-
Subscribe -
Community
-
Top Commenters
-
Popular Threads
-
Recent Comments
- I see that Coinstar now also offers vouchers for Starbucks or iTunes. How do you redeem the voucher with Starbucks? Do they give you a Starbucks card? Regardless, I've got a lot of coffee...
- Dude, I'm just joking with you Applebees. Chill. And yes, the stylus is soooo 1994. But some things from the past are worth preserving.
- Steve: Great comment, to which I would add just a couple of points. First, to the extent that a corporation that may provide the "pipes" is separate from one that provides the content,...
- My top 5: 1. <a href="http://ww.autohotkey.com/">AutoHotKey</a> - This is a simple yet powerful macro scripting program for Windows. I created a simple one-line script...
- "Don’t understand how you iPhone zombies get along without one." It does not help your argument any when you insist on insulting users of another device. You arguments come off as...
DISQUS
Returning? Login
1 year ago
What about someone who uses his IM client's file sharing facility to transmit a file to other user? Makes a phone call using Skype? Watches TV with Joost? Runs World of Warcraft, which downloads patches using BitTorrent?
"No servers" might have been a reasonable rule back when most people just did web surfing and email. But these days a lot of ordinary consumer applications are technically "servers." So while Comcast may technically be within the letter of their TOS, I think the way they're interpreting the TOS richly deserves criticism.
1 year ago
That's basic network engineering and should be common sense among the network-literate among us.
That being said, we know what a server is: it's any application that continually moves massive amounts of data on a timeline that stretches from minutes to hours. So if we take the classical definition of a server and extract its traffic profile, we can update the definition to cover modern applications that have the same network impact without unfairly implicating modern applications that don't.
For example, the traffic profile of Skype is massively different from that of BitTorrent, but in some sense they're both "peer-to-peer." (The confusion of these two applications is deliberate, because file-sharing buffs don't want their favorite piece of code called "a piracy tool with a few legitimate applications.")
Skype transmits a very small amount of data - a stream of 64Kb/s, more or less, plus a little bit of routed traffic. BitTorrent, on the other hand, will transmit and receive as much as the pipe can handle. Last weekend I ran a BitTorrent download of Fedora 8 on the Comcast network at 4 Mb/s for an hour and a half. So from the traffic standpoint, no two applications could be more different.
The basic problem that any network provider has is the management of peak loads. During these periods, it's necessary to reduce traffic below the network's stability point, and at other times little or no management is required. So how is a network provider to state its policy of peak-load handling in such a way that a regulation-happy neutrino can't make demagogic hay from it?
That's the $17 billion question about NN.
1 year ago
It is not at all clear that P2P constitutes a 'server'
First, let's use Comcasts own FAQS. Looking today (November 16, 2007) at Comcasts own FAQS posted on their website:
FAQs / Product Information / Comcast High-Speed Internet / Connection
Do you block access to peer-to-peer applications like BitTorrent?
No. We do not block access to any Web site or applications, including BitTorrent. Our customers use the Internet for downloading and uploading files, watching movies and videos, streaming music, sharing digital photos, accessing numerous peer-to-peer sites, VOIP applications like Vonage, and thousands of other applications online.
http://www.comcast.com/Customers/FAQ/FaqDetails...
So, first step to not allow people to confront Comcast with their lies would be to : stop lying, Comcast.
It is interesting that a simple call to stop a large corporation from commiting fraud, and interfering with the rights of natural people to freely associate can be called making "demagogic hay"...
1 year ago
1 year ago
I'm not sure that that's a given. Who are you to be the arbiter of what 'deserves' bandwidth and what doesn't? And in any case, you certainly can't conclude that any and all disruption of high volume traffic is reasonable from those premises. By the way, I don't know of any ISP that throttles or blocks that Fedora Linux download if it's done via HTTP, although it uses the same bandwidth. Video streaming over the web also uses a lot of bandwidth. I defy an ISP to throttle, or block (or 'delay' as you like to call it) web video streaming in the same way as they do for Bittorrent and try to convince its customers of this 'common sense'.
