The FCC is to investigate Comcast's network management practices. Last month here I gave an expert view on how the EFF, and other campaigners who called for an inquiry, don't understand the problems. Now Peter Eckserley, a copyright academic at the EFF, has responded to my article. Let's recap the story first. To avoid …
I'm not convinced by the argument put forward here.
Yes, other peers will still attempt to open connections when there is packet loss. It dosn't matter. The bandwidth used by a few SYN SYN-ACK packets is insignificant.
"Packet drop applies back-pressure to conventional TCP sessions, but it does nothing to the data queued inside Comcast users' cable modems"
This is simply wrong. If there is enough packet loss the computers IP stack will send less data so there will be less in the cable modem's buffer.
Cable modem's don't have infinte upstream bandwidth. When someone on cable sends a big email the computer figures out how fast it can send by increasing the speed until it hits the limit imposed by the cable modem and packets start getting dropped.
Dropping packets to control bittorrent is just the same as setting a lower upstream limit.
The whole problem is fundametally because the cable companys in question don't want to be honest with their customers. The should just tell their customers that heavy users will have their upload speed limits set lower at peak times because of the limits of their infrastructure.
Once again you do your argument little good with your Bill O'Reilly-esque emotive rhetoric. After reading your first one and now this I am amazed you have the audacity to accuse the other side of some sort of religious fervor. You would say those of a rational bend much more readily if you stopped with your various ad-hominem statements.
Additionally the real issue seems to have flown over your head. While there may or may not be reasonable technical considerations for its' activity, it is the marketing and games they play that Comcast most finds itself at fault. There is no reason for Comcast to play word games to try and cover up the fact that they throttle/manage/shape traffic. But they did. And it isn't the customer's fault that they over sell their network.
If you can't handle a customer taking advantage of all you can eat, don't sell your service publicly as all you can eat.
If only the interenet were like a diner, as you suggest
I was looking forward to a non-emotional fact based analysis. I guess I was hoping for too much:
[quote] over-heated religious rhetoric[/quote]
With statements like that above it's obvious you are emotional about a topic that should be handled with reason and logic. You accuse the EFF of not reading articles on the one side, yet you don't seem to be reading the other side's points either. You fail to mention that random packet drop doesn't attack a specfic application or protocol, which is why it is acceptable to the EFF.
Now if you had suggested an alternative that deal reasonably with all bandwidth hogging applications fine, but your solution singles out Bittorrent, which is also not acceptable. What about other protocols that do this? What about when Apple gives you credit for sharing HD content, or when Skype shares upload bandwidth to help with multiuser video conferencing?
Notice I haven't talked about religious rhetoric here. I'm simply stating that you need to offer a solution that will solve the problem from a scale that is network wide, and not single out applications. So will you come up with a custom solution for every new app that is developed that attempts to use peer to peer technologies to work better? Bittorrent may be the only or the biggest problem today, but it will be something else in another year, and several other things the year after that. So now you've got custom spoofing hardware and software patched together to deal with every bandwidth sharing protocol.
This is like fixing the hole in the damn with a piece of gum. Pretty soon, your whole damn is gum.
While you accues the EFF of being too emotional about the topic, your response sounded equally emotional and irrational. Better to have the cable companies not oversell their bandwidth in the first place. They quote these great upload and download speeds, assuming only a single user on their network. They should have to assume usage of their maximum quoted speed in their advertisement, and not be allowed to sell bandwidth beyond that.
Please try to be less emotional about this topic in the future, or remove yourself from the debate.
You're missing the point
You're missing the basic point here--Comcast isn't merely THROTTLING BitTorrent traffic, it's KILLING uploads. It's one thing to reduce the speed limit; it's another thing entirely to close the road.
Why don't they just set up their own ISP to prove to the world how right they are about everything?
@ the author
"We can't provide the service you are paying for because torrent breaks the rules". Tough titty. Stop cashing in or re-invest in a structure that actually supports what you are seeling. Yes, I used "you", because obssessing about the subject and portraying the oposition as religious zealots (or nerdy day-dreaming teenagers like in the last article) babling drivel makes it obvious you cannot really counter their arguments to support your own, and paints the picture of someone trying too hard. Who's paying for these articles, anyway?
Traffic Shapping is not a techical issue, it is a consumer right's one. Disserting on why it happens and how it's done then hoping we will think it's ok won't cut it.
