Slashdot Mirror
Competing (Commercial) Visions For The Internet Future
Stirland writes: "This
article in today's NYTimes says that AOL's new
plan focussed on creating content for broadband could
have cable companies over a barrel.
It tries to compare programming on cable to 'programming' on the Internet.
It's an important article to read because it gives us an idea of what cable companies' potential plans are for the broadband Internet. Given that they're not regulated like DSL, and they're the fastest growing providers of broadband Internet access, this has profound implications for the next generation of the Internet.
This article omits the fact that Excite@Home tried this 'programming' approach on broadband. It failed.
Another reason this article is important: Contrast AOL's approach described here with Amazon.com and Microsoft's .Net strategy. These are two polar opposite visions of the way the Internet will develop. The media vision vs. computing vision. The interesting story here is that it isn't that one is 'open' and the other 'closed.' They're just open and closed in different places -- places, obviously, that suit the companies' strategies. Why should you care, and what's in it for you? These competing visions are currently duking it out at the FCC under open-access proceedings. So the future of the Internet is hanging in the balance."
Here's a mirror of the text of this website. Login/password haters of the world unite!
http://members.cox.net/infornography/nyt.txt
Earthlink, which bills itself as the number two ISP in the US behind AOL, does not restrict your use of servers on their home DSL lines, though they obviously won't supply you with any tech support for them.
But then, Earthlink does offer various anti-spam services as well as a new Popup blocker software available free to members, and while they provide dial-up software similiar to AOL's, you're just as free to not install or use the "simplified net interface" and just use the connection on it's own.
For one of the big guns of ISPs, they tend to be one of the better deals, at least if you're not in certain areas and in need of detailed tech support.
The internet is not meant to be a broadcast medium, nor is it very good as one. Ask online radio stations that not only must now pay high license fees, but also must buy lost of servers and bandwidth to stream audio to a few listeners, nothing like traditional broacast mediums (radio, tv, even broadcast cable and satelite) which scale much better.
That's absolutely true, if you look at unicast streaming methods. But there's definately technologies out there to get around that. Multicast is an excellent way to get over that nasty bottleneck that expensive bandwidth creates. The main problem with multicast? Not everyone has it. Its one requirement is that every device in the path be multicast enabled.
Check out this FAQ for a starter on multicast. Read up on PIM-SM, the dominant multicast protocol, and the future of multicast which is SSM. These are protocols that are designed specifically for one-to-many applications, which is ideal for things like audio and video streaming. Unfortunately, the only major OS with built-in SSM support is Windows XP. There's patches out there for specific Linux and FreeBSD kernels to add the necessary IGMP v3 support, but you won't see it in the main builds. Why? I wish I knew.
If your cable modem is DOCSIS 1.1 compliant, then it's capable of multicast. But most ISPs don't want to enable multicast. A lot of the time, they've never even heard of it, even though it's been around since the mid-80s. It's a requirement for IPv6, but Juniper and Cisco routers don't support it yet. I definately haven't seen any IPv6 multicast enabled applications.
Multicast is out there, and it's exactly the type of communication model that we need in order to scale audio and video streaming applications on the Internet. On 9/11, an audience of 2000 was watching CNN Headline News over a multicast feed from the University of Chicago. It was a single 300 kbps video stream that never ran into the issue of a bandwidth bottleneck that CNN's own website had. And quite frankly, that audience could have grown to over a million, and the University of Chicago's server never would have known it. It still would have been sending out a single 300 kbps stream, and still reached all those people.