Slashdot Mirror
Bittorrent To Cause Internet Meltdown
Gimble writes "Richard Bennett has an article at the Register claiming that a recent uTorrent decision to use UDP for file transfers to avoid ISP 'traffic management' restrictions will cause a meltdown of the internet reducing everybody's bandwidth to a quarter of their current value. Other folks have also expressed concern that this may not be the best thing for the internet."
Plz seed
Hello!
You can't blame your wife for stabbing you in the face when you keep locking her in the bedroom. When you threaten someone, their choice of retaliation may not be the smartest way to go about it, but then again, WHAT DID YOU THINK WAS GOING TO HAPPEN?
"When life gives you lemons, don't make lemonade. Make life take the lemons back!" -- Cave Johnson
Finally, I'll have a legitimate reason to slack off and not do my job...
On the other hand, how am I going to procastinate without the internet?
So what you're saying is that it may clog the tubes?
Someone get a plumber, quick!
If you're going to transfer files over UDP then you need to build some TCP-like protocol on top of it. The article doesn't say exactly how BT works in this respect, but he's probably right. There's no way that BT's protocol could be as sophisticated as TCP, given its 30+ years of development.
Most people don't appreciate how amazingly well TCP's flow control works in terms of maximizing link utilization in a way that is fair to all network users. We really don't need is an arms race of new, greedier protocols.
However, one thing to realize about P2P is that because there are often dozens of active TCP connections transmitting from one machine, fairness goes pretty much out the window anyway. An alternate protocol could conceivably improve on this by applying flow control to the aggregate throughput for the whole "bundle" of connections, rather than each connection individually. This would improve fairness and also increase efficiency because you wouldn't have a bunch of TCP streams individually trying to grow their windows, causing packet losses.
In the end this will be a good thing for the internet.
Forcing ISPs to treat all traffic the same, because they can't tell what is what, will be good for net neutrality.
You should get the bandwidth you pay for, regardless of what actually travels over it.
The terrorists, it's all their fault.
Love many, trust a few, do harm to none.
"They" said the same thing about once popular File Service Protocol (http://fsp.sourceforge.net/) way back when the net was young, pre-Napster, and before any massive internet infrastructure investment was made...
Comcast will just block UDP completely... duh it's not like they care if you can use 'your internets'
Informative? Try more like "troll" or "flamebait."
Using your stupid analogy, this would be more like threatening to raze the entire city to the ground because no one intervened to stopped the wife from being locked in the bedroom.
You know what I'd like to see happen? Anyone who is caught using uTorrent with this setting gets their broadband internet access contract torn up. Don't even pretend that most bit torrent traffic is legitimate and legal. For every Linux DVD image distributed by bittorrent, there is probably dozens of times that much data in blatantly bootlegged content being distributed.
The net has been about to melt down any day now for at least 10 years.
Quick, someone add more tubes!
Gamers, VoIP and video conference users beware. The leading BitTorrent software authors have declared war on you - and any users wanting to wring high performance out of their networks.
What a load of nonsense. The best solution (if there's a problem here in the first place) is for ISPs to drop any bandwidth allocation.
Obviously, isps aren't going to lay down and die. They'll simply throttle the offending users and throttle udp for residential customers. The problem with this is that legitimate applications like voip will be blocked or throttled with the excuse of "fighting the thieving pirates". I hate packet tampering as much as anyone else here, but without qos rules everyone loses.
The whole point of 95% of torrent transfers is that the users are greedy and don't care about anyone but themselves. They couldn't care less about stealing to the point of killing one golden goose (the music industry), what makes you think they give a hoot about killing the internet?
Well, since The Register don't seem to want to print my comment*, I'll repeat it here:
"I think this is a bit of scaremongering that's missing one vital point:
When an ISP throttles UDP packets because somebody is using excessive bandwidth, they'll be dropping packets *from that source*.
So while .torrent moving to UDP is going to affect VOiP and games, the effects of that will be *restricted to the person using excessive bandwidth* via bittorrent. There's no reason it would affect anybody else, and I doubt ISP's are going to be dumb enough to block packets at random.
Unfortunately that kind of blows the articles entire premise out of the water."
Myx
* Posted at 12:40pm, ten minutes after the article appeared, at a point where there were no other comments on the article. 3 hours later there are 37 comments, but no sign of mine. Now it may be that they've just been overwhelmed with comments, but I'm a suspicious soul at times...
If they drop TCP packets or try to fake reset the connection, it's obvious and provable. If they silently discard UDP packets, that's just normal network behavior.
I was thinking. The people doing Skype have a point. I don't care if my packets take three, four, or five times as long to get to me. I ONLY care about the total time to get the 700 MB (or whatever). It could take a path five times as long as the voice communication, I don't care.
Then it hit me. A CD-ROM. They're always falling off the shelf above me.
As I lay there rubbing my head I thought: if my roommates voice call is so important, he wont mind peering into a mesh network at the same time, to keep me from using his bandwidth with his ISP. The thing about mesh networks is, they're nowhere near as direct as your ISP. But with bittorrent it just and simply DOESN'T MATTER. It literally doesn't matter if it it take 8 seconds to get from me to my roommate to his neighbor to their roommate to their son downstairs in the basement to their neighbor across the lawn who is one of my seeds. ALL that matters is not getting throttled.
Folks: Bittorrent is the killer mesh application, and we need to get people who want their precious skype to realize that just by joining the mesh, they can improve their voice quality! It's the perfect trade-off.
UDP does not guarantee delivery. If ISP's want to, they can simply start dropping UDP packets once the total amount exceeds a certain threshold. This should be almost trivial to implement.
Sure, just blindly dropping all types of UDP packets will also degrade VoIP services etc, but certainly this does not need to impact "the entire speed of the internet".
Since VoIP and other "normal" uses of UDP do not need terribly high bandwidth, the problem can be easily solved by imposing a maximum UDP throughput per IP and simply dropping any UDP packets past that limit. That way, VoIP will still work just fine but other services "abusing" UDP will just be effectively capped by the unguaranteed delivery.
I'd love to see lawsuits about this as well, as UDP does not guarantee delivery so you would hardly have a basis to complain when ISP's drop such packets, especially as long as they deliver *most*, but not necessarily all such packets.
Every expression is true, for a given value of 'true'
In the end, is this truly going to change how the ISPs work? The article mentions that one of the reason for doing this is because of Bell Canada. I personally don't think Bell Canada is going to care - neither will Comcast and all the other ISPs who throttle and shape the network (can we say all ISPs at this point to at least some extent?).
Really, all this will do is continue to drive a wedge between the "evil" file sharers and "enlightened" ISPs who will attempt to use the other 95% of the users to make claims to whoever will listen... Personally I don't do much file sharing and I'd get real pissed if my broadband speed dropped to a quarter of what I actually pay for.
Obviously if you're going to try to transfer large files over UDP you're going to need to develop some way to ensure reliable delivery - which is exactly what TCP does. TCP has years of work behind it, so the odds of you coming up with something just as good as TCP are fairly slim. So I can certainly understand why folks out there would be somewhat apprehensive about this decision... There is a distinct possibility that the new protocol will waste tons of bandwidth or do something horrible to existing equipment or summon up a shoggoth. There is certainly the possibility that damage will be done.
But is that necessarily bad for the Internet? ISPs are regulating the hell out of TCP traffic. They're shaping and compressing virtually every packet that crosses their networks. They're blocking ports and resetting connections. They are intentionally preventing their customers from using the bandwidth they've bought in the way they want. Doesn't that count as damage?
Maybe this is exactly what the Internet needs to drive home the point of network neutrality. Maybe if ISPs get stuck with a new, horribly inefficient protocol that they can't mangle they'll realize it was a bad idea to abuse TCP.
"Work is the curse of the drinking classes." -Oscar Wilde
Is more in cause of the frickin' meltdown of the internet with about half the content of web sites being Flash-based animated and (GAH!) audio adds.
Flash video is also irreparably defective.
Disclaimer: did I mention I hate Flash?
The problem with UDP rate control is:
a) Unless you make it TCP fair, you stomp on the user's OWN traffic, which is already a big problem for BitTorrent clients which fill up DSL and cable-modem buffers. And if you DO make it fair, then it doesn't matter.
b) It doesn't stop ISP traffic management, it just forces their devices to be inline.
c) The biggest offender, Comcast, is moving away from P2P blocking anyway.
Test your net with Netalyzr
They use UDP. And it eats bandwidth like none other. Oh noes!
It's all fun and games till someone divides by 0. Then it's hilarious.
Correct me if I'm wrong, but I think actually we can reduce the bandwidth by switching to UDP. With UDP you need to implement your own transmission control, and if BitTorrent can make its own lightweight implementation, the transmission control overhead caused by TCP will be reduced.
Actually I see UDP a better alternative for BT because you don't need to make sure every packet is transmitted successfully, given how BT seeding works.
I see TFA's point is not that UDP increases traffic, but they are harder to be throttled by ISPs. Well then why don't the ISPs upgrade their own infrastructure to handle the increased traffic and charge their users accordingly to cover the cost? Blame the current economy?
Again, I don't know much in this area so I may be wrong.
Colorless green Cthulhu waits dreaming furiously.
...but nobody wants to pay for it. It's been said many, many, many, times before but the average user doesn't have any concept how much bandwidth costs for the circuit to a carrier alone, much less the hardware required to light it. I work with carrier-level Cisco gear, a single linecard alone is in the 50k price range. A single router I work with has 8 of those. It takes at least 2 of those routers to handle a few small to medium size towns, (30k subscribers). That's just the price to give you a connection back to the local building, of course I'm omitting the cost of the wiring to your home, the equipment required to power it, etc, etc. We haven't even discussed how much the transport out to the internet begins to cost. I think a lot of ISP's are beginning to see that it's probably a failing business model, and because of that they are making some-what drastic changes to try and make it sucessful. Things like bittorrent, youtube, etc are what make the web truely great, but at the same time they very well could be the downfall in the current state of the internet. You of course could always get your own internet circuit but even a T1 will be at least $300 per month + construction costs and appropriate gear to utilize it.
