Jump to content

Bittorrent declares war on VoIP, gamers By Richard Bennett


moogly

Recommended Posts

From the horse's mouth

http://forum.utorrent.com/viewtopic.php?pid=379206#p379206

Tech blogs

http://torrentfreak.com/will-utorrent-really-kill-the-internet-081201/

http://www.networkperformancedaily.com/2008/12/bittorrent_over_udp_end_of_the.html

http://jmkupferman.blogspot.com/2008/12/bittorrent-using-udp.html

http://arstechnica.com/news.ars/post/20081201-utorrents-switch-to-udp-and-why-the-sky-isnt-falling.html

http://www.slyck.com/story1804_BitTorrent_Switching_to_UDP_Transfers_Nonsense

My response

Already addressed, really. Firon said it before, but it bears repeating: if he were serious about what he was saying, he would have actually asked BitTorrent directly for more details. Your average tech bloggers were easily able to get in contact with Simon -- why couldn't Mr. Bennett be bothered to do the same and get his facts straight? Sensationalism -- or irresponsible journalism at best.

Link to comment
Share on other sites

The thing that caught my attention most was his claim that uTP came about as a response to ISP throttling. Where does he come off saying that without knowing how long uTP has been in the works? My IRC logs show the devs talking about it since as early as July 2007 (and I'm certain it was discussed before that, while I was offline). Something similar in essence to the current form of uTP was already being worked on by ludde (meaning it was in the works already by somewhere in 2006). This was certainly before Bell started throttling their users. Baseless claim indeed. (And before anyone argues that Mr. Bennett said that BitTorrent set uTP to default to counter mounting ISP throttling, I'd like them to explain why the devs would waste their time on such a large undertaking if it were to be relegated to being just a side-feature)

Then the other stuff about congestion control, well, I'm less familiar with (not being a networking expert of any sort), but that he could claim that uTP would kill the Internet because it's discarding TCP's native congestion control is... silly (euphemism). Clearly, he didn't know the details behind uTP if that's his claim (or he's simply misleading his readers). Either way, reporting it the way he did is (IMHO) rather disingenuous.

Link to comment
Share on other sites

Ultima asks the relevant question: "Where does he come off saying that without knowing how long uTP has been in the works?"

Actually, I got it from remarks on this site and from DSL Reports (see: http://www.dslreports.com/shownews/New-UDP-uTorrent-Takes-Aim-At-Throttling-99366 ) where uTorrent users themselves were saying the default transport was changed to help them out with Bell Canada: "They're implementing their new protocol on top of UDP because it lets them bypass TCP's congestion control and retransmission mechanisms. It gets around throttling because it doesn't fit Bell's profiles."

It turns out there's quite a bit of history to the use of UDP for torrents, but it tends to be entwined with management avoidance techniques such as header obfuscation or hidden behind the secret "BitTorrent DNA" scheme that's been mentioned but not disclosed for a long time.

So it's not unreasonable to surmise that the effects of a change like this are linked to the intentions for it. That was a wrong surmise in this case, as BitTorrent, Inc. is trying to go straight and leave its piracy-enabling past behind. But we're all dogged by our histories.

In this case, a simple press release from BitTorrent, Inc., before the alpha was released would have prevented any confusion about their intentions.

I've written a follow-up article which you'll find interesting. It's not a love-letter to BitTorrent, but it does strike a more moderate pose.

Link to comment
Share on other sites

Yeah, I actually saw your blog a few hours earlier, but never got around to re-replying in this thread. I can't say whether I agree or disagree on the deeply technical details, since I'm no networking expert (as I mentioned earlier :P), but after a bit more reading around, I can at least appreciate your point of view. I still trust that the devs and engineers working on uTP understand (most of?) the consequences behind what they're doing, but time has a funny way of letting objectivity seep in :)

For anyone interested (and I hope you don't mind my posting it), Mr. Bennett's blog can be found here.