A.C.: because I still can.
You are missing a point.
Specifically that this was concealed from it's users, that they are still lying about it, and that it effects other applications, Lotus Notes for one.
If they have a problem, the sensible solution is to improve their (pretty crappy) network architecture in the long term, but in the short term, they need to have a competent strategy (phony packets ain't it), and they need to be straight with their customers.
Expert view- eh?
Got to love the expert-view. Intelligence is lovely, but it's most attractive in women because they tend to use it socially. Isaac Asimov (leaving the chair of Mensa) commented that he didn't like the tendency of men to become "intellectually combative".
Two points in the article I would draw attention to:
1/ You write that Comcast users have possession of rare file chunks, which makes bittorrent scoop up their bandwidth. Well, that is .. unsubstantiated. Expert conjecture?
2/ You write that Comcasts upload rate of 384kbit/s, makes it attractive to downloaders. Even though the download rate is a mediocre 4 Mb/s. Why would downloaders prefer a service with mediocre download speeds (but high upload speeds)? Perhaps you meant uploaders..
Certainly most users would support more bandwidth (both up and down). And don't think anybody cares about how the experts achieve it. Like the road network at the explosion of car transport, the network needs more expressways. Limiting passenger movements is at best crisis management. Move up or step aside.
And in case Isaac Asimov is reading this, I love your work.
I've said it before, and I'll say it again: Bandwidth throttling by type-of-service = good network management (VOIP and gaming need priority over browsing and downloads). Bandwidth throttling by source or destination = evil corporations trying to extort money (SBC shouldn't throttle Google, just because they are partners with Yahoo!).
I can’t believe these guys are taking it so far. I even signed the net neutrality petition back when it was about the AT&T extortion plan, but even then it was with a heavy heart, as I somehow knew it would evolve into the typical internet phenomena based on paranoia, pride, ignorance and noise. Why must zealots always turn every good cause into a religion?
256 things i hate about you
You don't understand how torrents work. How clients limit open connections, connection initiations, how trackers split torrent users into groups, how upstream and downstream is limited by clients, ratios of 1:1 and time spans, ,user usage patterns for torrents, the effect on downstream data rate of not limiting upstream , relationship of rare parts to torrents,
"upstream traffic moves considerably faster on its network than it does on a standard DSL connection."
" Comcast-resident system's possession of rare file parts, which is even worse"
"the user's system responds immediately, regardless of congestion, TCP window sizes, or load...BitTorrent downloaders will gravitate to peers on Comcast over those on DSL"
There's a great bit on that terrible pdf, where they simulate web browsing using 60% of available upstream bandwidth on the test network. They also class incomplete downloads as aborted and failed, not understanding hashing of the file and parts.
RE: I'm not convinced by the argument put forward here.
My thoughts exactly, well said.
They offered a service
Now, because they have oversold they are withdrawing the service but not reducing the price. Airlines are penalised for overbooking and have to take responsibility for their actions. The same should apply to ISPs. They should try being HONEST about their service levels and not resorting to underhand tricks.
Richard, I'm with you on this. Way to push back at those "information just wants to be free" kool-aid drinkers.
Controlling bandwidth is one thing Forging packets is another
Whilst Comcast might need to shape traffic to manage bandwidth, I can not accept that they should be allowed to send high level protocol packets claiming that they come from their customers. If they can do this, what's to stop them joining in other conversations your computer might be having.
Oh look, that's the end of an eMail - quick lets inject an advert into that SMTP data stream.
Before we finish transmitting that webpage, let's send a redirect to their competitors site.
Throttling bandwidth throttles techological developments.
Rather then spending money on routers that drop packets or reset connections, wouldn't it be a wiser use of money to construct a network that is capable of handing the demand of the modern internet, which is now demanding much greater upload bandwidth with the development of, Bittorrent, personal web radio streamed from a home server (slim devices), VOIP, video conferencing, etc. etc. etc.
This is not 1990 anymore, where the only thing you use upload bandwidth is for sending e-mails and requesting web pages.
By implementing bandwidth controls rather then improving network performance all comcast is doing is taking a crowbar to the kneecaps of technological development by assuring there will never be enough bandwidth to go around.