ISP's have been managing UDP traffic for years now, this won't change anything. Any of the deep packet inspection boxes (Packeteer, Allot, Sandvine, Ellacoya, etc) can identify the traffic whether it is UDP or TCP as can open source tools like Ntop. Encrypting the traffic can of course disguise what's in the packets, but the overhead hurts transfer speed. In addition, several of the new generation of traffic shapers don't even care what layer 4 protocol you're using, things like Netequalizer just looks at the two IP end points of a given conversation and treats it as a flow regardless.
This all sounds familiar, this promise that all computers would collapse and the internet become a smoking ruin that could never be used again. As long as I can log on and keep reading these prophecies of doom, I'll know that I still have plenty of time left before I need to go to the hardware store to get the materials for my The End is Nigh sign.
from http://www.dslreports.com/forum/r21500602- :
Re: Is this a good thing for the net?
Yawn, here comes the typical argument... bandwidth is bandwidth, either way you look at it. All p2p does is open several simultaneous connections, splitting the user's bandwidth. Unless you horribly misconfigured your client to open up, say, 1000 ports.
It's not as if the user is using any more bandwidth than if they were conducting a regular http download. P2P actually is better for a network, as (given enough peers) it completes downloads significantly faster than normal centralized server methods, thus getting heavy users off the network noticeably faster (obviously, unless the user is dumb enough to allocate their entire upstream bandwidth to seeding).
As to bypassing the "TCP congestion control" you speak of, do you think Bell's solution is ANY better? The throttling of particular packets by itself violates the principles of TCP. Not only that, it also throttles/cripples MANY legitimate applications, such as secure VPN's or other encrypted connections.
Do you REALLY want that as an alternative to this so-called "problem" of p2p? I've said over an over, the ideal solution is to gracefully scale back speed for ANY upload/download if the said user is using their full bandwidth for more than 20 minutes during peak hours. This actually solves the problem, unlike throttling schemes like bell's, which render many legitimate applications useless. Let's face it, even Comcast here in the states has been forced to take a long hard look at their policy on Sandvine. Soon enough, we can only hope Bell will as well...
Do I even support the above solution? By itself, absolutely NOT!! IMHO, the ideal solution is to upgrade the core and its routers. However, that takes time and capital that companies like Bell are rather unwilling to spend; they'd rather (ab)use their position in the limited Canadian ISP market to deploy band-aid solutions like throttling p2p.
I find it disappointing that ISPs don't meter usage. It would help cut down on spam and viruses for example if users suddenly realized that something was costing them a lot of money and wasting bandwidth. I mean all our other services are metered. As for myself there are months when I download huge amounts of anime and then there are other months where I download next to nothing yet I still pay the same amount. This fact alone means it's more beneficial for me to download like a nutcase and ruin it for everyone else. Granted the only catch is that ISPs would hopefully charge reasonable rates with a certain flat fee to maintain the line. To folks to believe otherwise, I suspect you're not willing to give up your free lunch to the expense of others. The Internet is a limited resource as some ISPs are learning the hard way. Given the choice between metered usage versus throttled / controlled / broken Internet, I'd pay for metered anyday.
The internet evolved as a gentleman's system in the comfortable confines of the ivory towers of academe, but now that it's an essential part of daily life for more than a billion people, the time has come to get realistic about its management.
First off the internet originated from ARPANET a military funded library project...it was hardly an evolutionary gentlemen system. Secondly, the things the internet is two most used for are social networking sites, and pornography. Essential part of daily life? I hardly think so for those billion people.
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.
What the hell is he even talking about? OK children and junkies don't need the internet, gotcha.
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.
When is the last time anyone here heard of someone complaining about lack of bandwidth because their neighbor was using too many torrents? Never?
Most of the time when I hear crap like "By most estimates..." with out any sign of a source to back it up, I attribute the remainder of the sentence the same amount of credence as the sound of my coworker's ass cheaks flapping together after an especially hanious fart.
Maybe he's right, but with out anything to back up his opinion, he just looks like some shill who is lobying for some organization with a strong financial incentive for not seeing net nuetrality laws and being allowed to run deep packet inspection.
The best I can find is Ellacoya's June 2007 report that put P2P at 37% of total bandwidth. http://www.businesswire.com/portal/site/google/index.jsp?ndmViewId=news_view&newsId=20070618005912&newsLang=en
A wee bit shy of the 50% the author is claiming.
Another obvious way to see what the impact is would be to look at a tool like http://www.internettrafficreport.com/30day.htm to see if the change to UDP and expected rise in bandwidth actually effects TCP communication. If it is as gloom and doom as the author makes it out to be, we should see a steady rise in lost packets as the P2P users upgrade to the new UDP defaulting version.
This report from March 2008 http://www.foxnews.com/story/0,2933,342988,00.html sites Arbor Networks (they bought out Ellacoya in early 2008) claims P2P traffic represents about a third of internet traffic.
I'm all for making a plan to be able to react if a problem is detected. But lets not get all worked up over someone's questionable theories.
-Rick
"Most people in the U.S. wouldn't know they live in a tyrannical state if it walked up and grabbed their junk." - MyFirs
This sounds like the basest kind of scare mongering, relying on a basic ignorance of the way networks work.
UDP is not any less filterable than TCP. To even make this argument, the reasoning is so contorted as to be silly. In either case, one uses a router to inspect packets and decide what to do with them. ISPs will simply go as deep through the envelopes as they like; they already do. With that knowledge they will do whatever is allowed by law. At present, almost anything is. If they abuse that power too foolishly, then it will start to be taken away from them.
And in the meantime, whoever they filter will tweak to retaliate, and it will always be a race. As far as I can see, this is just the ISPs (or their proxies) stopping at one random lap and crying how unfair it all is.
Why ignore the real issue here? If you sold a teenager in Topeka unlimited use of a large pipe, but now cannot handle her actual unlimited use of her large pipe, then you just need to start cutting better deals.
It's as simple as that.
If the teenager cannot actually use her fat pipe, 100% of the time, then stop lying about what it is you have sold to her. Either charge more or advertise less. It's as simple as that.
When I as a CEO, and millions of others like me, buy #MB upstream and #MB downstream, and utilize it 100%, 24/7, no one quakes over the calamity of the internet backbone melting down.
All of this discussion over filtering is really a discussion of pricing. And the fact that we are talking about it in the wrong terms is creepy.
Believe me, you do not want a bunch of unaccountable telecom bureaucrats playing god with the backbone. You want a free market making these decisions.
Tired of Political Trolls? Opt Out!
Perhaps if the big telco's spent a bit more on upgrading, widening and developing their networks rather than just pounding penny profits to shareholders this wouldn't be an issue. When are they going to realize that the people paying $35 ~ $200 a month for services which today cost about 10% of the charge are the real shareholders, and are the only real reason they are in business.
flinging poop since 1969
Richard Bennett could not more obviously present himself as a shill for comcast and the like. The article is complete nonsense, attempting to portray bittorrent as the enemy of the Internet.
This idiot drones on about the "ungentlemanly" conduct of using UDP for such purposes, but conveniently avoids the fact that comcast/sandvine caused this mess by injecting face TCP resets, to break bittorrent's TCP connections. Well, what does he expect would happen?
Obviously, UDP is not a good choice for bulk transfers as it lacks congestion control, but lets be fair about where the fault lies. This is not something that can be worked around at the application level, and after being pushed into this corner, there is little else that can be done to work around their abuses of the TCP protocol.
when you enter "google" into google, you can break the internet!
One would have guessed Spam is more of a problem. Or YouTube. Or the internet TV some networks are pushing now. Or maybe even WoW. Why? Because EVERYONE does it! Everyone gets spam, everyone is on YouTube, 11 Millions playing WoW... ok, nobody cares about IPTV, but still.
No. It ain't the spam that clogs my mailbox (and no, spam is no longer text. It's effing HUGE pics!), it's not megabytes of videos being streamed to every other computer connected to the net, it's not WoWheads spending 20 hours a day in front of their machine. It's Bittorrent. Something that less than one in thousand internetusers uses with some dedication that could be called "heavy use".
How about doing something against spam and botnets? I'd wager my salary that would instantly take care of any congestation, too. At least if that's the real problem we're trying to deal with...
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
BitTorrent is believed to be harboring weapons of mass destruction. These weapons are believed to be capable of destroying all of the internet tubes. The government has no choice but to authorize the ISPs to use lethal force to prevent these terrorists from succeeding.
You are using English. Please learn the difference between loose and lose; they're, there, and their; your and you're.
While UDP itself fundamentally has no congestion control, this doesn't mean that the transport protocol layered over it by the application (Almost all uses of UDP have an additional application-based transport protocol layered over them, such as RTP or Sun RPC) doesn't also have congestion control.
Was this meant to bypass congestion control (probably not, TCP congestion control is fine and traffic shaping is usually fine if done right), or was it done in response to Sandvining which has NOTHING to do with congestion control and everything to do with the ability of an attacker to shut down a TCP connection with a Man-In-The-Middle attack (fake RSTs)?
I'm assuming it's a defense against MITM attacks, in which case it doesn't mean the end of the Internet as long as the new congestion control approach is sensible. Unfortunately we can likely expect lots of bugs and misbehaving applications so this could be bad.
retrorocket.o not found, launch anyway?
His example is that multiple users behind ONE of these DSLAM ports have congestion. Per your example, the congestion is as soon as they reach their 64 kb/s policy.
However, everyone in these bittorrent debates pretends that the DSLAM port is the bottleneck. In a highly interconnected environment like a world full of bittorrent and other users, there are many other places for congestion far from the simple consumer-to-ISP policy enforcement point. TCP congestion control helps with congestion anywhere that multiple flow paths intersect. The Internet does not magically prevent intersections and congestion except at consumer head-ends.
I think part of the answer is for Quality of Service (QOS) to be implemented properly and universally. Would even the most rabid BT user/provider care hugely if his traffic throughput took a back seat to video for the space of several fragments? It still wouldn't be shut down completely. The win is that when no higher-QOS traffic is on a network segment, BT gets the whole "tube."
