For sure, that makes sense, you've said it well. What I'm wondering is: the content provider's ISP isn't losing money here? If not, then I'm real surprised that multicasting hasn't been more widely adopted.
I think we'll see more application level multicasting if it doesn't become more supported at the IP level.
Thanks for the information:)
Aha, I see what you're saying. My comment was hinting at a situation like this:
Currently company X is streaming some content to 1,000 users. If X were to multicast that content out, it might only get sent to 4 routers, which would in turn multicast it out to more routers, and so on. So X's ISP would be a little unhappy that they X's outgoing bandwidth is reduced.
Have I missed your point? I'm interested, so fill me in.
I agree, I think that for the vast majority of organizations, NAT takes care of the address space problem. Given that a single IP can have 64k ports, that's 64k mappings to internal machines. So a company with 640 machines can have 100 simultaneous connections per machine.
This is true to an extent. IPv4 routers could support multicasting if they wanted to. I think the bigger issue with support for multicasting is that it breaks down conventional ISP economics of charging for bandwidth.
One way they can switch without significant down time is to roll out the changes over time. Essentially they have two options:
Dual stack: routers that support both IPv4 and IPv6. The routers speak v4 to v4 routers, and v6 to other routers.
Encapsulation: routers can encapsulate IPv6 packets in IPv4 packets and then tunnel the encapsulated packet to other IPv6 routers via IPv4 routers.
If you're interested in having these performances hosted at the Internet Archive, contact etree at archive dot org or post to the message board on the website.
this is exactly what the archive is doing. the link provided above is to project gutenberg, which is available as a collection from within the archive. they have other collections as well including a cool open source one where people can contribute public domain books already digitized.
http://www.archive.org/texts/texts.php
Getting that stuff would be pretty interesting. And old Usenet as well. Does anyone know of any such archives?
To see those old BBS screens again would be a trip...
Slashdot Mirror
For sure, that makes sense, you've said it well. What I'm wondering is: the content provider's ISP isn't losing money here? If not, then I'm real surprised that multicasting hasn't been more widely adopted. I think we'll see more application level multicasting if it doesn't become more supported at the IP level. Thanks for the information :)
Aha, I see what you're saying. My comment was hinting at a situation like this: Currently company X is streaming some content to 1,000 users. If X were to multicast that content out, it might only get sent to 4 routers, which would in turn multicast it out to more routers, and so on. So X's ISP would be a little unhappy that they X's outgoing bandwidth is reduced. Have I missed your point? I'm interested, so fill me in.
I agree, I think that for the vast majority of organizations, NAT takes care of the address space problem. Given that a single IP can have 64k ports, that's 64k mappings to internal machines. So a company with 640 machines can have 100 simultaneous connections per machine.
This is true to an extent. IPv4 routers could support multicasting if they wanted to. I think the bigger issue with support for multicasting is that it breaks down conventional ISP economics of charging for bandwidth.
One way they can switch without significant down time is to roll out the changes over time. Essentially they have two options: Dual stack: routers that support both IPv4 and IPv6. The routers speak v4 to v4 routers, and v6 to other routers. Encapsulation: routers can encapsulate IPv6 packets in IPv4 packets and then tunnel the encapsulated packet to other IPv6 routers via IPv4 routers.
If you're interested in having these performances hosted at the Internet Archive, contact etree at archive dot org or post to the message board on the website.
And I forgot to mention that on several of their servers they use P2P options including Onion Networks (Open Content Networks) and BitTorrent.
And the Internet Archive has eight servers of SHNs, featuring over 2,200 concerts!
this is exactly what the archive is doing. the link provided above is to project gutenberg, which is available as a collection from within the archive. they have other collections as well including a cool open source one where people can contribute public domain books already digitized. http://www.archive.org/texts/texts.php
Getting that stuff would be pretty interesting. And old Usenet as well. Does anyone know of any such archives? To see those old BBS screens again would be a trip...