Slashdot Mirror
Web Creators Call Internet Outdated
ElvaWSJ writes "Several networking pioneers are dissatisfied with the Internet's underpinnings, and some are offering remedies to ease the strain that bandwidth-hungry services put on technology networks. Along with other projects here in the US and around the world, numerous companies and organizations are looking to rewrite the underpinnings of the internet. This piece looks at new concerns from old hands at networking, with comments from folks like Larry Roberts and Len Bosack. 'Mr. Roberts's concern over the Internet's infrastructure stretches back years. Even while at ARPAnet, he says he was unsure how long the technology could work, especially since the system didn't ensure that information packets would arrive at their destination. His fears crystallized in the late 1990s when he saw companies begin to use the Internet to make phone calls and consumers begin to dabble in online video.'"
The Internet is dead.
Seriously, there really isn't anything that wrong with the Internet. Sure, it may not work perfectly, but how can you ever expect to connect so many diverse systems together in one unregulated mass and have it work perfectly? If you want a better system, go use Internet2 and leave the rest of us alone.
www.timcoleman.com is a total waste of your time. Never go there.
From TFA
To tackle the problem, a slew of start-ups are producing gear and software to accelerate Internet traffic or to increase the network's capacity. These include companies run by Messrs. Roberts and Bosack, as well as Riverbed Technology Inc. and Big Band Networks Inc. Other companies, such as BitGravity Inc. and Limelight Networks Inc., are creating "parallel networks" -- essentially scaled-down versions of the Internet -- to escape the glut of traffic on current networks.
Of course, the gentlemen crying wolf are the same people who run companies who can sell you stuff to fix the problem. There's no new problem here. The tubes, according to business people, always seem to be in a sorry state, about ready to crumble the moment the wrong person clicks one more time on that link promising Brittney Spears porn. And yet, I have been able to get my email every morning since 1993 when I got my first email account.
Typical fearmongering article designed to drum up new business. Mod me up, give me my karma now, and move along, nothing to see here.
"All great wisdom is contained in .signature files"
The solution according to Roberts:
...
"Last month, his start-up, Anagran Inc., introduced a piece of gear called the flow router that he says can help modernize the Internet. The equipment analyzes Web traffic to discern whether it is an email, a movie or a phone call and then carves out the bandwidth needed for transmission."
No thanks.
The solution according to Bosack:
"Last month, his company, XKL LLC, unveiled a system that allows businesses to connect to underground cables that have nearly 100 times the capacity of current telecommunications pipes."
That would be really nice, how about making use of all the dark fiber first.
All in all, we see the people who were involved in the creation of the Internet now got into the private business and use all possible means of pushing said business forward. It's almost sad they did so good job the first time, that now they have created solutions in search of a problem
Even while at ARPAnet, he says he was unsure how long the technology could work, especially since the system didn't ensure that information packets would arrive at their destination.
So long as you're running packets over copper, or fiber, or radio waves, or any other physical medium, you're going to have the possibility of packet loss. Oops, I unplugged the cable.
I always thought that was the brilliance of IP: once you admit that packets will always be unreliable, you can build a platform on top of that which does what you want. Pretending it can be 100% reliable is a fantasy, and it doesn't help us build better networks.
The web is the same way: no database geek would have ever thought of throwing referential integrity out the window. But Tim realized that there would always be the possibility of not being able to connect, so we have the 404 page, and the web is flourishing.
If Larry has an idea for a way to guarantee packets arrive, that's great, but somehow I doubt it's physically possible. And as long as we don't have it, the best way we know how to build networks is to allow for the possibility of failure, and deal with it.
Even web clients are smart enough to say "Sorry, can't seem to connect to some-server.com right now", but if cable TV goes out all I get is a blank screen. And if my network starts to get flaky, I can pause an online video and come back later when it's fully downloaded; I can't do that on TV. Is online video really that bad? On everything except bandwidth, we're doing pretty darned good, and bandwidth is being solved as we speak.
I'll let the "average person" part pass (since I just don't know any better), but the megabits in question are at the last mile--there's a bottleneck where all those last mile circuits feed into an uplink that doesn't have nearly as much bandwidth as they do in the aggregate. And that's by design; the whole point of a communications network like the telephone network or a packet-switched network is to make the most of a limited resource by sharing it.
If you want everybody to be able to watch the same exact thing at the exact same time, all the time, you run a wire from a central station to everybody interested (or even easier, you transmit a radio signal). The bandwidth requirement is just whatever your broadcast requires. However, if you want everybody to be able to watch a completely different thing at the same time, all the time, and that thing could be provided by anybody else in the network, then everybody needs to be able to get full bandwidth to any other site all the time. And for that you have to build the equivalent of a wire running from each subscriber to every other one. What's the factorial of 300 million?
The only way to get it to work is to weaken the assumptions; e.g., sometimes you won't be able to watch TV over the internet because too many people are already doing so. Or the number of choices you actually have is small, relative to the number of subscribers (which allows for some tricks to share bandwidth). Or you can't really watch whatever you want whenever you want.
Yes, on a set of on-the-air channels shared with whatever other subscribers may be in the vicinity.
Gee, if only we had some method to control the transportation of packets. I envision it starting with something like a handshake between two hosts so each would know that the other was ready. Then you'd want to assign sequence numbers to each packet so the recipient would know if a packet had been dropped. The recipient might have some way to acknowledge each packet, so the sender knows that the recipient received it. And there might even be some way for either endpoint to tell the other that it's finished with the conversation, allowing timely cleanup of network resources.
Nah, I'm dreaming. If such a magic "transport control" protocol were possible, surely the inventors of the Internet would have figured it out by now.
Chelloveck
I give up on debugging. From now on, SIGSEGV is a feature.