There is much more of a workable QOS in IPv6 than in IPv4, where it was more or less grafted on after the original design.
Let's get moving to IPv6.
Well, apparently, you only have to fool the majority of people for a little while.
Right now the biggest problem ISPs have with bittorrent is not the bandwidth it uses, but the massive packet load it puts on their border routers, DSLAMs, etc. This is 99% due to allowing hundreds of simultaneous incoming connections. bittorrent clients have never been able to throttle incoming bandwidth aggregated across hundreds of connections very well, and the result is thousands of packets worth of backlog on the ISP's border routers, DSLAMs, etc, which blows them up.
Changing to UDP does not solve this problem, it just makes things even worse. The bittorrent people need to get their act together. The ISPs are just going to respond by changing their bandwidth filtering to be whole-IP-based instead of connection-based, and the result will be that people trying to use bittorrent will wind up with crappy, unreliable links for EVERYTHING they are trying to do over the internet.
I don't know why people using bittorrent should expect the ISPs to 'play nice' when they clearly don't give a damn about the mess bittorrent makes of the ISPs own infrastructure.
-Matt
These sorts of heavily biased, misleading articles are why I can't stand reading The Register any more. It's all GNU-bashing and items on defence spending. I don't remember it being this bad a few years ago. In any case, I don't want to read it any more.
Look, I'm a huge BitTorrent guy, I run uTorrent every day and pretty much max out my connection for about 8 hours a night.
That said, I don't understand this attitude of "If you sold me 10Mbps, then you'd better be able to give it to me 24/7"
The electricity in your home is also sold to you on the assumption that you can run any and all household appliances from it.
But guess what? If you switch everything on, because you're allowed to and can afford to, and so do your neighbours, and everyone else on the grid... the grid collapses.
There were certain statistical assumptions built into the grid and the pricing model. One of those assumptions was that their customers wouldn't take up 100% of their allowable usage ALL THE TIME.
This is hardly a new thing.
Anyone who doesn't think that numerous products, Narus (http://en.wikipedia.org/wiki/Narus) being one example that comes to mind, won't rapidly generate IPDR to track and immediatly filter/blackhole such excessive UDP dataflows is sadly mistaken. I remember when ping floods were the death nail. It rapidly became a non-issue. If TOrrent wants to make a brillany decision that immediately shines the megawatt spotlight on them being the nasties in the network neighborhood they will, in very short order, see the million monkey army come up with that quick fix brillant solution to quickly put them out of business.
In today's world of high speed networking, I'm honestly surprised that UDP is even allowed to traverse the backbone. Short of the million monkey solution, rather than allowing a runaway child to melt down the big boy back bone providers, I would see them black holing UDP and forcing all protocols to use TCP. Sure it will kill some audo and video streaming. It will also decrease utilization at the backbone by 75% and improve all the "real" traffic (HTTP/HTTPS/SMTP) performance by an order of magnitude. Sucks to be a media content provider. They beter start looking at TCP alternatives if they want a sustainable business model. Its not like people haven't been considering the possibility (http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=01295064, http://en.wikipedia.org/wiki/Streaming_media). RTSP will just become the defacto standard rather than an alternative. .02.
My
Armaments, 2-9-21 And Saint Attila raised the hand grenade up on high, saying, 'O Lord, bless this Thy hand grenade' N
Its true for now that most links are sold purely by bandwidth, so your statement about getting what you've paid for is valid -- unless what you're paying for includes by contract a cap on total use or continuous use, or something else. Most home network access contracts contain those use case caps.
What really caught my attention as I read your note is that comment about the number of open tcp connections. As I read it, I'm surprised that hasn't been used as a valid limit on use by contract.
If you think of things that add cost and hurt performance from an ISP perspective, total number of sessions is one of them. It increases the load on routers and adds overhead and latency to network (or it can, unless more money is spent on bigger and faster routers).
Limiting concurrent sessions is probably just around the corner.
I currently pay for the highest bandwidth version of consumer net access offered in my area. Its more that sufficient for downloads, but because I work from home it is slower than I'd like for uploads. I use a hosted server at a co-lo site so that my business system isn't carried over a consumer line, but I still pull down a pretty huge amount of data sometimes.
I'm completely convinced that if my ISP wanted, they could slow my links by about the third week of most months. I believe they don't only because they don't offer a higher level of service than I'm paying for. At least for me, they've always been fair and responsive.
If anyone is limited by an ISP when they've got an agreement to pay for services that doesn't support those limits being applied, they should take legal action. If not, they should pay more attention to what they purchased and either refuse to buy what's offered or live inside it.
If you can't live with what's in the agreement, lease a line and pay for your own service to somewhere. What you'll find is that without the aggregation that's done by the big ISP's, you'll never be able to afford the on-demand use you want to buy.
I want massive bandwidth on demand too -- I want to download 4gb movies in under 30 minutes whenever I want. I do not, however, want to pay for a leased line capable of doing that. When I buy into a shared provisioning system (a consumer isp arrangement) I'm agreeing to live within that ecosystem and share the cost of that high bandwidth as well as sharing that bandwidth.
The contracts are obscure and don't come right out and say so. Maybe it would be better if they did.
The problem with quotes on the internet, is that nobody bothers to check their veracity. -- Abraham Lincoln
I will admit to the occasional download using BT. I justify my downloads as they are available for free from the web site of the show, but the plugins are not available for me to view the streams.
Anyway, I get letters from my ISP that are copies of the complaint sent to them by the studio. The studio is threatening the ISP to try to force them to do the bullying of the "violator".
I am not in any way condoning what the ISPs are doing and am in no way condoning uTorret's use of UDP to try to circumvent the issue under discussion in the original post.
Just my $0.02 worth.
It enforces bad design -- most client/server applications should be either stateless or session-based, rather than connection-oriented.
What? Why? Why is a connection based application "bad design"?
A "session" is just a hack to give you connection-like qualities over a connectionless protocol, such as HTTP. If you want connection-like behavior, and you're free to design your own protocol, why not just use connections? Why make your life more complicated?
There are plenty of advantages to connection-based applications; no need to re-authenticate on every message, for one. Clear indications of when a client is done talking to you for another (the connection is dropped, whereas in a connectionless world you never really know if the client has died or is just about to send you another message).
Even most "connectionless" applications, such as your web browser, try to cheat and get connection based behavior. Renegotiating an SSL connection for every request would be remarkably expensive for the server, so the server and browser try to reuse the same connection as much as possible. Fire up wireshark, and you'll see your browser will only open two connections when you read Slashdot, and all the images and text all come across those two connections.
Isn't the main issue against ISPs not the throttling of flows but the injection of fake TCP RST (Reset) packets and other such illegalities. Illegal from the PoV that the protocol specification and IETF have not ratified this behavior. http://www.eff.org/wp/detecting-packet-injection
What is needed is a "secure TCP" where all control information within the packets at TCP level are cryptographically signed like with a MAC checksum. (No SSL does not fix this problem! SSL sits on top of TCP and ISPs a messing with flows at TCP level)
This will however require kernel support at both endpoints. Maybe it could even be implemented as a series of TCP options.
I would not support actions that lead to an internet meltdown that the article suggests, /. is presuming that no flow or congestion control will be implemented via this UDP alternative to TCP (which remains to be confirmed and certainly the forum in the article indicates THERE WILL BE flow/congestion control implemented with the UDP alternative).
Why don't they just use a form of TCP protocol but apply techniques from "VJ Compression" and add a MAC code it should be possible to get the overhead of sequencing .
TCP header (no options) is 20 bytes. UDP header is 8 bytes. UDP provides you port info already for connection tuples. So you have a 12 byte budget to implement:
* reliable sequencing
* flow-control/retransmission (window control/timestamps/selective-ack)
* MAC/protocol security (the real reason IMHO why they need to resort to UDP in the first place)
That should be ample budget. We specifically don't need hi-performance delivery just a protocol that scales well enough in the domain of 30mbit at the users access point communicating to between 1 and 30 others.
Routers deeper inside the network cannot keep tabs on millions of IPs and who uses how much, they already have much to do.
This article (And the summary, of course) is spinning this way out of control.
This is taken DIRECTLY from the guys behind uTorrent:
Source: http://forum.utorrent.com/viewtopic.php?id=49813
Hey look at that, this actually stops uTorrent using more Bandwidth than you have. If your ISP sells you an 8Mbit connection and you use all 8Mbit of it, surely that's all well and good? If the ISP can't handle you using all 8Mbit, then they shouldn't sell you it. Simple-as.
If (And it's a big if) this actually does cause any kind of "Internet meltdown", it'll be because the ISPs oversold on what they can actually deliver - it's not your fault, my fault or uTorrent's fault.
Hopefully ISP's will either stop overselling their bandwidth or update their infrastructure to cope.
Progress++;
+1 IDisagreeSoHeMustBeATrollOrAnAstroturferOrAShill
The author of this article could have called or emailed Bram Cohen before writing this article, but then he wouldn't have had such sensational tripe to garner page views. If he had, he would have known that he has got it completely wrong. The switch to uTP is actually to make BitTorrent traffic more friendly to Internet traffic. You see, BitTorrent is trying to sell a content delivery service based on their client and the #1 complaint from their customers (businesses with content to deliver) and their customer's customers (end users) is that the BitTorrent DNA client seeding/downloading in the background hurts the performance of other applications. That's unacceptable if you're trying to sell an unobtrusive alternative/complement to traditional CDN.
Yup, good ol TCP is what is causing the problem. That's because BitTorrent breaks the assumption in TCP that one application needs only one TCP stream to do its work. To solve the problem BitTorrent acquired advanced congestion control techonology and it's inventors from "Plicto." The congestion control technology lets BitTorrent work without causing crazy latency for other applications on the box. BitTorrent is the responsible party here, recognizing the need for congestion control and implementing it in their protocol. Compare that to the author of this article who saw that BT was using UDP and assumed it was a naive attempt to get around ISP blocks.
