Bittorrent declares war on VoIP, gamers | Hacker News Reader

Bittorrent declares war on VoIP, gamers

theregister.co.uk

17 pointsjmackinn9 comments

The author here comes across as too biased and passionate to be taken seriously. I understand that he wants to hype up his story to draw in readers but his "The sky is falling" attitude obscures any of the real issues and considerations involved.

Bittorrent, legal issues aside, is being throttled by many ISPs. It is only natural for bittorrent to fight back. There will be casualties, but I can't see ISPs throttling all UDP traffic, breaking VoIP. The internet won't break in some “Congestion Collapse”.

I also disagree that P2P applications would be indistinguishable from VoIP without deep packet inspection as the author suggested. The network usage statistics for VoIP and P2P must look quite different and should be easily distinguished.

Seriously. Its a 5 liner in a tc script on any linux box to make udp streams behave nice and share the bandwidth without driving up each others latency.

I think ISP's can figure it out. The fact that many isp's are also telcos presents a far greater threat to voip than bittorrent, tcp, upd or otherwise.

"The author here comes across as too biased and passionate to be taken seriously."

I find this the problem with a lot of the Register's articles. The reg can be humorous, and informative, but it often becomes too zealous to be taken seriously.

Gaming and VoIP both tend to be very low bandwidth. Seems like it'd be very easy to throttle all UDP traffic down without hurting either of them. Latency is way more important for them than bandwidth.

Followup: http://news.ycombinator.com/item?id=382152

Why the sky is in fact NOT falling: http://forum.utorrent.com/viewtopic.php?pid=379206#p379206

Flagged as FUD.

Seriously, read the other links... the whole thing is just flamebait.

Isn't UDP, by definition, unstable packets that could be dropped at anytime?

Any packet can be dropped at any time, for TCP or UDP. So the protocol stack or application needs to deal with it and re-send, using a backoff policy.

Its worth looking at IM2000 and the pertinent questions that D. J. Bernstein asks about a future mail infrastructure.

http://cr.yp.to/im2000.html

Reply on news.ycombinator.com