@"religious rhetoric" posters
Although I think that Richard Bennett is correct in his analysis, I do have to say he could have toned down the fervor. Usually when El Reg ads spin to the news, it is humorus, not vindictive or combative like this.
Let's repeat a few points
Some of the commentors seem not to have read the article (or the previous one) and are raising objections that have already been answered, so a few facts:
2. VoIP is not a bandwidth hog, it's an application with very modest bandwidth needs that is destroyed by unchecked BitTorrent.
3. Comcast's high upload speed relative to DSL is what makes seeders running on Comcast more attractive to other downloaders than DSL-based seeders. This brings connections into the Comcast network.
4. The bottleneck on the upstream side isn't bandwidth per se, it's the packet rate. So a number of connection requests use up the cable modem's contention slots before raw bandwidth is maxed out. It's not about bandwidth, it's about duty cycle.
5. TCP packet drop doesn't address the fairness problem at all.
I hope that helps.
More of the usual
comments then..." I have a right to rip off as much copyright stuff as I can and you have a duty to subsidise my infrastructure... "
Much as I personally dislike the idea I sometimes think the only way many of the ills of the net will be cured is to introduce charging by the packet... Then you rip off merchants are paying something towards the relative cost of what you use and what I use, and when some silly SOB gets grabbed by a botnet and used to turn out torrents of spam then they get billed for their carelessness...
It's easy to fact up a sellout argument.
You find what works and ignore the rest.
The author is obviously a shill as he avoids the obvious.
It is more profitable to avoid lawsuits from and/or please a copyright lobby than it is to maintain a proper network.
Network upgrades promised are just talk and I'm sure conspicuous by its absence in the article that DPI is part of it.
RE: More of the usual
Oooohhh, lets pull the "if you use torrent´s you are a pirate" card. Now I am scared.
Newsflash for ya: Linux distros and games updates are legit and distributed by torrent. Torrent makes ideal for the distribution of any big file. Why can´t we buy movies like we buy music from online stores? Torrent distribution could enable newcomers into the market as it wouldnt require massive investments in bandwidth to make the service viable and flog their DRM´d crap. Think of the indie movie industry, same as garage bands toss their mp3s in the wild to get noticed.
Try harder. Bittorrent (and it´s possible children, as there are a couple others in development based on it) was a leap forward for p2p capabilities and it wont go away because it will cost the telcos to adapt.
Paris icon, because she isn´t very bright either.
I think the critical point here is that Comcast only resets bittorrent uploads when there is congestion. Assuming they only do it when there is congestion, Comcast really isn't harming anyone, since they only reset when there isn't any spare bandwidth available to upload anyway. When there is congestion, it is better to interrupt protocols that handle delays gracefully, like bittorrent, than to interrupt real-time protocols like VoIP. If Comcast didn't discriminate by traffic type when handling congestion, bittorrent uploads would probably finish slightly faster; but VoIP might be completely unusable.
I think there are two analogies that might help explain the issues.
1. You go to a popular club, but when you try to enter, a doorman blocks you and tells you to wait in line. As patrons leave the club, the doorman allows new people into the club, and eventually you get inside. Did the doorman "block" you or "delay" you? The doorman certainly blocks people from entering, but only when the club is full; so in effect, the doorman is delaying you until there is no congestion.
2. You go to a supermarket, fill your cart, and go to the checkout. You notice a sea of cashiers with long lines in front of them, and resign yourself to a wait. While waiting, you notice that some of the lines are moving much faster than others. You are about to change to one of the fast-moving lanes when you notice the sign: "express lane, 8 items or fewer". As a checkout neutrality advocate, you complain bitterly to the management that if they dropped the restrictions on the express lanes, people with full carts could check out faster. You'll also complain about how counting the number of items in your cart is an illegal invasion of privacy.
In both of these cases, the congestion control mechanisms aren't theoretically necessary; the easy solution is to build a bigger club or hire more cashiers. Also, although the supermarket specifies the rules for the express lane, you don't know for sure why a doorman blocks you, the doorman might be discriminating against you for some other reason. It is better to patronize businesses that are not congested, if at all possible; but it is pretty much unavoidable.
