..says Richard Bennett. 'Sensationist nonsense,' BitTorrent tells us...
Over the weekend, we noted
that uTorrent has released an alpha client that uses UDP for P2P traffic delivery. UDP makes up roughly 2% of all Internet traffic today, and generally isn't used for data delivery because it doesn't guarantee either the delivery or the quality of the data being delivered (nor is it quite as easily managed). While the migration to UDP could potentially make filtering of P2P more difficult -- raising the hackles of some anti-network-neutrality ministers and ISPs -- BitTorrent tells us the decision was aimed at actually making BitTorrent more friendly.
According to posts
at the uTorrent forum, the new version lays uTP, the micro transport protocol, on top of UDP, which provides for better flow control and prevents the kind of TCP RST packet attacks Comcast has used to throttle upstream P2P traffic. Robb Topolski, the DSLReports user who first discovered Comcast's packet forgery tactics, thinks the shift to uTP/UDP is a good one:
It's a very good thing for the network. This new protocol YIELDS to other streams. In other words, it's less aggressive. The idea, eventually, is that background file transfers are handled like -- well -- background transfers -- similar to the way that background processes take a lighter toll on the CPU while you're actively using the computer. P2P users have the same concerns -- this change keeps their interactive uses snappy, and during crunch time it ought to help others as well.
Yet Richard Bennett, perhaps the Internet's most vocal opponent of network neutrality, pens a piece over at The Register
proclaiming that the shift will result in an Internet meltdown -- and worries that network neutrality laws would prohibit ISPs from taking on this new throttling challenge. The author quietly states his case by suggesting that those who would support the use of UDP for P2P transfers (like apparently, BitTorrent creator Bram Cohen), are little more than selfish junkies, mindlessly braying for the looming apocalypse:
Some of the people who use this system are spoiled children with no more concern for the greater good than junkies looking for their next fix. They can’t be allowed to spoil it for the rest of us, and the only practical means to prevent their doing so is to unleash effective management upon them. . .The best way to ensure that uTP doesn’t kill the internet is to throttle it at the source, and any law that stands in the way of ISPs exercising that level of management is deadly to the internet.
Simon Morris, head of Product Management at BitTorrent, thinks Bennett is a little confused.
While it's true that a UDP-based P2P network will cause some problems for ISP throttling (something confirmed
by our users), that isn't the goal. "We’re doing this to implement our own more sensitive congestion control on top of UDP," he notes. "We felt that TCP’s congestion control was problematic in that it relies on looking for packet loss as an indication of congestion. TCP spots the problem only after it has occurred."
"By contrast, our prototype UDP-based protocol (called uTP) detects congestion by measuring transmission times between peers – if packet delivery *slows down* then we infer that congestion may be about to occur and immediately throttle back on delivery speed," he says. "The point here is a protocol that is more sensitive, NOT a protocol that is more greedy," insists Morris. "The idea we'd "declare war" is unfortunately sensationalist nonsense."
Yet according to Bennett, "even the downloading fiends who haunt the message boards at Broadband Reports" can see the use of UDP for P2P transfers will result in an Internet implosion. Any thoughts, haunting downloading fiends?Update
: The reviews are in!GigaOM
notes that when it comes to Richard Bennett, "a little scaremongering can go a long way to make the case for an ISP-based network management clampdown on P2P traffic." Torrent Freak
proclaims that a significant chunk of what Bennett's saying simply isn't true (but it does get the Register plenty of hits).
The Industry Standard
says Bennett offers broad assumptions with few references.
Writer Robert Hallock pens an interesting explanation of TCP vs. UDP in a counter piece
to Bennett, suggesting "we wait to attack the potential problem with data and evidence rather than suppose and conjecture our way into the unforgiving embrace of network non-neutrality."