The people who work at BitTorrent are smart enough to know that you can't beat your ISP by making a new protocol. The ISP sees all and can control all, even if it may lag behind the changes. That's why BitTorrent has been working to make changes where it can make a lasting difference, in the political layer of the network.
mod up
Both BitTorrent and Comcast are working in the IETF ALTO working group, which is intended to improve the use of bandwidth and other resources by P2P.
Having been in these sessions, it is clear to me that BitTorrent has no interest in melting down the Internet and is well aware of the implications of what they are doing. Note that if worse comes to worse, UDP can be blocked too.
Is the IP selling constant bandwidth, or peak bandwidth? I'd argue that if you have a business account, it's constant bandwidth. But when they sell residential accounts, what you buy is peak bandwidth.
Buying a car that easily goes 100 m.p.h. doesn't mean you've bought the right to constantly go 100 m.p.h. - even though you've paid taxes and registration on the car and taxes on the fuel that helped build the roads.
Most residential bandwidth users don't care about constant bandwidth. All they want is peak to be there when they require it. And all most drivers require from their 100 m.p.h.-capable engine is quick acceleration on the on-ramp, or to avoid a hazard. In selling you that, there's no ethical requirement that you also be provided the option of running what you've bought constantly at the peak capacity.
If you want that, you need an arrangement with a race track ... or a business-class Internet account.
"with their freedom lost all virtue lose" - Milton
If the ISPs are so interested in saving bandwidth, why don't they turn on multicast on their routers.
Oh wait, that would cost money.
Seems to me the problem isn't bittorrent. I happen to pay for a 7.5mbit down, 1.5mbit up connection. So does everyone else with my provider. But, I guess the internet providers pass out bandwidth the way banks pass out money - that is, all the customers have a certain bandwidth to share and the provider simply hopes no more than 10% of their customers (or whatever the number is) will make a run on the bandwidth all at the same time. How about you give and guarantee me the bandwidth I pay for? You know, the bandwidth in all your advertising and contracts? I never use more than my 7.5/1.5mbit connection, so there should be no reason that me maxing out my connection should interfere with anybody else - if it does that's because it's really shared bandwidth (not the bandwidth that I bought which should belong to me, used however I see appropriate), and perhaps we should look at fixing your infrastructure problems and business model instead of pointing at my usage habits?
I'm not sure why BitTorrent is getting the bad rap here... The real problem is companies like Comcast selling connection speeds that their network can't actually support if most people were to use them. Instead of investing in there network or perhaps selling what they could actually support they start filtering content.
Comcast *shut off* two of my friends connections for downloading the new Fedora ISOs via BitTorrent. This is just getting out of hand.
If a user have bought a 2Mbit connection then it does not matter to the ISP what kind of trafic that goes on that link, so it should not impact any other users if I use udp instead of tcp. It is not as if I am allowed to send more then 2Mbit just because I use udp. (It is all just ip packets for the router anyway).
There are 2 situations where using udp might be a problem.
1: If the isp sell you a 2Mbit connection but can't handle that on their internal network. In that case the tcp connections might(We don't know which slowdown solution that uTorrent uses) back down faster then the udp, and thus give more bandwidth to upd.
But does this really happen? I live in Denmark and the 5 different danish isps I have had so far, have always been able to deliver the full speed i bought on their internal network. So if this is a problem for isps in USA, then its time for the Isp to upgrade their hardware. An temporary solution might be to lower the speed of all customers duing peek time.
2: If you share a single internet connection, and don't have any internal rate limit software this might be a real problem, but the solution is simply to limit the speed in the uTorrent client.
I forget the statistics, but let's admit it -- it's highly likely that BitTorrent developers, on all platforms all know, understand and encourage the fact that some enormous fraction (far above half) of total BitTorrent traffic is used to illegally distribute copyrighted content.
I'm not going to make a value judgment on that for the purposes of this post, although I have an opinion on its moral correctness -- but many of the features introduced by BitTorrent clients (protocol encryption being the biggest example) seem to exist solely because BitTorrent is being blocked by ISPs.
BitTorrent is being blocked by ISPs because it is hugely taxing on network infrastructure, and is almost always being used to do something illegal. I can count on one hand the number of times anyone I know has downloaded Linux ISOs over BitTorrent, but a colleague's recent purchase of as many 1TB hard drives as his SATA controller will support for the express purpose of "downloading every Bluray rip he can find" is, I believe, reflective of the general use of the protocol. Maybe not its intended use, but certainly its defacto use.
BitTorrent developers have to realize that their protocol is almost exclusively for illegal purposes. By constantly adding features to "get around" restrictions placed by ISPs on the protocol, they're actively supporting such activity.
It is insane to think that broadband shouldn't be oversold by a contention ratio. Why shouldn't it? Almost every ISP advertises speeds "up to X megabits per second", key word being "up to". Most people even in normal use will never see those speeds -- because their usage habits never require it of the system. Or if they do, it's for a fraction of the time that a BitTorrent session will max out a connection. The most taxing thing I've done that I can think of is load a page of several hundred ~1MB photographs in Firefox. It took a long time to download, but that's only a 200MB transfer that happened once -- not someone downloading multiple HD movie rips at all hours of the day, all the time.
If the majority of BT usage was for legal purposes, the traffic using the protocol would drop down to near-zero and it wouldn't be an issue. It's only come up because of the connections to illegal activity.
What we need is a legal solution to the problem. Let people use their connections for whatever they want -- we just need a more reliable way to punish those who break the law, and to do so more reliably. The exact mechanics of this solution can be debated from now until the end of the Universe -- but a technical solution to a social problem will never be successful. Whether it's a change in business models, more policing, aggressive prosecution of large-scale Internet distributors, or something else, there needs to be a change on the human side of things.
Otherwise, they're likely right -- the Internet, in its current state, will not survive. And it will be BitTorrent's fault.
Wait - let me see if I get this straight...
ISPs, the gateways to the Internet, are doing bandwidth management at the TCP layer and above. Since bandwidth on an IP network is properly measured at the IP layer (duh), obviously the ISPs are not actually doing bandwidth management. They are doing something, but it is not bandwidth management.
Some software operates on not-TCP. If that software is high bandwidth, the ISP's "bandwidth management" (which is not really bandwidth management at all) fails. And that is the fault of the software?!?
"We built a toll road, and in order to eliminate traffic jams, we strictly regulate the number of blue cars entering the road. But people are driving cars that are not blue, and causing traffic jams. Those people driving not-blue cars are flooding our road!"
Stop-Prism.org: Opt Out of Surveillance
Comment removed based on user account deletion
the nice thing about capitalism is that it's not my job to worry about the health of the internet. I have a contract that clearly states what I can expect for the monthly service fees I'm paying and this means I get to download at 16Mbits ALL DAMN DAY. Should I have priority over VoIP and gaming? I don't know and that's really not my job either. Let the ISPs figure it out, they are the (sorry to use the word) stakeholders in this.
___
No power in the 'verse can stop me
Mislabeling it is not. If you sell 1MBps with a 25GB/month cap, then you need to be advertising your "1MBps peak bandwidth, 0.01MBps constant bandwidth" service, not misleading your prospective customers.
Practically every ISP should be overselling peak bandwidth; because people don't all use it at the same time, your only choices are to let them use as much as they can (overselling) or to throttle them. But both peak and aggregate bandwidth are important; if you're not providing much of the latter you shouldn't get to imply otherwise.
Someone get a plumber, quick!
Let's call Joe the Plumber ;-)
Catalin Braescu
Ofaly.com
It says what is really happening - assuming it is true, of course.
Why do they say that this will happen? One has paid for a certain width on ones internet, so one should be able to use it. If they don't let us use what we have paid for, then they have sold us something that dont exist...?
Dr. Peter Venkman: This city is headed for a disaster of biblical proportions. Mayor: What do you mean, "biblical"? Dr Ray Stantz: What he means is Old Testament, Mr. Mayor, real wrath of God type stuff. Dr. Peter Venkman: Exactly. Dr Ray Stantz: Fire and brimstone coming down from the skies! Rivers and seas boiling! Dr. Egon Spengler: Forty years of darkness! Earthquakes, volcanoes... Winston Zeddemore: The dead rising from the grave! Dr. Peter Venkman: Human sacrifice, dogs and cats living together... mass hysteria! But really. The Internet isn't a damn packet radio network. Transmission errors are largely a thing that has been taken care of via providers putting infrastructure in place to make retransmits unnecessary. In fact, BitTorrent over UDP shuld sould like a good thing to Service Providers....unless they are over-subscribed kinda like this whole subprime thing. Wait. Comcast...node...saturation....no wonder they're spreading fire-and-brimstone FUD. They HATE spending money. On the other-hand, we have hashing. So if the connection is adequate, then ISPs shuld praise this sort of thing for lowering their network overhead.
Every week there's another post about how "X" is going to cause the Internet to come to a grinding halt and cause a meltdown. Yet each day more and more videos are posted on YouTube of kittens riding roombas, and somehow the Internet survives.
Call me when the sky is MEASURABLY and DEMONSTRATIVELY falling.
What? That makes no sense.
Unless I missed something and nowadays downloading at high speeds suddenly endangers the lives of those sharing your connection, you may want to rethink that comparison.
Comment removed based on user account deletion
and get off your lawn? Care to back up any of your claims with actual information? "OMG, Flash sucks!!!!!!" doesn't mean shit.
How is Flash video defective exactly?
Too bad you don't like newfangled audio/video on your websites. A lot of us find it very useful, especially for tutorials and such.