That is the most absurd definition of 'server' I have ever come across. As a counterexample, (personal) mail servers typically don't generate much traffic.
So does HTTP, FTP, and every other protocol I know of. You have repeatedly tried to paint Bittorrent as some ill-behaved network application because 'it tries to uses as much as the pipe as possible' but this is not some unusual feature, or even an undesirable one in general.
1 year ago
[As in "What's "torture", huh, huh, huh? Waterboarding? If I stick your head under a faucet for 30 seconds, is that torture? How about 15 seconds, HUH, what about that? There's no permanment damage, so that doesn't count, right, right, right? And didn't you ever go swimming, IS THAT TORTURE?!!! ..."]
Near-drowning people is torture.
Transmitting files for hours is a server.
1 year ago
Saying this:
Do you block access to peer-to-peer applications like BitTorrent?
No. We do not block access to any Web site or applications, including BitTorrent. Our customers use the Internet for downloading and uploading files, watching movies and videos, streaming music, sharing digital photos, accessing numerous peer-to-peer sites, VOIP applications like Vonage, and thousands of other applications online.
And then, actually blocking Bit torrent traffic is lying, and fraudulent.
As Tim Lee has stated above, above many applications behave like servers, and to indicated that bandwidth usage defines a server is to use a the word 'server' in a non-standard way.
In any case, Seth, please address this question: Is Comcast's FAQ, still listed on their website as of November 16, 2007 truthful or not?
Given that customers may have entered into contracts based in part on the representations made in their FAQS, shouldn't those customers have a claim against Comcast?
1 year ago
What Comcast does do, and what all ISPs do in some way, is limit the amount of bandwidth for BitTorrent (or other similar style) uploads. Comcast in particular does this because their network would otherwise become the principal source of BitTorrent uploads for the entire Internet, and customers who use other applications would find their performance suffering.
BitTorrent throttling is an example of John Stuart Mill's dictum that morality consists of achieving the greatest good for the greatest number.
Tim Lee's statemtent above simply shows that if one defines "server" broadly enough, all things are servers. That's simply sophism and hence not very useful toward the proper discussion of network management policy.
1 year ago
A person of common sense.
1 year ago
Yes. What Richard Bennett said. Comcast does not prevent use of BitTorrent. They try to traffic-manage it so as to keep down bandwidth and apparently to keep the transfers within their network. This is a good thing.
Comcast does not enforce the no-servers provision in their contract with an extreme, draconian interpretation of what is a server. In fact, the opposite, they seem to only take action with applications which cause serious network problems. This is also a good thing.
Bandwidth decisions have to be made, or else the most resource-intensive will crowd out everything else. And that would be a bad thing (for everyone except the users of the resource-intensive applications, and maybe not even them if they get into a tragedy-of-the-commons).
1 year ago
I will not hold myself out as someone with extensive technical background in this area, but I would not that many how do have extensive technical competency disagree with your and Richard's opinions as stated above.
In particular I would refer you both to Ed Felten's post on this issue, which says in part:
For starters, Comcast’s measures are not aimed at heavy users but rather at users of certain protocols such as BitTorrent. And not even all users of BitTorrent are targeted, but only those who use BitTorrent in a particular way: uploading a file to non-Comcast users while not simultaneously downloading parts of the same file. (In BitTorrent jargon, this is called “seeding”.) To get an idea of how odd this is, consider that an uploader who is experiencing blocking can apparently avoid the blocking by adding some download traffic.
It would likely be easier for Comcast to simply measure how much traffic each user is generating and drop the heaviest users’ packets, or just to discard packets at random (a tactic that falls most heavily on those who send and receive the most packets).
Beyond its choice of what to block, Comcast is using an unusual and nonstandard form of blocking.
There are well-established mechanisms for dealing with traffic congestion on the Internet. Networks are supposed to respond to congestion by dropping packets; endpoint computers notice that their packets are being dropped and respond by slowing their transmissions, thus relieving the congestion. The idea sounds simple, but getting the details right, so that the endpoints slow down just enough but not too much, and the network responds quickly to changes in traffic level but doesn’t overreact, required some very clever, subtle engineering.