In the physical world, we usually manage to deal with congestion without requiring a bunch of new laws. If a supermarket is always crowded and we can't use the express lane, we'll switch to a different supermarket. If we always have a long wait before getting into a club, we'll switch to a different club. Comcast's networks are regularly congested, so much so that it was worth adding equipment just to make the congestion more bearable. If you don't like Comcast's congestion or how they mitigate that congestion, switch to an ISP who makes better tradeoffs; I don't think it is necessarily a better solution to force all ISPs to use the same congestion control algorithms (or none).
But its not just bittorrent
Comcast throttle bandwidth on users regardless of their activity, and that's why I have little sympathy for them.
I wouldn't care if they only throttled back applications that are known to cause bandwidth problems and are in violation of their terms of service.
But this simply is not the case. They throttle back ALL connections, regardless of activity, if the user appears to want to make full use of their connection during the times they are awake and not at work - internet prime time.
This could be playing a game, downloading music from iTunes, or maybe streaming movies from pay-per-view or on demand websites. All of which are activities they use in their advertisements to sell you their internet package.
The problem is when they sell to more customers than they can sufficiently supply bandwidth to, they end up with the need to throttle all customers during peak hours - and that is not acceptable.
You don't buy 5, 7 or 12Mbit internet connections that are only available at those speeds during unsociable or working hours. You buy them to use during the evening and at weekends. If they can't support those speeds they have no business selling them to as many people as they have.
And that applies to all providers. Advertising internet packages with speeds they can't possibly maintain even for short periods of time is nothing but fraud.
The answer is to be honest with customers and inform them if they want internet speeds during sociable hours that are above 384Kbits, they need to pay real money. Otherwise the provider can't possibly invest in the bandwidth they need to support those speeds.
So you do what my provider does and offer two packages. The first is more affordable, but clearly states they will throttle bandwidth and place limits on the amount of content in gigabytes that you can download each month - or face an agreed increase in your bill.
The other, significantly more expensive option, is to have true unlimited connections at speeds of 3, 5 or 7Mbits. They can't realistically go above this speed because they can't realistically buy sufficient bandwidth to do so. But they will upgrade hardware and infrastructure to support anything up to that limit.
I bought an unlimited 5Mbit connection and no matter how much I download it is never throttled back - unless that activity is in violation of my terms of service or one of a very limited number of activities such as using torrents.
And that's all Comcast and other providers need to do, be honest and offer separate packages at realistic pricing.
Charging by the packet might seem a good idea, but so does free ice-creams for children. The market will be driven by where the money goes, so those ISPs that offer uncapped bandwidth will be able to charge more for the service. Maybe the internet should become two-tiered, with packet-charging for old grannies, and flat-out uncapped service who want to pay a premium for it. Look at me, data prices are coming down. Truth or dare. Phone calls don't get more expensive for those who make high volumes of calls. It gets cheaper. User won't take packet-charging, they'd rather walk.
The packet-charging issue has been around since the early days, and it won't suddenly happen just because of an El Reg mention. Sorry El Reg!
Why doesn't Comcast have their own Torrent box?
A simple solution to me would be to have ComCast have their own BitTorrent box at the head end of the cable. That way it doesn't need lots of uplink speed, and all the users can have a go with the downlink speed they like. I suspect that this is a much nicer solution than shooting out the tires of people who exceed the speed limit by 1MPH (forging reset packets).
It looks like the Tortoise (the Slowskys aka DSL) is going to beat the Hare (ComCast).
p.s. I have DSL, no problems!
Sorry, dude. Asimov died in 1992.
"Therefore, it's acceptable for Comcast, as a matter of reasonable network management, to employ TCP Resets to prevent BitTorrent doing harm to the web browsing, standard file downloading, and VoIP sessions that are the typical behavior of the Comcast customer."
So your argument boils down to Comcast's behavior is acceptable because for them to do otherwise would be difficult and/or expensive.
Perhaps, instead, they should stop over-subscribing their network.
Can you elaborate on this point: <blockquote> "The bottleneck on the upstream side isn't bandwidth
per se, it's the packet rate. So a number of connection requests use up the cable modem's contention
slots before raw bandwidth is maxed out. It's not about bandwidth, it's about duty cycle."
Is your argument, in sum, the following?