I've been thinking about this topic for a while. It almost always degenerates into a "I paid for X mbps, I should get to use it 100% of the time" vs. "You're killing my connection, and my XYZ traffic is getting hit even though I'm a good consumer, we should pay for each bit we use, and let the market sort it out." What if we implement a QoS service level based largely on the existing pricing model. When you subscribe, you get a certain bandwidth of traffic that you are (almost) GUARANTEED (as if you were (almost) leasing a T1 to yourself) The ISP doesn't mess with it. The rest of your traffic is "best efforts" at between X and Y mbps. Let the ISP shape the "best efforts" bandwidth in whatever way they feel brings the best average consumer experience. Let the customer choose if they want to use their guaranteed traffic to surf the web, run VOIP, Games, BT etc. That way I'm not limiting your BT, and your BT isn't killing the voice quality of my phone. Everyone talks like QoS, shaping, and throttling is a bad thing. I've used all 3 tools on my own LAN to IMPROVE the connection of my network for ALL it's users. Sure some HTTP traffic gets delayed while Voip jumps the queue, and when there's heavy surfing, BT slows down. Network bandwidth is a finite resource. Burning it up like fossil fuels in the 60s is a bad long term idea. I can't afford a guaranteed bandwidth connection at home. I'd much rather participate in a MUCH bigger shared and shaped pipe than be stuck with what I can afford to buy all for myself.
Something about more sessions to track uses more cpu and memory. I'd imagine it has to do with the tables required to track the sessions.
The same is true in your workstation. There are performance limits and real serious points of diminishing return as you increase the number of concurrent sessions. Some operating systems are better than others at dealing with this, but there are limits in all cases.
The problem with quotes on the internet, is that nobody bothers to check their veracity. -- Abraham Lincoln
If you want true net neutrality get rid of broadband and make everyone connect, including businesses, using a 14.4 modem. Current infrastructure can handle that kind of bandwidth for every citizen. In ten years upgrade everyone to 28.8, future infrastructure will be able to support it I think...
Might as well, because in a net neutral world that's what happens. Not because of technical mumbo jumbo some techie tells his CEO about, but because of the bottom line cost. Equality reduces throughput for everyone, and increases it for no one. That's the way business works and thinks.
Analogy time: In 1990 the average professional woman's salary was 60% of what a man's was with similar experience. In 2000 it was 80%, but the average salary of women only increased 4%. Guess what happened, to make things more equal men were payed less. Money is a limited resource, bandwidth is a limited resource. Business isn't going to pay anyone more to make things equal, they are going to pay someone less.
Even if they upgraded infrastructure, people would just download more movies and audio. People who download from BT don't just get 1 movie or one song. They get whatever they can, which IS limitless, unlike the bandwidth they are using to get it.
Simple supply and demand economics really.
That said, I'm all for net neutrality. In fact I don't even support overall bandwidth cap limits. If your network can do 1000Mbps down, let me try to pull it off on my machine, along with everyone else on your network...it would be fun watching this thing seize like a 90y/o epileptic at a laser light show. After all, I miss Prodigy.
See http://forum.utorrent.com/viewtopic.php?id=49813
"What is in 1.9:
uTP, the micro transport protocol. This UDP-based reliable transport is designed to minimize latency, but still maximize bandwidth when the latency is not excessive. We use this for communication between peers instead of TCP, if both sides support it. In addition, we use information from this transport, if active, to control the transfer rate of TCP connections. This means uTorrent, when using uTP, should not kill your net connection - even if you do not set any rate limits.
What was in 1.8.1:
uTP, but connection attempts were not initiated by default, and there was no control over TCP as described above. You can enable it, but likely you will see the uTP connections not transfering much data, because they are pushed out of the way by TCP."
This sounds like congestion control of some sort to me.
LOL! Well.. good luck, I hope she's worth it, and with any luck the wife will never find out about her. Don't forget to burn the receipts!
https://www.eff.org/https-everywhere
That is an internal problem of the small business. They rent one T1 line for 50 people, which is obviously not enough for 50 people watching stuff on YouTube. So they need some policy to prevent excessive use. On a former job of mine we actually had such an event: ;-)
One guy was running eMule from his laptop and collecting Gina Wild videos
We found him through the router management software and told him to switch his P2P application off, problem solved.
The problem with most end-user ISPs is their excessive overselling:
They have a bandwidth per user that may not be much better than in your example, but they market the connection to every single user as if he had the line for himself. Then the ISP complains or starts throttling the bandwidth, if it is actually used heavily. Users feel cheated.
My point is that it is very much the fault of the ISP if they make empty promises. It is not the duty of the customer to figure out how much capacity the ISP actually has and restrict his usage accordingly. Actually, I think the behavior of some ISPs borders on fraud and I'd like to see them lose a lawsuit over it. AFAIK there is actually one brewing, some users have sued Comcast.
C - the footgun of programming languages
I seem to remember the pundits saying that there weren't enough applications that could make use of the broadband connections and all that bandwidth was going to waste.
If someone is passing you on the right, you are an asshole for driving in the wrong lane.
p2p applications should switch to SCTP.
http://www.ietf.org/rfc/rfc2960.txt
Having to work for a living is the root of all evil.
There are roads and plumbing and electricity and the Internet. These are services.
You pay for road access (tolls and taxes) but if the road is too busy, that sucks and deal with it. There is not enough room on any roadway to handle all potential customers.
You pay for water but if everybody flushes their toilet and waters the grass and takes a shower at the same time, water becomes scarce and that sucks but you deal with it. There is not enough water throughput to deal with all potential customers claiming their purchased usage.
You pay for electricity. Too many users = brownout. That sucks, deal with it.
The internet isn't a product, but a service. Like all services, it is oversubscribed. If there is a sudden surge of demand it becomes unavailable and that sucks but just deal with it. Go outside; it's hi-def too.
If there is a continuous over-demand for the service, the providers will build more infrastructure. And that will probably mean temporary limitations.
Just like road construction, or plumbing and electrical upgrades.
Don't get me wrong, I like to "download Linux" over bittorrent like everybody else. But I don't have a Service Level Agreement with my ISP, or my city road crew, or my utilities providers. I just deal with downtime.
-Alex
How much work could a network work if a network could net work?
If my ISP starts blocking traffic I consider valid coming to me, because it happens to be torrent traffic, they deserve what they get. Regardless of what type of traffic I am getting, if I use the bandwidth for A, B , or C it does not matter how I use the traffic, I still have 1gb per month download. Don't penalize me for using one form over another, or getting all my download at once instead of over the course of the month.
I hope enough people complain that this does not become an accepted practice amongst the ISPs to block torrent. Many technologies use the torrent download system, and do so because of better performance for people using the same technologies in conjunction with torrent
Only ISPs that have oversold their bandwidth should suffer. What a shame... Now they will actually have to invest to deliver what they actually sold to their customers.
Having actually read the article, the author has his facts wrong. His characterization of the technologies (UDP v. TCP) is incorrect and... prejudicial. The use of UDP in this application will avoid large amounts of TCP overhead and largely unnecessary TCP resends - for a net reduction of traffic.
Although the article characterizes this as an attempt to dodge throttling this actually looks like a good engineering decision. I'm not part of the design team - I don't know the rationale, but it certainly seems to have a good basis in technology.
If she can't use her fat pipe, she can use mine...
Would Jesus uTorrent?
I am not a networking master, but wouldn't UDP be a better choice to avoid the TCP syn/ack packets? As many connections as a P2P program makes, changing from TCP to UDP would help alleviate some stress on routers worldwide. Although in this case, more unreliable, hash checks and pieces would help prevent corruption amongst files.
"I'm a well-wisher, in that I don't wish you any specific harm."
Sounds like another RIAA/MPAA conspiracy to get people to stop downloading via torrent...I'll believe this when i see it.
Stupidity only gets you so far, then you've gotta try
but why would you trust the programmer/user to report the time-sensitivity of a piece of data? Do we trust injured people at the hospital to triage themselves?
BT isn't the problem, it's the internet providers offering more than they can handle. If I am capped at 5Mb/s then I should be able to use that regardless of my protocol and so should everyone else. Internet providers, upgrade your infrastructure and stop over selling!
First post! (just in case I am...)
The client application can't know about congestion in the core of the network, so it breaks control at the core of the Internet where there's less management. If you read the article fully, this would have been apparent.
Real-time applications use UDP because they have built-in bandwidth caps and they don't burst to the full capacity of the network like file transfer applications. It's dangerous to put P2P on UDP on a very large scale and it will require a massive change in the core of the Internet to deal with this new behavior.
Something about more sessions to track uses more cpu and memory. I'd imagine it has to do with the tables required to track the sessions.
A "router" doesn't have session state. The only information it keeps are routing tables, and those are not dependant on the traffic being sent, but rather on the network topology.
Most home users have routers that also contain address-translating, stateful firewalls, and this is probably what you are thinking of. Too many active connections can cause such a device to stop working, but in practice even the cheapest sold today can handle somewhere between 1000 and 2000 connections.
Actually, no, it's not a good thing.
I once brought a very large and very expensive intranet to its knees by accident. Basically I was using UDP packets to transfer a few gigs of data back and forth between a master node and a bunch of slave nodes. Neither the ATM interfaces on the Solaris boxes nor the switches themselves handled this well; the former wedged completely requiring cold reboot, and the latter just froze for long periods of time. You'd think it wouldn't matter, but evidently the designers had anticipated TCP would be the bulk of the bandwidth, and optimized for it.
Using UDP to move large amounts of data around on a network you have complete control of is risky. Doing so on the public internet is a very, very bad idea.
You're just another global warming apologist. Oh SUUUURE... it's BITTORRENT causing all this melting...
"Mind, as manifested by the capacity to make choices, is to some extent present in every electron." -Freeman Dyson
Consider doing such a transfer trans-Atlantic when there's a temporary loss of capacity due to (say) a ships anchor cutting a fibre route. Suddenly your connection is congested although your ISP has not been overselling you - there's just been a loss of capacity on the network and now the backoff features no longer work. So instead of stuff just getting slower the whole system starts to break. That's what the author of TFA is afraid of - whether the actual implementation in uTorrent will do this, I can't say, but it sounds plausible.
then my torrents will download slow!
If you're going to transfer files over UDP then you need to build some TCP-like protocol on top of it. The article doesn't say exactly how BT works in this respect, but he's probably right. There's no way that BT's protocol could be as sophisticated as TCP, given its 30+ years of development.