What Comcast is doing instead is to cut off connections by sending forged TCP Reset packets to the endpoints. Reset packets are supposed to be used by one endpoint to tell the other endpoint that an unexplained, unrecoverable error has occurred and therefore communication cannot continue. Comcast’s equipment (apparently made by a company called Sandvine) seems to send both endpoints a Reset packet, purporting to come from the other endpoint, which causes both endpoints to break the connection. Doing this is a violation of the TCP protocol, which has at least two ill effects: it bypasses TCP’s well-engineered mechanisms for handling congestion, and it erodes the usefulness of Reset packets as true indicators of error.
1 year ago
I will not hold myself out as someone with extensive technical background in this area, but I would not that many how do have extensive technical competency disagree with your and Richard's opinions as stated above.
In particular I would refer you both to Ed Felten's post on this issue, which says in part:
For starters, Comcast’s measures are not aimed at heavy users but rather at users of certain protocols such as BitTorrent. And not even all users of BitTorrent are targeted, but only those who use BitTorrent in a particular way: uploading a file to non-Comcast users while not simultaneously downloading parts of the same file. (In BitTorrent jargon, this is called “seeding”.) To get an idea of how odd this is, consider that an uploader who is experiencing blocking can apparently avoid the blocking by adding some download traffic.
It would likely be easier for Comcast to simply measure how much traffic each user is generating and drop the heaviest users’ packets, or just to discard packets at random (a tactic that falls most heavily on those who send and receive the most packets).
Beyond its choice of what to block, Comcast is using an unusual and nonstandard form of blocking.
There are well-established mechanisms for dealing with traffic congestion on the Internet. Networks are supposed to respond to congestion by dropping packets; endpoint computers notice that their packets are being dropped and respond by slowing their transmissions, thus relieving the congestion. The idea sounds simple, but getting the details right, so that the endpoints slow down just enough but not too much, and the network responds quickly to changes in traffic level but doesn’t overreact, required some very clever, subtle engineering.
What Comcast is doing instead is to cut off connections by sending forged TCP Reset packets to the endpoints. Reset packets are supposed to be used by one endpoint to tell the other endpoint that an unexplained, unrecoverable error has occurred and therefore communication cannot continue. Comcast’s equipment (apparently made by a company called Sandvine) seems to send both endpoints a Reset packet, purporting to come from the other endpoint, which causes both endpoints to break the connection. Doing this is a violation of the TCP protocol, which has at least two ill effects: it bypasses TCP’s well-engineered mechanisms for handling congestion, and it erodes the usefulness of Reset packets as true indicators of error.
1 year ago
Some of this comes from the fact that PR people aren't network engineers. So if you ask them a technical question, they're going to read from a script and not be able to answer it. To be sure, they have a bunker mentality - but really, given the politics, can you blame them?
If we're quibbling over the details of the method being used by Comcast, that doesn't strike me as much of an objection.
1 year ago
You're right that if they do block a protocol like BitTorrent they'd be within the letter of their TOS. However, as Tim points out, they're being pretty selective about what parts of their TOS they enforce. They need to be more transparent about what they're doing. I don't worry about it, though, because consumers won't stand for a company that hides behind a technicality. -JB
1 year ago
Stripped from all the playing of politics here, it seems like a very simple and reasonable action - BitTorrent users are a big problem, despite a few legit exchanges the vast majority are infringement, it's against the TOS, throttling them down keeps them from ruining the network performance for everyone else.
The supergeek-type complaint, that they must construct a perfect solution that is protocol-independent, is making the perfect the enemy of the good. Sometime theoretically imperfect solutions are deployed, if they get the job done in practice.
1 year ago
To address your second point, about selective enforcement of the server ban, I think that many of the commenters are right when saying that it's like comparing apples and oranges. Bringing up examples like Skype, while technically accurate, is really like comparing apples to 10,000 oranges considering how much bandwidth BitTorrent can suck up.