<blockquote> In contrast to other forms of
traffic, Bittorrent produces a large number of small synchronization (SYN/ACK) packets which
substantially increases contention at the DOCSIS MAC level (through collisions on the "contention
slot"? Or contention for mini-slots?). Packet drop has no appreciable effect on the number of these
packets and so such "throttling" is ineffective. </blockquote>
I would expect the rate of
contention to be a function of the amount of data to be transmitted and not the number of packets:
if you are constantly sending data you need to vie for the same transmit slots regardless of the
size or type of the individual packets. That is, the same data rate HTTP transfer should create the
same degree of contention as a Bittorrent transfer.
Second, the amount of contention is limited in
some way (it is not unbounded). How?
In the paper you cited, "Assessing the Impact of BitTorrent on
DOCSIS Networks" I see no comparison to performance degradation caused by other forms of traffic (e.g. HTTP). That there is contention when links are highly utilized is not under question. There is
no evidence in that paper that Bittorrent, as a protocol, causes more contention that other forms of
Please do not shy away from precise, technical explanations.
There is more to this article than what is presented...
true BitTorrent is based on *symmetric* capability...
however "Sandvine" is man-in-themiddle and ignorant of any high-speed connection,
(I have had mid-connection-resets with and without BitTorrent running)
I run BitTorrent and find the easiest means of controlling its bandwidth use is to rate restrict
the connections it handles symmetricly, only allow incoming at the outgoing rate.
this lets "packet drop" rate manage the connection sufficiently,
its unfortunate but a "sandvine" afflicted connection with man-in-the-middle resets breaks
beyond just the bittorrent protocol being affected, *any* 'secured' protocol isfair game,
as it is now standard practise for all bittorrent protocol to have crypto applied,
There is evangelical responses for and against this, what are the facts?
without following the evidence trail with factual data and repeatable results there is no
point to the argument,
The following is entirely my own opinion,
any ISP need have a set guideline and policy with regards network "Fair use",
and have this as a document available for all users,
I consider any "TCP reset" equivalent to the telecoms company pushing "hangup"
during calls made (would any telecom company actually do this to customers?)
I *have* read the RFCs and Specifications of the protocols concerned,
I am of the impression that there is police-state erosion of the US constitution
by the death of a thousand cuts...
can anyone inform me otherwise?
I just don't get it...
I think I've commented before on this issue. But first, I would love, absolutely love, to see a customer come forward and say, "I've been financially harmed by Comcast and their network shaping policy." I am a Comcast subscriber, and while I don't like the company or policies, they do provide some of the best service in my area. DSL providers seem to have problems with availability and upload and download speeds. You are buying a product that specifically says there are limitations. Just like any other ISP out there, the Ts&Cs say that the 6Mbps speed is an "up to" speed, and other factors will influence the top speed you see. Next, we are talking about a program that, for the most part, does not play nice with established standards. Taking Comcast to task for not playing to the standards you think should exist when dealing with a program that doesn't play nice is a bit of a hypocritical response. BitTorrent is a great tool, but I would say it was before it's time. It does a good job of spreading the load and decreasing a single entity's bandwidth usage, but it does that by off-loading it to residential pipes that are not designed for business-class usage. So a company like Blizzard, which sends out patches via BitTorrent clients they have built into the game engine, are reducing the 1800GB of data they have to send out with a 150MB patch to 12,000,000 clients and spreading it out to all 12,000,000 million clients, which helps reduce Blizzard's bandwidth usage and doesn't increase latency for gamers when connecting to Blizzard.... but it gives it all to the Comcasts, Qwests, BTs, and others of the world through pipes not meant for such use.
And then the flip side, whining about how, well, if Comcast thinks it takes up too much space, they should spend more on capital expenditures ignores the realities of the world. Just because you want more network capacity doesn't mean you can buy a couple more network routers and flip a switch. The company you work for pays dearly to increase the LAN and Intranet bandwidth, and they just have a building or two to wire up. While Comcast has to buy all the equipment then pay workers millions of dollars to dig up the ground, put the fiber in, cover it back up, troubleshoot, get new systems to handle the new backbone, etc. And without any promise that the new bandwidth will bring about more customers. Welcome to capitalism. Let me get you the handbook and you can review how building excess capacity without knowing you will have new subscribers to fill it is what we call a bad idea.
Business sense, people... not utopian, rose-colored glasses vision. Or, you can petition your local or national government to treat the pipes as a government property... and then we'll see how responsive things become.
I welcome unbiased expert opinions, but being paid to hold an opinion does not qualify, Mr. Bennett.
