Justin Chapweske wrote: > There are already use cases in live streaming that multicast will not > be able to handle well, such as DVR/trick play, access control, and > auditing. I think the window of opportunity for multicast doing any > sort of heavy lifting of media is gone. > Access control and auditing are easy, assuming you need it to be "good enough" just like all DRM/content protection schemes are. DVR and trick play are easy if you do all the storage and work on the client, harder if you want to pick up media for late joiners. Of course peer-to-peer networks can help a whole lot for that case.
If ubiquitous and cheap source-specific multicast existed, there'd be a whole lot of "long tail" broadcasters who'd use it to get their content out, and who wouldn't care about these issues, just like the operators of pirate radio stations care more about listeners than accounting details, but this is unlikely to be deployed widely for a variety of reasons. Peer-to-peer-based "application-level" multicast moves the "who does the work of redistribution" from "the IT and ISP guys" (who don't really care to be bothered with special-casing your ability to view things) to "the viewer" (who cares a whole lot). Even if it is less efficient for the network, it is the only way to align the economics. (Except in things like corporate and education networks, where the network operator's job is to get the in-house content to the in-house viewers, and so they have an incentive to do it in a way that loads the network as lightly as possible while doing so). Matthew Kaufman _______________________________________________ p2p-hackers mailing list [email protected] http://lists.zooko.com/mailman/listinfo/p2p-hackers