However, just because this is a practical difference does not mean that it passes any sort of legal test. Tim and others aren't pointing out the selective enforcement issue because of uber-geekiness, but rather because we value enforcing rules uniformly.
While BitTorrent and Skype are worlds apart, somewhere in between is a line that Comcast is drawing and that needs to be spelled out so that users can predict what their experience will be like before they sign-up.
1 year ago
If Comcast is only limiting seeders, does that constitute “blocking” Bittorrent? Most Comcast users including myself have no problems downloading torrents. Sandvine is noticeable only when the client shifts to seeding mode. Further analysis of Sandvine is needed, and legal question of whether Sandvine constitutes blocking Bittorrent must be answered. If Comcast’s FAQ and public statements are found to be deceptive, Comcast be forced to pay retribution. But I suspect Comcast will get off on a technicality.
2. Cord is correct that Comcast’s terms of service do prohibit servers. Tim makes a good point that banning servers is absurd in an age where myriad widely used protocols rely on peer-to-peer connections. But Comcast rarely enforces its server ban, so the ToS clause is better interpreted as stating “Comcast reserves the right to limit usage of server-based protocols that place a burden on the network” Selective enforcement of a contract, however tyrannical, doesn’t negate its validity.
Richard’s definition of a server makes sense to an ISP network engineer; the reason servers are banned is because their traffic patterns place a burden on networks designed for residential use. But from a legal standpoint, the technical definition of a server has nothing to do with the amount of bandwidth generated. A better definition of a server is a connection that accepts and transmits data to unsolicited inbound connections.
Bittorrent’s appetite for bandwidth is virtually insatiable—unlike HTTP traffic, which saturates your internet connection only until the download or web page has finished loading. But Bittorrent, in seeding mode, will keep on uploading and accepting inbound connections as long as peers are out there.
Comcast has made a calculated business decision that the benefits of restricting seeding to relieve network congestion outweigh the costs of angering users who seed regularly. Presumably Comcast did not foresee the magnitude of the hellish blogosphere response or the widespread negative media coverage. But, at least for now, Sandvine is here to stay.
3. Most technophiles would prefer if Comcast dealt with node saturation in a protocol-neutral manner, instead implementing a system for controlling or pricing bandwidth usage. Comcast execs are understandably hesitant to deploy a metered pricing system (which Comcast is rumored to have fully prepared) because no major American residential broadband ISP has ever priced bandwidth. Some cable providers like Cox have explicitly stated bandwidth caps but these caps don’t discriminate against peak usage or upstream traffic which is more likely to impact user experiences.
Comcast already terminates users who generate excessive bandwidth, which Comcast reportedly considers to be anywhere from 90GB to 500GB per month (probably closer to 200GB). But while one user may generate 50GB of HTTP traffic during off-peak hours each month, another may generate 50GB monthly of seeding traffic during peak hours. The latter user places a far greater strain on local nodes and is much more likely to impact experiences of other users.
Consumers are still entrenched in the all-you-can-eat unlimited mindset from the days of unlimited AOL dial-up in the mid 1990s. And considering its unlikely any DSL provider would resort to metered pricing, for Comcast to be the first U.S. ISP to charge residential customers for overages would be a risky business maneuver. It’s not hard to envision an AT&T; ad campaign touting its DSL services as having no preset bandwidth limits.
4. enigma_foundry criticized Comcast for using Sandvine instead of conventional traffic shaping which ISPs normally use for network management. But Richard Bennett has argued that, because of technical limitations in DOCSIS 1.1, conventional traffic shaping is not a workable solution to the specific problems the Comcast network faces. Bennett suggests that congestion at local nodes may be caused by excessive amounts of “Request to Send” packets, for which shaping is not feasible. Forging reset packets actually stops the Request to Send packets from originating in the first place.
