Ultra-HD?) that is 50 gigs a pop but not any time soon (besides a Blu-ray or HD-DVD image I guess).

    That reminds me, bittorrent better damb well work in this "new internet" if they ever get around to doing it. Otherwise there will truly be parallel Networks -- Bittorrent and Everything Else.

    To replace the internet, they would at least need to match its flexibility and openness, or it would never catch on. That means allowing people to connect stuff to it that isn't approved by some certification body. Bittorrent is safe. :P

  • Rhesus PositiveRhesus Positive Damn these electric sex pants! Registered User regular
    Most of what I know about how the Internet works is from this thread, so what I want to know is: where will all of the information go? Will it just be stored on external media until everything's operational, or something?

    I know that this question is probably as dumb as the "Where is the Internet stored?" thread, but it seems like a big deal to me - as well as the wiretapping stuff.

    robothero wrote: »
    damn rhesus, you're like a cyclical procedure of poor decisions
    PSNID: RhesusPositive
    I'm doing Movember for Men's Health! Donate if you can - thanks.
  • MKRMKR Registered User regular
    Most of what I know about how the Internet works is from this thread, so what I want to know is: where will all of the information go? Will it just be stored on external media until everything's operational, or something?

    I know that this question is probably as dumb as the "Where is the Internet stored?" thread, but it seems like a big deal to me - as well as the wiretapping stuff.

    The "internet" is just a term for the wires, devices, and protocols used to move information from one point to the other. The information will stay right where it has since someone first came up with the idea of linking two computers together (on disks).

    It's no different from a home network fundamentally. It's just bigger, and has a wider variety of stored information, and more ways of moving it around.

  • Rhesus PositiveRhesus Positive Damn these electric sex pants! Registered User regular
    Cool - thanks. So everything would be where we left it once all the updates were done? In that case, I would be for it with the caveat that it was done like when Britain changed to North Sea gas - a massive operation ensuring blanket compatibility.

    robothero wrote: »
    damn rhesus, you're like a cyclical procedure of poor decisions
    PSNID: RhesusPositive
    I'm doing Movember for Men's Health! Donate if you can - thanks.
  • MKRMKR Registered User regular
    Cool - thanks. So everything would be where we left it once all the updates were done? In that case, I would be for it with the caveat that it was done like when Britain changed to North Sea gas - a massive operation ensuring blanket compatibility.

    The only incompatibility I can think of is the storage of e-mails, which is usually done in an application-specific format. Hopefully some standard format for storing e-mail would be adopted, and all the major applications would provide a converter.

  • GlyphGlyph Registered User regular
    No longer constrained by slow connections and computer processors and high costs for storage, researchers say the time has come to rethink the Internet's underlying architecture, a move that could mean replacing networking equipment and rewriting software on computers to better channel future traffic over the existing pipes.

    Deja vu.

    MyBannerII-a.jpg
  • FeralFeral Who needs a medical license when you've got style? Registered User regular
    Most of what I know about how the Internet works is from this thread, so what I want to know is: where will all of the information go? Will it just be stored on external media until everything's operational, or something?

    Nobody's talking about taking the entire Internet down and replacing it all at once. They're talking about guiding the development of new technologies and replacing parts of it over time, with the understanding that the eventual result will be something that looks very unlike the Internet of today.

    I am comforted by Richard Dawkins’ theory of memes. Those are mental units: thoughts, ideas, gestures, notions, songs, beliefs, rhymes, ideals, teachings, sayings, phrases, clichés that move from mind to mind as genes move from body to body. After a lifetime of writing, teaching, broadcasting and telling too many jokes, I will leave behind more memes than many. They will all also eventually die, but so it goes. - Roger Ebert, I Do Not Fear Death
Sign In or Register to comment.