Isn't this partly why SCTP was created? (Of course not all OSes and routers / NATs know about it, so it would only be of limited practicality in this instance.)
No, they haven't. What starts as a claim that BitTorrent has declared war on VoIP turns out to be a claim that BitTorrent has made a change to their protocols that might impact VoIP users as collateral damage - but ONLY if ISPs decide to engage in a foolish and moronic packet blocking scheme that would also impact DNS and therefore is unlikely to ever be implemented.
The article is choc-full of half truths and downright lies. The reality is that ISPs have many choices in how they "shape" user traffic, and the most obvious solutions (outside of upgrading their networks) that are fair rather than discriminatory are also the solutions that will have zero effect on VoIP users.
You can identify heavy bandwidth users and throttle their traffic. It's easy. Each packet has this thing called an IP address on it, that records where the packet is from, and another one recording where it goes to. If someone's identified as making heavy use of bandwidth at a time of peak congestion, you can use this hitherto completely unknown attribute of every IP packet to throttle traffic to and from that customer.
Easy. But it's not a knee jerk "It's BITTORRENT that's destroying out Internets!" response. So you get the Richard Bennetts of this world - who clearly have an agenda that's anti-BT rather than anti-bandwidth hogging - pretending it doesn't exist.
Idiots.
You are not alone. This is not normal. None of this is normal.
"As long as the UDP doesn't reach 176 degrees, the PVC tubes should hold up."
Surely the issue isn't using UDP to get around ISP's 'traffic management', the problem lies with the ISPs who have forced BitTorrent to use UDP with said 'traffic management' policies.
If the ISPs didn't throttle bandwidth, BitTorrent would have no need to use UDP and the world would be a better place. If the ISPs didn't complain to their customers for using the bandwidth they supply but had networks that could support their advertised bandwidth, everyone would be happy.
Unfortunately as well we all know, using 10Mbps of the 10Mbps connection you paid for is bad, you're not allowed to do that - heaven forbid you use the bandwidth you're buying from them :o
Would this not only be a problem with transfers that are running while the cut happens? I mean if I start a udp transfer after the bandwidth is cut, then uTorrent should detect the lower transfer bandwidth and set the transfer speed lower. They do not have any way to know how the bandwidth between me and an random other peer, so when a new 'connection' is opened between me and an other peer, they need to determine the bandwidth. And in case of a cut, they should detect lower bandwidth.
And I think they really do need code to do dynamic available bandwidth detection(Similary in effect to tcp), because If I open a connection in my browser that uses half my available bandwidth, things would really fuck up(Both for uTorront and my browser) if uTorrent did not lower the speed.
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.
The problem here is that he starts with a valid premise: completely unmanged uTP is a threat to the continued smooth operation of the internet, and comes to the completely unsupported conclusion that any law restricting the ISPs is bad.
ISPs need to be allowed to control bandwidth flow in a fair fashion to provide the best service possible to their clients, but this doesn't mean we should let them do whatever they want either. ISPs should, for instance, not be allowed to treat packets differently based on who they're coming from/going to, or what their content is, but some ISPs have been talking about doing exactly that.
Over regulation is bad. This doesn't make no regulation good.
Drugs are illegal, yet it's a multi-billion dollar international trade, despite many more times that amount spent on eliminating it. Why? Because there're drug users.
Vast majority of P2P traffic is probably also prohibited by various IP laws, but if enough people want something badly enough, no power in the world can prevent them for long.
Guess who has a lower tolerance for BS? ..."Lynette" whose children are screaming at her because their Wbox 3 is 'broken'?
The developers of the next generation of consoles and game software will do so in the environment created by the ISPs.
Give me one good reason why a console game *needs* the ability to send out multiple streams of UDP data at the same time. 'Cause that's what ISPs will look for -- not individual protocols. You just look for an app that *looks like* a badly behaving P2P app and throttle the entire user's stream. Well-behaved P2P apps, like Steam, will be whitelisted.
And I don't see very many mothers continuing to scream when the service rep looks at why the connection is throttled and asks if their family is using any file-sharing apps and strongly hints that the ISP doesn't support piracy of movies and music and that they should either shut off their filesharing or upgrade to a package that supports it or be faced with possible termination if they're caught violating their ToS. Little "Lynette" here isn't likely to even be AWARE of most legitimate uses for P2P, and frankly most big ISPs (the ones making the filtering decisions for the little ones that sublease their lines too) don't have to give a damn about customer service because they're one of the only two providers in town. ...And the other guys are doing the same thing.
Game, set, match for the ISPs. It's cheaper to hire low-skilled workers in India to listen to screaming moms than it is to upgrade backbone bandwidth. You don't honestly think whining to the outsourced customer support rep of the local duopoly actually *means* anything, do you?
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
Everyone here seems to work under the assumption that heavy bittorrent users would be worse off in a world with download caps or metered bandwidth. I don't think so. It would force companies to compete to give good service with clear contracts where they actually tell you what they are really selling. This is bound to increase, not decrease, the availability of real bandwidth per dollar for most users, included us computer geeks.
How?
Is this some sort of strange free market fundamentalist idea that the market will provide solutions that make everyone happy, or do you have a good clear model by which free riders on the system stand to benefit under a system that holds them accountable and is explicitly designed to discourage their behavior?
I mean, I have to accept that what I am, from the view of lying "We sell unlimited bandwidth and are happy to serve customers of all stripes!" ISPs -- a "free rider" of sorts. I pay for the absolute cheapest DSL package available in my area, and I use about a third of its maximum capacity (upstream) every month. I probably transfer 10-100x the data of the average user of their *most expensive* package, and my ISP probably sees me as someone using "more than their fair share."
What ISPs want to do by metering connection is to create a world that provides a financial disincentive to heavy users. This gives them more breathing room to slow down upgrades, lets them attract more customers (who can't afford current rates), and lets them gouge the living heck out of the those who have high demands (restoring the usual "supply v. demand" pricing curve you get elsewhere) and turn high bandwidth into a premium luxury good that you can charge increasing rates for (like they do for business customers).
In a world priced to suit low bandwidth users and to dig more money out of high bandwidth users, how do high bandwidth users benefit more? Right now I pay less than average for my connection and use far more than average. I can see policy reasons for the public as a whole maybe benefiting from such a move, but how on earth does flipping the cost equation benefit me at all?
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
I knew a girl who would be ecstatic if her man bought her some sexy underwear. She likes to flirt sometimes, ya know? Girls like to feel pretty and wanted...
Support my political activism on Patreon.
You pay for electricity by the pound, not on a monthly rate. If you want to run an ISP like a power company, then you have to accept a rate plan based on how much you use, rather than based on how much you want access to.
The disconnect with ISPs is that they are overselling because they're selling access - except they can't deliver on it certain peak situations. For the vast though shrinking majority of their customers, the maximum access potential will never be reached. It's those few times, people and places that actually hit that max that are causing them problems - because in those cases they've sold a product they can't deliver.
Unfortunately, without changing to a pay-as-you-go scheme, they don't have many tools for reducing the problem users; forcing them to quit, suing them, or trying to ignore them being the obvious ones. If they raised their rates to match the power user use, they'd lose market out of the bottom bunch of people. Who is going to pay $100 a month to check their email at home? No one, especially with cell phones covering that these days.
Telecoms don't WANT to move to a pay as you go plan, though, because they're making a lot of money right now off of those customers who don't use the service much. Ideally, all their customers would be like that - paying more for less - and they will want to manipulate the market in that direction.
And that is why regulation is needed.
[Ego]out
The basis of your argument for why complaints won't mean diddly ("the company doesn't care") is the exact reason why this would blow up in their face.
Because the company isn't going to care enough too selectively whitelist services or even differentiate between 'well behaved' and 'ill mannered'. You know they aren't going to, because they aren't right now.
And because they aren't differentiating, they will catch enough people in their dragnets that they will manage to piss off the wrong people and pay for it.
Have you noticed lately how badly cable companies are getting their lunches eaten by the satellite companies or the phone companies having their lunches eaten by VOIP and cable providers? These are groups that used to have a monopoly in their area and are now fighting off fairly strong competition because they got complacent and arrogant enough to not care what the customer thought of them.
No, I'm not saying that they are quaking in their boots right now, but the one thing I am certain of is that the genie is already out of the bottle. People are ramping up their internet usage and they aren't going to be willing to settle for less rather than more. If the current providers can't provide service, someone else will. Maybe that'll be a direct competitor or maybe it'll be something completely different. But it will be something, because as powerful as the telco and cable companies are, they aren't powerful enough to shut out everyone.
BitTorrent is believed to be harboring weapons of mass destruction. These weapons are believed to be capable of destroying all of the internet tubes.
Maybe that's why the government has has deployed 4,700 troops domestically, ramping to 20,000 over three years, trained to respond to "weapons of mass destruction attacks".
Think that will be enough?
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
How *dare* you attempt to impugn the liberal practice of "shoot the messenger"!? He *allowed* the Messiah an opportunity to say something that revealed his true thoughts and beliefs!! Burning at the stake is too good for such a heretic!!
I'm so torn! I could go one of two ways...
The High Road Reply:
"You know, you'd be a bit more credible if you simply noted that partisans of all stripes play the 'shoot the messenger' game, and portraying it as a 'liberal practice' only reveals your own short-sighted, partisan bias. You could have pointed out that the whole 'is he or is he not a plumber because of a union-backed licensing program' was a silly argument to begin with by people desperately grasping at straws to discredit and already non-credible messenger that the McCain camp looked silly embracing in the first place when they could have just let the guy flame out on his own."
The Low Road Reply:
"At least his wife's cover wasn't blown, severely damaging her career, (arguably treasonously) endangering undercover agents, and destroying our line of info on Iran's nuclear program over an article calling BS on an obvious lie. 'Cause that's the kind of stupid thing you'd never expect security-conscious conservatives to do!"
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
Something about more sessions to track uses more cpu and memory. I'd imagine it has to do with the tables required to track the sessions.
Routers don't track sessions, so more sessions has no effect on CPU and memory.