Of course, if Comcast simply explained why they’re using Sandvine (assuming they have a good reason) criticism would likely be far more muted. When Comcast rolls out DOCSIS 3.0 in the next year or two, I believe that Sandvine will no longer be necessary not only because of greater node bandwidth but also because DOCSIS 3.0 is better suited for the realities of modern network management (namely, the ubiquity of P2P traffic). Sandvine is probably just an interim measure to keep costs down so Comcast can allocate more resources to overhauling its network.
1 year ago
Actually, my biggest criticism of comcast has been their dishonesty. Seth apparently believes that they are still being honest because because they are just "delaying" BT traffic. My point in posting the technical information from Ed Felten's website is to demonstrate their dishonesty, dishonesty which even Mr Richard Bennett appears to agree is wrong (see his comment over at freedom to Tinker)
Dishonesty here is key, because their dishonesty, when transmitted by wire, I believe starts to have some legal implications for Comcast, but IANAL, and I leave the legal eagles who will stick Comcast with a class action lawsuit to figure it all out, and I hope they do.
Personally, I'd like to see then lose a few million dollars for their attempts to stomp out BT traffic and then lie about it.
Recall that the testing that was done demonstrated that even BT files of just a couple of megabytes were blocked, too.
1 year ago
The technical details may have caused Comcast to unwittingly do something that many consider to be wrong.
However, notice that they continue doing it, and also that many who would switch internet providers because of Comcast's monopoly in their area can't do so. And many only have one other choice: AT&T; aka NSA.
So, given the monopoly (or duopoly) and the manifest harm being done, the stage has been set for: government protection of individual liberties, which, when the oppressor is a large corporation, comes in the in the form of regulation.
1 year ago
Are we agreed that Comcast's motives are almost certainly related to legitimate traffic shaping? All else is then armchair-quarterbacking.
They HAVE NOT attempted to stomp out BitTorrent traffic. They have attempted to keep BitTorrent users from overwhelming everyone else. There is no way government action can solve that, because it's a technical problem of bandwidth.
1 year ago
Are we agreed that Comcast's motives are almost certainly related to legitimate traffic shaping? All else is then armchair-quarterbacking.
First, the issue with me always was the deceitfulness of Comcast, and I don't believe that you can say that they were being anything but deceitful about their interruption of the BT traffic. If you tried to seed a 2 megabyte torrent your traffic was still interrupted. So Comcast was in fact censoring BT more than other traffic that used the same bandwidth. So their FAQ was dishonest.
Second, if there existed a robust, competitive market with three or four choices for most folks, they could decide, to paraphrase Trotsky, "vote with their feet"
However that competitive market does not exist, therefore the questions become:
1. Was harm done to consumers by the non-competitive market place?
2. If harm was done, what is the least drastic means to intervene to mitigate against that harm?
I believe the answer to the first question is: yes, harm was done, inasmuch consumers were unable to participate in BT uploading. (See Thomas Friedman-one of his "Flatteners' is 'uploading'*)
The question then becomes 'What is the least drastic means to prevent that harm?' and in that regard, it seems prudent to legislate 'transparency' rather than net neutrality, to fix this particular problem.
See this post:
http://enigmafoundry.wordpress.com/2007/10/25/c...
1 year ago
1 year ago
Their FAQ was not dishonest because they are clear they do network management, and to expect otherwise is absurd.
If there was government regulation which required Comcast to publish something like "We are using Sandvine appliances set at threshold ABC this week", what would change in practice? There would still be X file-sharers trying to use Y bandwidth. "Transparency" isn't a magic word. In fact, I suspect all the heavy used would then tune to just under ABC, and if Comcast then moved the threshold lower, a mighty scream would go up "LIARS! Lack of TRANSPARENCY! You *said* it was ABC you're lying, lying, lying ...". I can very well see Comcast making a decision that they're damned no matter what, so say as little as possible (yes, I know the pundits have been saying otherwise, but the actions do not match the rhetoric there - it looks far more like find something to flame them).
