5 ms·
I really like Bram, but what is revolutionary here? I was pretty heavily involved in p2p field back in the early '00 and I've read extensively on the subject.
by huhtenberg 14y ago
I really like Bram, but what is revolutionary here?
I was pretty heavily involved in p2p field back in the early '00 and I've read extensively on the subject. Even back then there was plenty of research into peer-casting, including peer clustering by proximity metrics. The idea is bloody obvious, and it all inevitably boils down to constructing an efficient and resilient overlay networks, which is a very well researched domain. The reason there's not much of it implemented is because there is always a simpler (read - dumber) solutions that worked just as well in practice. Think YouTube vs. Joost.
Can anyone with a more recent exposure to p2p stuff comment on whether this is indeed an innovation or is it just a PR spin on a patent application?
- hack_edu 14y agoThe patent is about breaking peers out into small sub-swarms to look after each other. The more healthy peers in each group focusing to keep their's groups healthy while also keeping up with the larger swarm. The greater BitTorrent protocol doesn't do this and, judging by the patent filing, other p2p streaming applications don't either, hence why they fall down upon high demand.
- synctext 14y ago> other p2p streaming applications don't either,hence why they fall down upon high demand. This patent does not claim anything novel. This idea of "clubs of peers" or groups has been around for a while. This well known paper from Torino, Italy, written a few years ago: "overlays are maintained by peers organized in clusters that represent sets of collaborating peers", http://dx.doi.org/10.1109/TMM.2010.2077623 http://dx.doi.org/10.1109/TMM.2010.2077623 Disclaimer: my research team is part of the open IETF Internet Standard on P2P streaming which is directly competing with this patented technology, http://datatracker.ietf.org/doc/draft-ietf-ppsp-peer-protocol/ http://datatracker.ietf.org/doc/draft-ietf-ppsp-peer-protoco... Anybody else has an even older citation of prior art?
- bramcohen 14y agoThe traditional overlay approach involves a bunch of full trees, and uses multiple trees as a way around dealing with leaf nodes being unutilized. My approach uses multiple groupings, which do not overlay, screams within them, and does something completely different for the last hop. They're completely different architectures. I have trouble taking seriously any paper which says that it makes heavy use of multiple description coding. If you have congestion control, skips should be an extreme and bad event.
- mammalfriend 14y agoThere have been several p2p low-latency streaming companies in the past with swarming, some of which where acqui-killed by CDN companies like Akamai. But since I haven't read the app, I won't go as far as to critique the filing. I think the big challenge for this is that traditional bandwidth is becoming cheap, and anyone who cares about QOS can pay for it and/or build out their own distribution network.
- hack_edu 14y agoCurious what companies these were? Haven't heard of any of this.
- bramcohen 14y agoThose other technologies have latencies measured in dozens of seconds or minutes, while mine is measured in seconds.
- malandrew 14y ago"... while *mine* is ..." No one else at BitTorrent was involved in coming up with this solution?
- malandrew 14y agoI don't know why this got downvoted. It was a legitimate question. I've been involved in patent filings and unless you are a one-man operation, there are always other people that contribute to the key insights needed to come up with a novel, useful invention.
- Confusion 14y agoPatents aren't just intended for revolutionary things. They are also intended for mundane, but actually useful, applications of those things.