There is a long-standing principle documented in RFC-793 that TCP traffic is an end-to-end connection; having a third party forge reset packets (aka "Reset Spoofing") in order to disrupt network traffic is widely and correctly regarded as a malicious form of denial-of-service attack.
With regard to the claim that "BitTorrent strives for a symmetric interchange of data, offering as much (or slightly more) in the upload direction as in the download direction."-- the BitTorrent clients I am familiar with, such as Azureus and uTorrent, default to limiting the upstream bandwidth to about 10% of the user's total upstream bandwidth, in order to avoid significant congestion of other outbound network traffic. Of course, that can be adjusted to suit the user's preferences.
With regard to the notion that the only choices for managing network bandwidth are random packet drop or reset spoofing, Mr. Bennet seems to be unaware of techniques used in many firewalls and freely available in Linux and BSD operating systems as part of IPFW or PF+ALTQ which include Weighted Fair Queuing (WFQ), WF2Q+ (http://redriver.cmcl.cs.cmu.edu/~hzhang-ftp/TON-97-Oct.pdf), and other variants of hierarchical packet scheduling found in ALTQ.
Lots of people use these today to prioritize VOIP, ICMP and DNS traffic over FTP or peer-to-peer traffic, and these QoS mechanisms scale up to at least T3/OC3 bandwidth on consumer-grade (CPE) Cisco routers or P2-grade Intel boxes, and these mechanisms do not involve forging traffic or even slowing down lower-priority traffic if the network is not being utilized for higher priority traffic.
Re: Good analysis
@Brent Gardner: "I've said it before, and I'll say it again: Bandwidth throttling by type-of-service = good network management (VOIP and gaming need priority over browsing and downloads)."
Why? No, really, serious question: "Why?". Why should a specific application protocol have priority over another? Could it be that your choice is dictated by the fact you game on-line and you use Skype (or alternative)? If so, your statement reeks of hypocrisy.
As for myself, I limit my Torrent application to 50/12 off-peak, 25/6 peak time. Torrent applications can be set to be "friendly" to the network. Can your game/VOIP application do the same? Or do you require "full speed" to best enjoy the experience?
And what happens when VoD (video on demand) rears its ugly head? Will people allow the ISPs to sabotage their pipe because someone wants to make a phone-call without paying long-distance changes?
This approach by ComCast smacks of a head-in-the-sand, short-sighted solution.
Won't WiMAX fix all this?
I mean, if Comcast gave all its subscribers a WiMAX box, they could all mesh into a new internet layer providing the extra bandwidth required without the expensive task of digging up roads?
Comcast needs to be realistic; they think they've got problems now, with Bittorrenters hoovering up all the spare bandwidth and making the network run at capacity. What about when your average consumer joins the feast at the trough when IPTV goes mainstream? Face it, more and more information is going to flood the internet and they need to expand the infrastructure rather than dilly-dally around trying to restrict the flow of information.
With the WiMAX example above, they also pass on much of the cost of expansion to the consumer.
I want my bittorrent. Make it happen!
even if that includes paying more ...
I would have thought...
"the routers to do packet-drop in real-time are more expensive than the aynchronous Sandvine system."
...the simplest way to achieve this is to just buy inexpensive, underpowered routers and let the internet grind to a halt in the sea of lost packets until tcp/ip stacks throttle themselves so far back that your network runs at a constant speed again.
Now, it's not very <i>controllable</i>, but what do you need now, eh?
What a shame...
..reading such Luddite stance here on Register.
An ISP is just responsible for providing the capacity the customers have purchased and it is up to the customers to use it as they please.
Bittorrent is only one recent and advanced communication protocol. If the ISP has problems delivering capacity for it imagine how they will cope with all the future even more aggressive and more efficient transfer protocols.
ISPs, stop overselling and start investing in your infrastructure so that we can always enjoy future technical advancements on the Internet.
So the bottom line is
Not only are BitTorrent users nicking content they should be paying for (ok of course there IS legit stuff on there, but how I bet it's a fraction of the illegal downloads), but they're also screwing up the general network for the rest of us?
Wow, they really are nice netizens aren't they??
Almost every service, from water & electricity to lans and broadband internet connections are never specced for everybody using all the capicity at the same time.