1 year ago
I understand that seeding, after you've downloaded a torrent, stops. I don't think it matters whether or not Comcast has 'deliberately stopped all bit torrent traffic' but it is much more important that this interruption is in fact unique to bit-torrent--i.e., it does not occur if I upload using ftp, or download using http or ftp.
So this throttling uniquely attacks bit torrent, while NOT altering ftp or http traffic, and it will stop a torrent of a small (2 meg) file while leaving a 5 gig ftp download unscathed.
The net effect of limits on seeding was not to prevent exchanges wholesale, but to throttle them. Let's get it straight - IT'S A KLUDGE! It's not a perfect way of doing bandwidth-limiting. But it would be utterly unrealistic to have a no-kludge government regulation.
Yes, legislating against stupidity is a dead end. But you are portraying this as ONLY a bandwidth measure when in fact it deliberately targeted bit torrent, regardless of the size or bandwidth consumed.
Second, the requirement for transparency would require some careful deliberation, but I think the requirement for transparency would, for this particular wrong, be more appropriate than legislating net neutrality.
Bit torrent is an important tool in the dismantling of highly centralized and oppressive corporate power structures. Therefore, any effort to stop or delay adoption and market penetration of bit torrent has a political dimension. That does NOT equate to some conspiracy theory that Comcast sat down and decided to "stomp out bit torrent" which I doubt happened.
But the effect of their decision was to weaken an important tool a tool which will ultimately play a part in depriving certain economic structures of their present power to do things like:
1. wage a campaign against the First Amendment;
2. wage a sustained and well-funded campaign to suppress the truth about global warming, (see exxonsecrets.org) This campaign may very well have already doomed tens or even hundreds of millions of people to premature deaths.
1 year ago
Sorry. This is utterly ludicrous. The BitTorrent developers are very eager to sell themselves in service of those corporate power structures, and have been making deals with them as much as possible.
BitTorrent is a network problem because in the real world, it's the cause of an overwhelming amount of bandwidith use. That's just an empirical fact. And I keep repeating that ground-level problem solving is not the same as geek-rant perfection. This shouldn't be so tedious, and it's a lesson that it is.
1 year ago
No it is not ludicrous just because you say it is. I will grant you that its potential is very far from being fully realized.
I would look who have been trying to make p2p illegal, and I would posit that as my evidence that many feel threatened by bit torrent. The RIAA and the MPAA have both tried to stamp out p2p services such as bit torrent.
So, Seth why, exactly, have they done so?
Secondly, many linux distros use bit torrent as an effective cost-shifting method to distribute their iso's. The cost gets shifted from the distros bandwidth charges to the unmetered service that downloaders/uploaders using bit torrent have already paid for.
1 year ago
The "cost gets shifted" is exactly right - but the problem is the network wasn't built for such cost-shifting. THIS CANNOT BE GEEK-RANTED AWAY! It must be addressed in the real world, with practical economic limits.
1 year ago
That does not change the fact, which you haven't explained, that the MPAA and the RIAA have tried many times to make p2p, like bit torrent illegal.
Any 'deals' they may make with developers will not change the basic architecture of bit torrent--a decentralized network for distributing digital content not under anyone's control.
Microsoft is an example of one of those highly centralized corporate power structures whose value web 2.0 will collapse, unless they evolve away from fixed product, and move to services and skills (like IBM is doing)
The big four labels are another example. How have cd sales been the last 4 years??
1 year ago
The same argument you make about decentralized network has also been make about the Internet itself, and we see how THAT worked out (and saying someday in the future the revolutionary potential will be reached is not a convincing argument).
1 year ago
Further, this article caught my attention, about the market for shrink wrapped software declining, and this is a result of peer production processes, processes whose results are distributed by bit torrent.
Further when you say: The "cost gets shifted" is exactly right - but the problem is the network wasn't built for such cost-shifting. THIS CANNOT BE GEEK-RANTED AWAY! It must be addressed in the real world, with practical economic limits.
I agree with you that that network was not designed for this cost shifting, but that's part of the reason why it is worth doing--It's a good thing!
1 year ago
Article mentioned above: