News

Vuze Study Points To P2P Interference From Cogeco

While Bell and Rogers have attracted much of the Canadian net neutrality attention in recent weeks, a study conducted Vuze, an online video site that uses the BitTorrent protocol, has placed another Canadian provider – Cogeco – in the spotlight.  To better track ISP network management techniques, Vuze created a plug-in that allowed users to measure network interruptions.  Interruptions – referred to reset messages – might occur in the ordinary course of network activity or might be the result of false messages used to hamper peer-to-peer file sharing.  Vuze managed to collect an enormous amount of data – 8,000 users worldwide generating over a million hours of data.

Perhaps unsurprisingly, the ISP with the highest percentage of resets is the same one that has been in the regulatory spotlight for doing so – Comcast.  Yet the number two is a surprise.  Canada's Cogeco, which offers service in Quebec and Ontario, has a reset rate far higher than any other Canadian provider (indeed higher than anyone other than Comcast).  None of the major Canadian providers fare particularly well (Rogers, Telus, Shaw, Videotron all rank among the top 68 of 108 providers), however, the Cogeco certainly captured Vuze's attention.  Although acknowledging the methodological limitations of its data, the company has written to Cogeco to ask it to "outline the specific network management practices your company uses, including, particularly, whether your network management practices include the use of false reset messages."  It will be interesting to see if Cogeco responds, particularly given the regulatory response to false reset messages in the U.S. and the growing demand for greater transparency of network management practices.  If it turns out that the company is actively interfering with network traffic, the CRTC may find itself dealing with another complaint before too long.

Update: A representative from Telus writes to provide assurances that they do not throttle P2P traffic. 

Update II: Ars Technica notes that research from the University of Calgary calls into question the relevance of reset data. 

19 Comments

  1. Antoine W. says:

    We need our politicians
    We need our politicians to stand up about this issue. “Market forces” will not resolve the problem, since the entire market is using this tactic.

  2. mr peepers says:

    not news
    This really isn’t news to anyone who has ever used P2P applications on Cogeco – they were one of the first ISP’s in North America to start traffic shaping on such a large scale and they have been doing it for about 3-4 years now. About 4-5 years ago, Cogeco used to offer the best service in all of Canada, perhaps in North America – 10mb/s uncapped/unthrottled amazing connections, then they added throttling and hard data caps (60GB on a 10mb/s connection is laughable) all the while raising prices.

    Obviously I left Cogeco a long time ago and have influenced countless others away from using Cogeco services which is really all we as consumers can do.

  3. CRTC Complaints
    I’m a Cogeco customer who’s stuck in an area where I have no alternatives. Seeding with a properly tweaked torrent client is impossible with Cogeco, and I’d like to lodge a CRTC complaint… how does your average citizen go about this?

  4. Bad service.
    It amazes me that a company will advertise one thing, and do the flip once they have everyone on board.

    Why are we paying for something that they cant give us.

    Yes, not too mention service that is getting worse.I had recently been stung by the service bug, where it took them over a month to port my phone number when I moved. There is no remediation what-so-ever. All I ended up with us 1/2 a month of free service.

    Once things settle financially, I will be looking elsewhere,such as teksavvy.

  5. grahamcharles says:

    crtc
    keep the focus on the crtc. market forces will always move towards maximize the bottom line, the oligopoly can’t be blamed for doing what the system requires of them. it’s simple economic theory that service/product quality drops as competition drops. to regulate this effect is the point of the crtc in the first place, it’s why they exist at all, they are grossly incompetent and in a sense are conspiring with the oligopoly against canadians.

  6. valid?
    Actually all this stirring up makes me laugh inside…

    From a statistical point of view, their method if flawed big time…

    They say they “managed to collect an enormous amount of data – 8,000 users worldwide…” Enormous ?? 8000 only ??? Out of the millions of users ??? Quite laughable, even more so that those who really follow these threads and get those plugins are actually those who have problems with their Torrents;

    I download and upload to my maxxed speeds on Cogeco and all those I’ve seen having problems were people that couldn’t setup their program adequatly;

    So even though their little study with a ground breaking 22 Cogeco clients involved points at many ISPs, I call BS on the statistical and mathematical value of their test since they admit in their pdf that they “cannot conclude definatly that any particular networkoperator is engaging in artificial or false RST packet behavior”

    Finally all that hype is created by uncertain data that is released by guessing people about an issue that clearly isn’t of public domain (it’s their network and they can do whatever they want on it, not happy about it ? Start your own company and build your own network…)