Network Address Translators have to track sessions, but those are at the edges of the network. Network hosts (i.e. your PC and the one you're pulling bittorrent data from) have to track sessions, but again those are at the edges of the network.
Your ISP's routers only have to track sessions if they're doing some sort of session-based QoS. Few ISPs do that, and none should.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
So, in other words, people complaining about BitTorrent users overutilizing the network should read their contract, see there's no minimum guarantee of service or line speed, and get stuffed instead of trying to bully other people into using the network in ways that would make life more convenient for them?
Despite being a heavy BT user, I can't feel anything but horror at that line of thought because that line of thinking is behind every single tragedy of the commons situation EVER.
"There's no law saying that I can't."
"It's my land, so I'll do what I want with it! Don't you tell me what to do with it."
"I don't care about people downstream; I'll do what I see fit to do with MY water."
Et cetera. I mean, that's the situation as stands, but actually encouraging that line of thinking only encourages those on the short end of the stick to consume as much as possible to "get theirs." Pretty soon, it sucks for everybody.
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
UDP senders don't know when they're flooding the connection, so they just keep doing it. A TCP stream would back off to try to be fair, but your UDP stream is just going to keep on blasting at full speed.
If that's the way that it's implemented.
But UDP is just a pass-through of the underlying IP transport. You can build anything you want on top of it - including recreating as much of TCP (or an equivalent tuned better to your own usage patterns) as you want.
However: Suppose they DO just ignore flow control and hammer away: This is being done as a workaround for ISPs who throttle Bittorrent traffic. So it looks to me like the ISPs just brought it on themselves.
uTorrent looks to me like the users' nuclear option - a threat that creates the incentive to abandon war and come to some peaceful arrangement.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
After you sold it to me, it's simply criminal to then say
No, that would be a civil lawsuit.
(ianal, tinla, sorry to nitpick)
I'm sorry but the "Bittorrent (or video, or pr0n) is going to kill the internet" BS is touted in the US about as often as Wall Street Hit's up Washington for their allowance. I don't mean to troll here but honestly I'm growing weary of the FUD.
To the ISP's of the world. If your tubes fill up, Build/Buy/Design bigger ones. (Oh and I don't want to hear about how large the US is, Canada is larger has more space between cities and does a better job of getting true broadband to people than we do.)
Perhaps we would be better served by trying to find ways to accommodate the uses of the net, instead of spreading fear, and sticking our heads in the sand, while the rest of the world passes us by.
I'm sorry, I'm to tired to be witty at the moment so this message will have to do.
If my ISP is selling me unlimited internet and they decide not to deliver, I want a rate cut.
I think you should argue that you should pay for the share you're getting. If you get bandwidth limited by a constant k, but paid for unlimited bandwidth, it would only be fair to pay a limited part of an unbound whole:
lim_{n \rightarrow \infty} \frac{k}{n}
(i.e. 0)
Lies! You know a girl!?
Belief? Hope? Preference?The Existential Vortex
Good utility analogy. It's time to stop pretending bandwidth is not a utility.
I thought that UDP had less overhead than TCP, so how is it worse, exactly, the article is very vague. If the ISPs Can't supply the bandwidth they have on the label, then its there fault, not the application developers fault, that the rest of the users cannot access the internet? Sure they will need to use more money, but it wouldn't hurt anyone except the CEO if they invested more into bandwidth.
O.o
BitTorrent claims it is actually trying to reduce congestion.
By "reducing the value of bandwidth", do they mean "cutting into our profit margins by actually using the bandwidth that we only promise because we are sure you won't use it"?
Because screw that.
uTP was explicitly stated -- by its developers -- to be an end run around providers' reasonable network management practices, and Richard is absolutely correct when he notes that it could cause severe network problems. In fact, Ricard understates the case, because he neglects to mention one extremely important point. By switching to UDP, BitTorrent will not only compete with VoIP and some video and audio applications but also with DNS. This could well be catastrophic, because DNS (domain name service), as ISPs know all too well, is a "critical path" protocol in virtually every application. If DNS is slow, EVERYTHING ELSE that users do will also be slow. Remember, most network applications, including Web browsers, have to stop and wait -- unable to do anything else -- until they resolve one or more domain names. So, they'll hang frustratingly if DNS packets are dropped due to congestion. And what underlying transport protocol does DNS use by default? UDP. (It can use TCP as well; however, it does so if, and only if, it has a lot of data to transfer. And TCP, due to its complex handshaking and "slow start" flow control, is much less efficient and much slower.) So, what we're talking about is not just congestion but sand in the gears of the entire Internet. Also, because uTP does not conform to any explicit congestion management protocol that could detect congestion BEFORE packets are dropped, the only way it would be able to detect congestion in the network would be after packets were dropped. Which means that by the time it did anything -- IF it did anything -- to mitigate the congestion it caused, it already would have damaged the network. Finally, do you actually trust P2Pers -- who already, in the vast majority of cases, are brazenly engaging in illegal activity -- to be courteous to anyone? There's no honor among thieves, folks. YMMV, but personally I wouldn't want to be on the same cable segment with someone using this new version of BitTorrent.
The ISPs started this war, I say bring it on.
An SQL query goes to a bar, walks up to a table and asks, "Mind if I join you?"
Using UDP in BitTorrent can't make that much of a difference.
It's only natural to move BitTorrent to UDP, considering that the bulk file transfers don't have the "session" nature, and also that Bittorrent already has to deal with transmission errors. Let's cut the fat...
The main problem with TCP is managing the memory of the non-ACKed transmitted packets, and the flow control is intended more to share the machine's connection with other TCP connections, than anything else. TCP doesn't know about the Internet, about "congestions"... The "Congestion control" of TCP Reno, Vegas, etc, is more of a kludge than can and should be enhanced. Using UDP we can even create more conservative policies to prevent congestions.
The bandwidth available to each machine is controlled by the gateways and routers, by the ISPs between the bitTorrent and Skype users, deciding when to drop the packets. The way routers can decide that, and consequently implement a certain QoS to its clients, does not depend on TCP or UDP use, the packets will be allowed and dropped just the same.
Now think about this: if UDP had this problem, of taking up band irrestrictedly, it would not just cause havoc on the Internet, it would first cause problems to the other UDP connections in the machine itself!! If UDP really had no flow control at all, it would consume your own machine's bandwidth before causing a "congestion" problem way up in the network. Last time I checked, it's the OS who tells the programs when it is ready to receive a new package, and never the other way around. There is sure SOME kind of limitation to UDP in any OS, even if it's nothing as sophisticated as the Renos around.
Flow control is not "congestion" control. What TCP has is congestion fear, and not control. It doesn't control it's fear yet... We need extra protocols to handle QoS properly, for the routers to tell the clients when are congestions going on the network... IPv6, MPLS and RSVP now!!! The ISPs need to put their money on these new technologies, and not blame the users to use the network exactly the way they are supposed to: just using as much as they can. The routers can drop the packets whenever they like, and that's they way TCP/IP works... Enforcing QoS up in the middle of the network, and not asking the users please to hold on.
UDP and IP multicasting go well together for transmitting the same data to many hosts ... couldn't it help reduce the torrent traffic after all?
The number of TCP sessions *makes no difference at all* to an ISP - assuming they are not doing deep packet inspection of course. Apart from that case, all IP packets are routed, with no inspection of the TCP header, so there's no way the number of sessions could make a difference.
If you want to use no. of TCP sessions as a weak indicator of P2P activity, go ahead, but BitTorrent developers will develop workarounds (maybe this is one of them) - ultimately everything may end up as encrypted UDP traffic that's really hard to traffic shape even using DPI.
I generally agree that you get what you pay for - there are shared pipes all over the place, so if people use BT 24/7 and fill their DSL/cable access link, they should expect to pay more than people who simply do email, surfing and occasional video clips. Traffic shaping and usage caps are just different approaches to making behaviour fit the costs of running the network.
Yes, I do have 10,000 sheep. But this land is common, I have a right to let them all graze there if I want to.
Sorry, your internet connection is part of the commons? Mine is paid for privately, and I can graze all the fuck I want on the paddock that I rented FOR GRAZING.
shame the ones I know dont dress like they do. Lowest common denominator -- what ever gets them by.
Semi-automatic amateur armchair Australian philosopher; conjecture ready at any moment...
Bandwidth is and always has been priced according to average usage. No ISP in the world has ever had enough transit to cover the full bandwidth of even 25% of it's customer's total combined link bandwidth, and that doesn't mean they're ripping off the customer; it's simply indicative of the fact that customers do not need to be filling their link all the time in order to find it valuable.
That said, the sky is not falling. If bandwidth demand goes up, ISPs will (and do) deal with by providing two options:
1. Maximum total monthly downloads with per-Mb charging or a hard cut-off thereafter.
2. Higher prices for unrestricted connections.
Bandwidth gets cheaper every year and does so more quickly in the face of high demand, so whichever you choose it shouldn't be too painful.
... not real time. So at least 37 people posted before you and that's as far as the moderators had got since publication. Sometimes I have to wait overnight to see how others have responded.
I'm sorry if I haven't offended anyone
will happily disconnect bittorrent using customers, as thats the most profitable path to take. For people like me, I'll simply lose internet due to lack of competition (that is the result of the free market in my location).
All the isps need to do is to periodically disconnect the top one percent of customers based on bandwidth use.
There's no way ISPs are selling more (how is it called? Link width? Connection speed?) than they have, is there?
The technological solution has been there for 10 years already, the ISPs/vendors are just too lazy to utilize it. It's called DiffServ.
All you need is to classify your traffic into 3 classes and use separate queues for them, with strict priorities:
- High Prio - <= 1% of the end user's theoretical maxed out traffic
- Medium - <= 10% of the theoretical traffic
- Low Prio - the rest
Games & VoIP would go into High Prio (Low Delay).
Anything unqualified (Web browsing, e-mail, etc.) would be Medium (good enough for Interactive traffic)
BitTorrent and big downloads would be Low Prio (Bulk).