Why, partially becasue yuo will have huge amounts of capacitity idle for the majority of the time adn therefore uneconomic.
If you want 2meg internet all the time with no traffic shaping then you are going to have to spend more than your 12 quid a month (or whatever your subscription is). Thats why companys spend stupid amounts for their network connectivity.
@Aubry Thonon, VOIP (& gaming) require a higher priority otherwise you end up with unuseable phone conversations due to dropped or delayed packets (and in the case of games characters mysteriously changing position in strange fashions).
Higher priority means that these packet are less likely to be dropped, and sent in preference to others. Without QOS (Quality Of Service) and (prioritised data) the services would be almost completely unuseable on a network gongested with streams & tP2P traffic.
In the case of downloading a torrent, it really dosn't matter wether your (possibly copyrighted) torrent takes 6h 4m and 17s or 6h 7m 08s, because game & VOIP data is prioritised.
Unfortunitely so many bittorrent (emule, limewire or whatever) users do not act in a considerate manner, they want their mp3s and they want them now.
Just a side questions just how many linux distros does a person need to download via bittorrent every month ?
Caveat Emptor (or, always read the small print)
As I have commented repeatedly, Comcast (and other ISPs) Ts&Cs are very, very clear on these issues.
If you didn't read them before purchasing the product, whose fault is that ? Yours, you stupid freetards.
EFF: Comcast are *spoofing* packets. Suppose if the postal service, during busy times, rather than delivering your letters with some delay sent "I don't want to talk with you" letters in your name instead.
ISPs: If they sell "up to 8Mbps" services, they can hardly fault customers that use "up to 8Mbps". If the electricity company sold unmetered electricity at "up to 10KW" that's exactly what people would use.
Docsis modems: If traditional TCP traffic management doesn't work, how come the modem buffer doesn't overflow as the PC tries to send 100Mbps to it (through the ethernet link)? Answer: it does, and the dropped/delayed packets cause TCP to throttle back the connection.
@Aubry Thonon: I don't game, and rarely use VOIP. The issue is that these are "real-time" protocols. A phone conversation where you get a delayed packet is pretty useless. See AC's response.
@b166er: In many ways what you suggest is a good idea, but I doubt that technology will ever allow us to communicate over the air as fast as over the wire/fibre, whatever. The issue is that when you have a wire, (especially coax), you have the entire radio spectrum to yourself. When you are on WiMax, you have a limited section, and interferance and atmospheric conditions to worry about. Also, the mesh topology doesn't work for the long haul. It is fine to find the best path to have an IM conversation with your neighbor on the other side of town, but the truth is that most internet traffic is travelling thousands of miles. That would never fly with radio hops because of latency. So it all has to go out over the "backbone" anyway, which results in the same congestion.
Some kind of control is needed...
... to ensure everyone gets a fair share. Traditional mechanisms tend to share on a per TCP connection basis. Most P2P apps have exploited this by simply having large numbers of simultaneous connections. (I don't think this is news to anybody). It's pretty obvious that providers are going to try to limit the impact of such applications on their other customers. Their sales and marketing people have ensured that details of this is buried in legalese in their terms and conditions but it is there.
@ the @Aubry Thonon comments
Although VoIP and gaming may require QoS to work properly why should VoIP and gaming users have priority over anyone else? Why should my internet experience be compromised just so you can get free calls or play games?
OK, if you pay more for it then fill your boots, but if not then tough, its a shame but use a normal phone and get some real life friends.
The bottom line in all of this is that you need to be careful what you wish for. At the moment there are very few controls / throttles on internet traffic and its a given that you get about half the bandwidth you are "promised". If you buy an "up to 8Mb" connection expecting more than 3.5 to 4Mb then you need a dose of the reality stick.
And although linux distros, game updates and some other legal stuff do get distributed over P2P would anyone actually claim that legal P2P traffic makes up more than a few percent of the total P2P traffic? With a straight face and expecting more than a piece of coal for Christmas that is.
The upshot in all of this is that the ISPs will do or be forced to do the following:
-Guarantee bandwidth. However on what is "up to 8Mb" now they will instead offer a guaranteed 3Mb. And then probably throttle it down to 3Mb if you could get more. At best they will just guarantee 3Mb at leave it as is.
-limit uploads/downloads across the board. And no-one will get a better deal than they do now, you can guarantee that.