  7. valid?Written by me

    Why don’t you enlighten us then on proper set up

  8. Response
    “me”, some users are reporting they’re not being affected, but if you go to the Cogeco forum on BroadbandReports, you’ll see that MANY users have been complaining about it for years. Cogeco employees haven’t responded in months regarding these accusations, and it coincided (as others have pointed out on this forum) with the time they enacted a new 60GB a month cap that’s strictly enforced with temporary disconnections. It looks as though a small subset of people either aren’t being affected, or they’re just not downloading as much as those that feel that they are. I switched to Bell Sympatico (who sadly are also throttling now) about 6 months ago when Cogeco wouldn’t help or admit to this and my torrents immediately skyrocketed with the same exact software that was loaded with the same exact torrents.

    You might also note that, out of millions of users, a small minority are downloading torrents, let alone seeding them and the majority of people seeding them wouldn’t have the knowledge to notice this on thier own. Also note that this was a test amongst a small group of people and I’m sure more will be done in the future now that this is public knowledge.

    I’d also ‘setup [my] program adequatly [sic]’ and there was no question, especially comparing the differences, that Cogeco was doing something fishy.

    Why don’t you leave it up to Cogeco to publicly admit or deny these practices, now that the ball is in their court?

  9. R. Bassett Jr. says:

    Side note:
    Last year I recieved a letter stating that my cable bill will be increased due to all of the new services that Cogexco was providing. However, I live in an area where exactly NONE of the services they listed are available.

    To make a long story short, rural Cogeco customers are subsidising their big city counterparts and if they don’t like it, Cogeco give them one option: cancel your service.

    I’m not sure if this qualifies as a complaint, but it sure as hell isn’t a fair practice, given that Cogeco has absolutely no intentions of ever providing digital phone, highspeed internet, and video on demand in my area (or most rural areas).

    Were it not for my wife’s love of TV, I would have cancelled my account when I recieved the letter…

  10. Methodological Concerns
    Soooo, some methodological concerns:

    1. They don\’t really define how they arrive at the median rate. The median is the middle result, in other words, in between the worst and best results. Is it the median hourly rate? The median rolling hour window rate? Median per minute rate? It isn\’t defined.

    2. What about the averages? The Inter-quartile ranges? How about some histograms? Perhaps they could release the raw data.

    3. How about some confidence intervals? Some p values? You know, how on TV you hear that a survey is accurate to within 3.8% or whatever? That is because they did the math to determine what the probability of the result being due to chance was (since you can\’t say anything with 100% certainty).

    4. They really should have pooled the info that came from the same provider but with different ASNs, unless there is a reason for not doing so. Sounds like they were just lazy.

    5. Talk about small samples, 8000 is pretty reasonable, but 20-25? Flip a coin 20 times, it probably wouldn\’t take too many attempts to get it to land on \”heads\” or \”tails\” 15 or more times out of 20. This goes hand in hand with point 3, point 3 describes the mathematical way of being able to say \”bullshit!\”.

    6. Teksavvy shows up worse than Bell does… which makes it all that much more suspicious.

    Yes, what I am suggesting requires a statismagician, or someone that can think like one, but if they just released the raw data, a statismagician could go ahead and make better determinations.

    Regardless, the evidence here is still better than that of Cold-FX.

  11. response to cally
    I’m on BBR every day and I only see the same 10-12 people coming in to rant about their torrents…

    Cogeco haven’t been answering because most don’t know the specifics about it or are just tired of being yeled at by ranting people; they’ll gladly ignore any post speaking of torrents out of boredom…

    The people nowadays want everything now, and for FREE, considering Cogeco hasn’t charge a dime in the past years and just recently went on to join the other ISPs in including a normal cap to the amount of bandwidth allowed by user is only normal business practice.

    You’re also speaking of a “small” subset of clients aren’t being affected… I think you got it wrong, Cogeco has way over 450 000 internet clients… How many of them do you see on BR complaining about it ?

    Your exemple on your own torrent experience proves nothing, sorry to say that.

