Inline ... Kon Wilms wrote: > On Sun, 2005-05-22 at 21:28 -0400, Tom Barry wrote: > >>To really make this sort of "hyperdistribution" (neat word) work >>over the Internet I still think they should make a version of Bit >>Torrent that works well using multicasting. For popular shows >>this would still commercial allow appointment TV with vastly >>decreased bandwidth requirements. > > > This makes no sense. You propose using the internet to duplicate what > you can already do with a broadcast station, i.e. mass distribute > content to 1~n receivers -- without any of the associated multicast > routing problems along the way. > > Why not use the more likely candidate in a more ingenious way. > > You can't just say 'it will be faster if we stuff multicast in there'. > No, it won't. It may even be worse. Bittorrent's strength is in the fact > that clients have different blocks, which allows them to swarm content > amongst each other. Your data rate remains high because the file is the > network and as peers complete blocks the network topology for the file > in terms of bandwidth changes, so it is possible to constantly optimize > the download rate amongst peers. With multicasting, this is not the case > since all clients (should) get the same blocks, and arrival time (or > even arrival) is not guaranteed. You may have a network outage at a > point in the network and subsequently request a chunk from a neighbor. > The chunk starts transmitting and at the same time the outage is > resolved. What do you do now? How do you cope with packets not stomping > each other? Multicast routing protocols won't solve these issues for > this particular distribution model. > > Bittorrent does have those strengths. But I qualified my post e popular appointment TV, the intent being something being "broadcast" that has enough interest that many many consumers will be interested in it. The problem with BT is that for every byte received, one has to be sent by someone. It may be distributed by many someones but it is still one-to-one for that piece of the file. If you know there are VERY high odds of MANY people being interested in that piece than adding multicasting still makes sense. This is true whether you are sending from a single server or multiple peers. >>This wouldn't be full HDTV but should at least compete well with >>the current efforts from the broadcasters and would work already >>for many broadband customers. > > > Compete with them using what content? > > >>Of course all this assumes the cooperation of the prime content >>owners, which is maybe unlikely. Otherwise it will probably have > > > You answered your own question. > Of course, but it had to be said. ;-) > >>to wait a bit for better anonymous hyperdistribution technology >>and continue to use pirated media at first. > > > Better than what? You can already bittorrent anonymously using Tor or > other anonymous proxying and services. > Of course. But it is not clear whether Tor, I2P, etc. are yet either anonymous or efficient enough for "commercial pirate" use. Probably not. > Everyone screams 'lets bittorrent' like its 1997 and the dot bomb boom > is upon us, but why is it that no-one brings up the two problems all of > these unregulated distribution mechanisms share: > > 1. You *cannot* guarantee any arrival time for the content on the > 'network', so it makes the distribution mechanism useless for prime time > content unless the content owners start seeding themselves. Which > defeats the purpose. A cheap distribution mechanism for content owners? > Think again (and definitely not if the movie or show is sub-par)! > Getting hung up in guaranteed arrival time is the TV model. We have already seen how zillions of PVR users are allowing some slack in order to get other benefits. But this also allows variable bit rates, error resends, more efficient VBR coding, etc. > 2. You *cannot* guarantee any delivery bitrates or content availability > on the actual 'network' (even after you have seeded it), because the > delivery of content is based on the popularity and willingness of the > users to seed the network after they have completed their downloads. > This means that content owners need to provide hubs to keep the network > seeded. Which once again, defeats the purpose. Reruns? Forgedaboutit! > I think multicasted P2P2Tivo would be efficient and find many interested consumers even for many reruns. But maybe not for infomercials. Oh well. Internet TV has less need to fill all the air space. I'll stop here. - Tom > The original article is just another hype-fest that doesn't take into > account the technical implementation problems faced here. More kool-aid, > anyone? > > Cheers > Kon > > > > > ---------------------------------------------------------------------- > You can UNSUBSCRIBE from the OpenDTV list in two ways: > > - Using the UNSUBSCRIBE command in your user configuration settings at > FreeLists.org > > - By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word > unsubscribe in the subject line. > > ---------------------------------------------------------------------- You can UNSUBSCRIBE from the OpenDTV list in two ways: - Using the UNSUBSCRIBE command in your user configuration settings at FreeLists.org - By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word unsubscribe in the subject line.