The traffic itself can be classified by the applications on the socket interface.
You would be surprised to hear that some of it already happened even before DiffServ - eg. Telnet/SSH on Linux labeled it's traffic as Low Delay before ages.
Anyway, the max. percentages can be easily enforced by re-labeling any user's traffic to a lower class if it exceeds the percentage on a daily/monthly average basis. So the end-users and application software vendors clearly have the incentive to comply and not to abuse the system.
That's it, and you have the fastest possible delay for gaming, reasonable performance for browsing, and you can utilize the surplus resources for bulk transfers at virtually no effect to the other traffic.
It's all old, available and proven technology, just start using it please...
All it needs is ISPs wanting to cooperate instead of negligence.
If we used Explicit Congestion Notification on top of it (also a technology that has been available for 7 years), there would be almost no congestion at all. All the pipes would be fully utilized, yes, but without congestion.
Actually, instead of a technological solution a rather easy and fair business solution is also available - instead of capping the already existing unlimited users, offer a discount to the users who volunteer to be capped.
You're just making that up, and it makes no sense. The "core of the Internet" can regulate TCP traffic and UDP traffic precisely the same way, by dropping packets. And that's exactly what it can and will do with no changes.
The "core of the network" treats each packet as its own universe. It looks at the source, destination, and perhaps packet size and type of service, and that's it. It doesn't care if it's TCP or UDP.
I can assure you I am not making it up. Routers and other network devices typically don't drop UDP packets because UDP end points don't respond like TCP end points which halve their bandwidth whenever a packet isn't delivered. It's also hasn't been necessary to drop UDP packets because traditional UDP applications are typically small bursts of data or low/fixed bandwidth applications like VoIP and online gaming. Now this is all about to change when you get a huge bulk transfer application like BitTorrent which accounts for a significant portion of the Internet's traffic.
"Routers and other network devices typically don't drop UDP packets because UDP end points don't respond like TCP end points which halve their bandwidth whenever a packet isn't delivered."
That's simply not true. Routers and other network devices typically treat UDP and TCP precisely the same.
"It's also hasn't been necessary to drop UDP packets because traditional UDP applications are typically small bursts of data or low/fixed bandwidth applications like VoIP and online gaming."
Nevertheless, they've dropped them the same way they've dropped TCP packets. They do this simply because there has not been (and still is not) any good reason to treat TCP packets differently from UDP packets.
"Now this is all about to change when you get a huge bulk transfer application like BitTorrent which accounts for a significant portion of the Internet's traffic."
Nothing will change. The traffic will be UDP instead of TCP, and routers won't care one way or the other, just as they never have.
The only devices that will be affected are unusual "invasive" devices, such as those specifically used to throttle P2P.
No, routers do not treat UDP and TCP the same. Richard Bennett is a network architect and I trust his assessment of the situation more than I trust your's.
Richard Bennett never said that Internet core routers do not treat UDP and TCP the same. You are misunderstanding what he is saying.
I am not asking you to trust me over Richard Bennett. I am asking you to not misrepresent Richard Bennett's argument.
*Your* statement -- "The client application can't know about congestion in the core of the network, so it breaks control at the core of the Internet where there's less management. If you read the article fully, this would have been apparent." -- is a misrepresentation of Bennett's argument. In fact, the client can know about congestion in the core of the network with UDP the same way it does with TCP, by inferring congestion when there is packet loss. TCP's congestion control is implemented by the endpoints, not the network core. (With the exception of ECN, but that's so rarely used that it doesn't really matter.)
Yes, he has said that UDP is treated differently than TCP, and it makes sense to treat them differently. I know the man we have spoken by phone and in person. It makes no sense to drop UDP packets since it doesn't produce a response from the end-points.
The reason you drop TCP or UDP packets is not to produce a response from the end-points. It's because the link is congested and you simply can't carry all the traffic. Well-behaved applications, both TCP and UDP, respond to packet loss by reducing their bandwidth consumption.
"I know the man we have spoken by phone and in person."
Well, you've now set the bar impossibly high. In order to satisfy you, I have to refute a claim to which I have no access based on evidence and argument to which I have no access.
The truth is, so long as a protocol layered on top of UDP detects congestion and responds with appropriate backoff, it will be just as Internet core friendly as TCP.
"The reason you drop TCP or UDP packets is not to produce a response from the end-points. It's because the link is congested and you simply can't carry all the traffic"
You do it for both reasons, but the FIRST reason you do it is to get the end-points to voluntarily backoff first before you reach a point of crisis. That's the whole point of Random Early Detection which is to get the clients to backoff before the network gets completely saturated in which case really nasty things happen. Routers drop packets before they reach 100% congestion not to trim the bandwidth; but to get end-points to back down.
There's another good reason to do this because it balances out the load between single-flow TCP applications. When a TCP end-point backs off to 50%, it's giving a new TCP flow a chance to take up the slack and speed up until the two TCP flows reach a state of equilibrium. UDP end points lack this behavior.
As for Richard Bennett's claim that UDP isn't generally dropped by routers, this is the reason he is concerned about this change in BitTorrent. This will force routers to start dropping UDP packets just like TCP packets and it will have a bad affect on other UDP applications that have a good reason to use UDP e.g., they're low/fixed bandwidth or they only send short/small bursts of data which means they benefit by bypassing the overhead of TCP. Bulk file transfer applications shouldn't be bypassing TCP for UDP and even if BitTorrent is well intentioned, such a large change could be risky for the Internet. That's the whole point Bennett is trying to make.
"There's another good reason to do this because it balances out the load between single-flow TCP applications. When a TCP end-point backs off to 50%, it's giving a new TCP flow a chance to take up the slack and speed up until the two TCP flows reach a state of equilibrium. UDP end points lack this behavior."
You have this completely backwards. TCP endpoints can only detect congestion through packet loss on the connection. UDP endpoints are free to back off for *any* reason, including but not limited to packet loss.
Because every TCP connection is basically independent, you cannot use information gained from one connection to backoff on another. With UDP you can.
A UDP application can backoff as aggressively as TCP, less aggressively, or more aggressively. Nobody yet knows how BT will rig its backoff algorithm.
They've hinted that it will backoff even more aggressively than TCP does, using congestion information from one connection to increase how severely they treat congestion information from other connections.
They've also hinted that they'll rig BitTorrent to act *in* *total* much like as single TCP connection, rather than like hundreds. This is good for the user, as using BT won't make his web pages and games get 1/100th of his bandwidth but 1/2, like they should. But it's also good for Internet's core, as one BT client looks more like one connection than hundreds there too.
"As for Richard Bennett's claim that UDP isn't generally dropped by routers, ..."
I don't even see where Bennett claimed that. And I've counter-claimed that routers generally treat UDP and TCP precisely the same. I believe this is a mis-statement of Bennett's claim. Can you cite any public claim of his to this effect?
I think Bennett was simply assuming that their motive was evil -- to get around congestion control -- rather than benign -- to fix a defect in TCP that it treats each connection as the distribution unit rather than each application. (Which is bad when an application with 100 connections tries to live on the same connection as one with two.)
"'There's another good reason to do this because it balances out the load between single-flow TCP applications. When a TCP end-point backs off to 50%, it's giving a new TCP flow a chance to take up the slack and speed up until the two TCP flows reach a state of equilibrium. UDP end points lack this behavior.'
You have this completely backwards. TCP endpoints can only detect congestion through packet loss on the connection. UDP endpoints are free to back off for *any* reason, including but not limited to packet loss."
Looks like you misread. I said routers drop packets (before total congestion), TCP endpoints detect and react by cutting speed in half which gives slower TCP streams a chance to rise to equilibrium.
"Because every TCP connection is basically independent, you cannot use information gained from one connection to backoff on another. With UDP you can. A UDP application can backoff as aggressively as TCP, less aggressively, or more aggressively. Nobody yet knows how BT will rig its backoff algorithm."
Yes it can backoff; if the routers dropped UDP packets the same way they dropped TCP packets. But they don't, so there's your problem.
"Looks like you misread. I said routers drop packets (before total congestion), TCP endpoints detect and react by cutting speed in half which gives slower TCP streams a chance to rise to equilibrium."
I understand that, and I agree with this. However, this says nothing whatsoever about UDP. In this case, the exact same thing will happen with UDP. The routers will drop the packets early (because they do not treat UDP differently from TCP) and the UDP endpoints will backoff (because they will be programmed to do so).
"Yes it can backoff; if the routers dropped UDP packets the same way they dropped TCP packets. But they don't, so there's your problem."
Yes, they do. At least, they should. RED is not as effective for UDP as it is for TCP (as many UDP flows are not responsive) but that is not a good reason to exempt UDP from RED just because it might not be responsive.
If there are people out there who exempt UDP from RED, independent of any QoS indication that the packets are precious and not in limited internal networks where it is known that UDP traffic is precious, they should fix their networks.
This article and many like it are just attempts to distract attention from a more fundamental issue. Why is Moore's law essentially almost null and void when it comes to provisioning of internet bandwidth? How do things need to be organized so we get the same sort of benefits in this area as we get in such diverse areas as printer performance, large flat screen monitors, processor speed and capability, memory cost, hard drive capacity, etc?
The only area of this part of the economy that seems to be mystified by the idea of Moore's law are the monopoly and duopoly ISP's. Qwest proudly proclaimed they would guarantee a certain level of bandwidth without price increases for some number of years. Are they kidding? Their costs keep dropping toward zero per bit and they think they are being generous because they will restrain themselves by not increasing prices! What a load of crap.
Nobody thinks bandwidth is cost free (see below for some details). What is clear to anyone involved in digital electronics for the past fifty years is that except for monopolies and government interference the price for a given level of service will always be dropping precipitously. The fact that it isn't for internet access is evidence that there is something rotten going on.
About that cost free issue it is worth noting that even for things that are not free, like roads, it does not make sense to make every road a toll road. The same sort of idea may make sense for a modest and increasing level of bandwidth. The FCC will be considering exactly this topic in the near future. The free lunch for fat cat monopolists may be in some danger.