    You might also want to re-read the document Vuze presented, they didn’t differentiate anything network-wise, and even more unrelevent, the plugin, being placed on the user,s machine, is the wrong way to do the testing and only proves something might be worng with either the connection, the network adapter or whatnot… Not only with the alleged throyttling of the ISP… It counted every kind of packets, TCP, UDP etc; everyone knowing anything about the technical involvement in the method of throttling will tell you the test are totally inadequate and this just creates a big fuzz for people not understanding networking since all they wanna do is point a finger at someone for their own problems and yell/rant…

    And if you’d read more about the subject, Cogeco had already answered to this to the CBC I’ll even provifde you the link : [ link ] so please do us a favor and read up before posting again…

    on a side note : Thank you Mickael for bringing this topic up 🙂

  12. wow lots of typos
    Sorry for all the typos…

    And I totally agree with Dan for the methodological concerns…

  13. Telus User says:

    Throttling
    No, of course they’ll say they don’t throttle BitTorrent. Meanwhile, they are more than happy to ax Usenet in the process which they did this month.

  14. Vonage User says:

    What Is The Implication On VoIP?
    I’m currently a Vonage customer (VoIP) over Teksavvy.com

    There are numerous posts regarding throttling of TCP port-specific traffic. Additionally rumour has it that website throttling is inplace for various service sites (youtube, itunes, etc.)

    If such aggressive throttling is in place, does anyone know if this will impact non-BELL customers – i.e. those of us who have chosen to go VoIP?

    Advance thanks.

  15. UofC?
    I’m curious as to why the UofC thinks that reset counts are not relevant. What is then? Whenever I download a TV show using Bittorrent, I noticed my iBrowser slows down considerably in it’s response. I especially noticed it when I Dl’s Canada’s Next Great Prime Minister from CBC via a Torrent. I’m on Shaw, and I KNOW they throttle Torrent traffic.

    What does the UofC then propose is a valid test?

  16. Complaining to the CRTC
    It’s extremely easy to register a complaint with the CRTC. Go to one of the pages below (English or French) and follow the instructions. I have done so twice already, for different issues than what is being discussed here, and got results.

    [ link ]
    [ link ]

  17. VPN
    I’m on Shaw too and without any doubt I’m witnessing their throttling practice every day starting at 2:00 PM. I’m a data analyst and I use VPN to download and upload big chunks of data. To process 100MB I need 30 minutes during the morning and over 1 hour during the afternoon although my company is paying for the fastest possible connection. The nature of my job makes me one of those “heavy” users but actually I’m not doing anything wrong other than using those crippled resources I’m paying for.

  18. ars technica has another article
    P2P throttling leading to net neutrality showdown in Canada
    By Nate Anderson | Published: April 24, 2008.

    [ link ]

  19. Just Cause says:

    Capping
    Too Many major ISP suppliers are unacceptably guilty of initial and subsequent false misleading adverting practices, and an immoral Bait and switch practice as well.

    Here is the undeniable reality.. Many corporations beforehand do not disclose the amount of capping that they do to their customers., or after wards, or lie as to much they supposedly cap. For example I have a Bell Sympatico connection or I can use a second party proxy connection, and next I get twice the download speeds with the proxy over the Bell’s capped services even during the non peak hours as well, such as all day Saturday.. not just weekends 4.0 pm to 2. am when Bell admits it caps their lines. Now that is a fact any potential bell customer should know now too.

    “AP Sun Jun 15, 9:45 AM ET At one time, the word “unlimited” meant unlimited.

    Sprint’s mobile broadband service is the latest to abandon the term and the principle in favor of a monthly cap designed to keep their heaviest users from overwhelming their network.

    But Sprint isn’t alone: its two 3G competitors also cap usage, and two wireline broadband operators are testing explicit caps as well. In the earliest days of broadband, service was either heavily capped, with ridiculously low limits–I recall DSL plans that had 1 GB monthly downstream limits for business-grade offerings–or totally uncapped.

    Now, the idea of capped service with metered rates, stern warnings, or cancellations above a monthly limit are fully in fashion. For the last few years, companies like Comcast and Verizon’s wired broadband division have warned users about excessive downloads, degraded their service, or canceled their accounts, often with little recourse, and sometimes denying it all the while. Enough states’ attorneys general and FCC staff and commissioners have been involved that what was implicit has become explicit, but with the related effect that caps have become much lower than what they were in the ad hoc days before these changes. Driving all this is not scarcity, because there’s plenty of headroom out there on the Internet, but two interrelated issues: service providers always dramatically oversell their service, and some users are actually abusers. ( But really how can one be an absuer when he pays for and uses what was advertised now?