-Insert clear terms that prohibit P2P software or anything that looks like P2P, and either charge, cripple or terminate anything that looks like P2P. Or maybe, out of spite give details of anything P2Pish to the RIAA etc.
There will also be bolt on packages for prioritising VoIP, gaming, BBC iPlayer etc. traffic. At an additional cost with no guaranteed results, obviously.
Now, I know not all of these are bad, and none of them will detrimentally affect lots, if not most of the normal interweb users, however I think those that are now shouting the loudest will get hit the hardest.
Truth in advertising...
Until 2004, they absolutely did advertise the service as "unlimited". They don't any more, but they also don't tell you that there is, in fact, a limit, because they don't have one. They have a nebulous soft limit, which is whatever the hell they say it is.
While I understand your argument about them controlling their network, there are acceptable and unacceptable ways to do that, and spoofing packets is *never* acceptable. Not that it helps them, of course, because every Comcast user is simply using encryption on their BitTorrent sessions now, which they don't reset because they cannot recognize them anymore.
If they offer me a speed instead of a limit, then I will adhere to that speed and not to a limit. If they sell me 4 megs down and 384 up, then by god I should get 4 down and 384 up, period. 24/7. All the time. If they cannot provide that, then they damn well should not have sold it to me. If they actually provided what they *claimed* to provide, then they wouldn't be in this mess. If they can't provide it, then they cannot claim it. Truth in advertising is the law, and I really truly hope that the FTC bitch slaps Comcast for their false advertising.
ISPs seriously needs to focus on increasing and improving on their network speeds, reliability and customer relations. Investing time and money in throttling services just isn't the answer. I can see a decade ago when web surfing meaunt little more than sendimng an email or loading web pages. Technology has evolved. We are a multi-media world now. Wasting precious resources on limiting what your customers can do is pathetic. Many other countries, Sweden, as an example, offer virually unlimited bandwidth to their people. Japan might be another example. These countries and many other have the foresight and vision to look ahead, not to point fingers at one application, like bittorrent and spend precious time and money to limit people's access.
Throttle Away Comcast
I am a Comcast customer in the U.S. I noticed some time ago, while using Transmission, that I was suddenly unable to upload to peers the Torrents I'd downloaded. I switched torrent clients and that solved the problem (at least for the moment). I read this piece with much interest. The argument is absolutely sound. P2P traffic has to slow down any network used. It's why the Onion Network developers ask the P2P traffic be kept off of that resource; it simply can't handle the load.
Given the validity of the fact that a certain kind of traffic slows down the network, the problem is this: Comcast sells a product by touting its speed levels. If one visits the Comcast page for US customers, one reads, "Stop crawling the web and start burning rubber with scorching speeds up to 4 times faster than 1.5 Mbps DSL, up to 7 times faster than 768 Kbps DSL, and up to 100 times faster than 56 Kbps dial-up!" Under the Terms, one reads: "Actual speeds may vary and are not guaranteed. Many factors affect speed." Ok. Caveat emptor. They'll do their best to get my service up to those minimums. In testing my connection, however, I have never reached an upload speed anywhere near Comcast's advertised maximum. My upload connection always hovers around 360 Kpbs.
Finally, nowhere is the Terms of Service on Comcast's website does the company state that it will shape its bandwidth, and/or interfere with the software users choose. U.S. law says Comcast may not shape its bandwidth, and must remain neutral is handling various kinds of traffic. The fact that the company was "caught" shaping traffic and then proceeded to lie about it resulted in the FTC's intervention. Comcast officials lied because what they did in response to Bit Torrent's unanticipated and extensive use of its network was to break the law and kill the traffic. It was plainly stupid, at best. I throttle my Torrent speeds voluntarily; I know others do, as well. It's a network, not a trough.
I don't think there's a simple answer to this dilemma. Using the US Mail to commit a crime is a felony. However, clerks may not open your letters at the Post Office to determine whether you are breaking the law. P2P software enables users to break copyright laws in various countries; Comcast is abetting this, and has come under tremendous pressure from those agencies that guard various copyrights. In my opinion, Comcast is between a rock and a hard place.
I am waiting for RIAA and MPAA to start offering companies like Comcast money to block P2P traffic. Paying network providers would be infinitely less expensive and less repugnant than suing college students and other individuals.