Link to comment
Share on other sites

By all means, Ultima, all links is good links. One more reference to the effects of 1.9 on management: "Wow this is awesome. Prevents my sandvine using isp from throttling.. full u/l speed to fellow [utp] peers. Ain't seen that in over a year. (i noticed bittorrent 6.1.1 clients w/ utp too.)"

That was posted on on this forum, by jizam.

Link to comment
Share on other sites

Indeed, there is no denying that it does thwart TCP RST attacks. It's only natural (considering how UDP is connectionless), and I've even acknowledged that fact on the very first page of that thread you linked to. I guess the main point of the first paragraph in my above rant was to say that uTP wasn't a reaction to ISP throttling, not to say that it doesn't affect ISP throttling. It's a very minor point for me to be picking at, but to say that the intent was to circumvent ISP throttling paints a image of uTP darker than if it weren't stated as such (well, really, the purpose behind uTP itself is in no way dark anyway, but whatever).

Link to comment
Share on other sites

An awful lot of the development work on Azureus, BitComet, and uTorrent over the years has been concerned with avoiding or defeating ISP management, so it's natural to assume, as many readers did before I waded into the discussion, that this was the prime objective of uTP. Every time an ISP rolls out a new management system it's assumed to be yet another example of the monopoly phone companies trying to bring back the days of the black telephone and all that. It's human nature to connect new events to historical patterns.

That doesn't mean it's right, of course, just that public opinion wasn't sitting on its hands waiting for me to shape it. This thread on DSL Reports was particularly illuminating: Notice, new uTorrent Alpha may be able to evade throttling

When this sort of discussion had been taking place here and at DSL Reports a week before my article, why was there so much excitement about what I wrote? All I did was break down the wall between the MSM and these discussions.

And one more thing, you ask why I didn't contact Simon before writing. That's a good question. I know Eric Klinker and would be more inclined to contact him than the marketing staff, because we engineers understand each other and the nuts and bolts of the implementations better than the help does. In this case, the guy who is best able to shed light on the things that interest me about uTP, Stas Shalunov, has been unavailable for discussion through this whole period from my writing last Sunday to today. I have talked to Eric and Simon and they've been helpful as far as they're able, but I'm a hardcore network protocol guy and I need to get my answers from the source.

Link to comment
Share on other sites

When this sort of discussion had been taking place here and at DSL Reports a week before my article, why was there so much excitement about what I wrote? All I did was break down the wall between the MSM and these discussions.

If you have to ask this, then you're underestimating the power of a sensationalistic title. Doomsday predictions rarely escape widespread attention, especially when they're in reference to something so widely recognized. It didn't have to be you -- it could have been anyone else writing an article in a similar fashion/tone as you did, and the reaction would have been the same. By breaking down the wall between mainstream media and those discussions in the way that you did, you can shape public opinion in a negative way.

A press release may indeed have been useful in helping people understand uTP, but frankly, the first post in the announcement thread was already quite suggestive of its purpose.

I wasn't really asking why you didn't ask Simon specifically, just why you didn't contact someone from BitTorrent Inc. who was in the know, before writing/publishing the article. Even if you couldn't get in contact with Mr. Shalunov, someone else would still be able to illuminate matters (as has clearly happened already). Simon is just one example.

Link to comment
Share on other sites

As far as I was concerned, the user reaction and the network effects were the story, not the corporate spin. The corporate intentions are much less important, in my assessment. But now that BitTorrent has made this is a big story, we'll have to keep it alive a while longer.

It's good publicity for the company in any event: I spelled their name right.

Link to comment
Share on other sites

There's a reason uTP wasn't put out in a press-released public-release cycle.

uTP has been in uT, albeit disabled by default for over a year. It's obvious from some of the reports in the announcement thread more testing is needed.

Bloggers are not journalists, but I appreciate your decision and welcome your foresight. I hope you stay around after the wave you created subsides :)

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...