    On the first issue, if an ISP has 500 people connected to a central office DSLAM (a DSL aggregator) with a total downstream bandwidth of 2 Gbps, there’s no universe in which a phone company makes available 2 Gbps to that location. Rather, they allot a fraction of that, which works when traffic is bursty, not continuous. Many people downloading or streaming a lot impact everyone in the same grouping. (I’ve seen this at home when I complained about my 3 Mbps DSL dropping to 500 Kbps at night. A Qwest technician explained I was lumped with heavy users, and with about 20 minutes of waiting on the phone, regrouped my line to another, less used pod of users, and my service has been fine since. The nice part is that was a logical change; no one had to walk over to a cage and move my wires around.)

    The second issue has provoked a lot of debate. But without explicitly labeling the limits on a service, a subscriber can’t technically abuse it. If you know when you sign up for Comcast that they limit your use to 10 GB and provide tools to monitor as well as an understanding of what that bandwidth would allow you to “consume” each month, it’s a very different matter than “all you can eat. ”

    Verizon had long promised unlimited Broadband Access for their 3G EVDO mobile broadband service. But it was well documented that unlimited had fairly strict limits. After an investigation by the New York attorney general’s office, Verizon agreed to change its disclosures, pay some costs to the state, and refund money to some subscribers. The company now fully discloses its 5 GB per month limit for combined upstream and downstream data. Verizon charges you 49 cents per MB ($490 per GB) when you cross that limit, and the company says that they use email, SMS, and a live data usage display in their connection manager to keep you apprised. Note that a single high-definition movie download might consume nearly 5 GB.AT&T, likewise, has a 5 GB cap each month on LaptopConnect, its 3G cell data offering, with unspecified behavior when you top that amount–additional charges may apply, but clarity would be helpful. They note in their PDF-only terms and conditions: “The parties agree that AT&T has the right to impose additional charges if you use more than 5 B in a month. Prior to the imposition of any additional charges, AT&T shall provide you with notice and you shall have the right to terminate your service.”Sprint has joined this club with first the leaked news and then official confirmation that starting July 13, 2008, its 3G service would also have a 5 GB cap. A spokesperson told me that off-network roaming–ostensibly with Verizon or Alltel, the only other major providers of 3G in the US using the EVDO flavor–is capped at 300 MB per month. Now these are all 3G providers, who have limited spectrum over which they have to make sure all contending users in each cell get approximately the same kind of experience. They can’t afford one user sucking down all bandwidth. However, we’re seeing the same kinds of limits start to be tested for cable-based broadband.

    Comcast is testing delaying traffic–slowing down packet transmission to throttle the bandwidth rate–in two Eastern cities they cover for the heaviest users of their service. This is an effective cap, rather than a cutoff. (Comcast has been delaying BitTorrent P2P traffic for all its users prior to this; this change affects all traffic, not just BitTorrent, and is being announced, instead of sub rosa.) In a town in Texas, Time Warner Cable is experimenting with offering different speed packages each of which is coupled with a monthly limit on usage. The lowest-priced package offers a ridiculous 768 Kbps downstream and 1 GB per month for $30 per month; the highest-priced is 15 Mbps downstream with a more reasonable 40 GB per month limit. Charges are $1 per GB above that. With cable companies traditionally and telephone companies newly offering television programming, premium channels, and on-demand video, the caps are another tool to prevent competition from over-the-Internet sources of things to watch. In a situation in which a few carriers control all the pieces, it’s unclear whether rate caps can stick. If both telcos and cable companies decide to impose such limits and restructure their networks, who do you turn to? People with broadband are unlikely to cancel it. In a monopoly or duopoly market, you can’t switch brands. There has to be a happy middle–a role that the FCC may help to negotiate. A 40 GB cap switched to 400 GB might serve precisely the right purpose without penalizing average users who have no other market choice. With Time Warner Cable charging a buck a gigabyte above their monthly limits in their test market, but with Amazon’s S3 service delivering it retail for as little as a tenth that, it’s not hard to see that carriers are looking to caps to solve network problems and make a little scratch on the side.” [ link ]

    Beware always of men and women, bullies, tormentors, control freaks, persons, civil and public servants, politicians, pastors, leaders, elders, Corporations, governments who falsely do, will try to enslave you, oppress you, exploit you even while they claim they are proclaiming the truth, democracy, trying to help you, etc.,

    Is 51:23 ..your tormentors {and} oppressors, those who said to you, Bow down, that we may ride {or} tread over you; and you have made your back like the ground and like the street for them to pass over.