Thanks to data collected by Measurement Lab (M-Lab) the public is now able to take a look at the frequency of these BitTorrent throttling practices. Among other tools, M-Lab runs the Glasnost application developed by the Max Planck Institute.

At TorrentFreak we have been keeping an eye on the changes in BitTorrent meddling among ISPs around the world, and today’s data is the latest installment. The throttling percentages for various countries and ISPs are based on M-Lab tests that were performed between December 2012 and December 2013.

Looking at the list of countries where at least 100 tests were performed, South Korea (74%), Malaysia (61%) and Singapore (53%) come out on top. Poland is the first European country with 35%, quickly followed by the UK with 28%.

Greece and Romania stand out in a positive sense, as only 7% and 9% of the tested BitTorrent connections were limited. The United States, Canada and Australia perform relatively well too, with throttling rates of 14%, 17% and 18% respectively.

United States

BitTorrent throttling in the US is not as prevalent as it used to be. After Comcast was exposed for interfering with customers’ file-sharing traffic, the throttling percentage took a dive from nearly 50% to only 3% in the period between 2010 and 2012.

Over the past year, however, Comcast’s throttling levels started to rise again to 12%. Cox is currently the worst of the larger ISPs when it comes to BitTorrent throttling with 13%, more than double what it was in early 2012.

Verizon runs the least interference on the BitTorrent traffic of its customers, 6% for Verizon business and 9% for regular accounts.

Worst: Cox (13%)

Best: Verizon (6%)

UK

In the UK, BitTorrent throttling rates are relatively high. Orange is currently the worst offender as it limits 38% of the tested connections. BT comes in second place with 35%, which is down from 65% during the first months of 2012.

At TalkTalk and BSkyB BitTorrent users are better off with throttling percentages of 10% and 17% respectively. For BSkyB this is a hefty increase, as the throttling rate is up from only 4% two years ago.

Worst: BT and Orange (35 and 38%)

Best: TalkTalk (10%)

Canada

Traditionally, nearly all of the major Internet providers in Canada have been heavy throttlers. Rogers in particular has a bad reputation, as it limited more than three-quarters of all BitTorrent traffic for years on end. This year, however, there are signs of improvement.

During 2013 Rogers interfered with only 20% of all BitTorrent transfers, down from 80% two years ago. Bell has shown the most progress though, moving 77% to 8% this year. Eastlink is currently the worst performing ISP in Canada, limiting 31% of all tested connections.

BitTorrent users in Canada are best off at Cogeco and Telus with 3% and 0% respectively.

Worst: Eastlink (31%)

Best: Bell (8%)

Australia

BitTorrent is extremely popular in Australia, but over the past year BitTorrent throttling has been relatively moderate.

iiNet is the worst offender interfering with 21% of BitTorrent downloads, up from 11%. Dodo performs the best of the larger ISPs, with a throttling rate of 12%.

Worst: iiNet (21%)

Best: Dodo (12%)

—

All in all we can say that BitTorrent traffic is still heavily limited, mostly in Asia. In all countries there are quite large differences between various ISPs. For BitTorrent users it is wise to ask ISPs about their traffic management policies and take a look at the M-Lab data before signing up.

If we haven’t mentioned your Internet Provider above, an overview of all the data is available here. Those who want to know whether BitTorrent traffic is being limited by their ISP can test their connection with the Glasnost tool.

Source: TorrentFreak, for the latest info on copyright, file-sharing and anonymous VPN services.

Over the past few years the situation has improved somewhat and in many regions of the world users now have more available bandwidth than they can consume. But for a minority there is no such thing as enough and maxing out a connection is a daily activity.

BitTorrent users are famously bandwidth hungry and as such they have been regularly targeted by ISPs looking to more effectively manage their capacity. While Comcast’s infamous torrent meddling is now six years ago, some ISPs are still looking to introduce controls to relegate the world’s favorite file-sharing protocol.

This week Aussie ISP Telstra confirmed it would be pressing ahead with ‘trials’ aimed at ensuring that its customers “enjoy the best quality of service.” The ISP says that traffic on its ADSL network has doubled on average every 12 months for the past four years so to cope it’s going to have to better manage its network. It appears they will do this by meddling with file-sharing traffic.

“This latest trial is focused on how our customers respond when network management techniques are applied in order to manage congestion,” explains Telstra director John Chambers. “Some trial participants will be asked to evaluate how speed differences on non-time sensitive applications, like Bit Torrent, impact their overall customer experience.”

While it’s easy to see how BitTorrent could be classified as less time sensitive than HTTP for example, one could imagine the response if Telstra said they would classify Netflix in the same way as torrents. They can’t do that of course, since Netflix isn’t even available Down Under, but if bandwidth is short now, things are probably going to get worse when it is.

So, preparing for this eventuality, Telstra are looking for some volunteers to have their torrent traffic throttled to see how it affects their Internet experience. The big question is this: why on earth would BitTorrent users volunteer to have their traffic meddled with?

It’s a bit like asking if drivers would put themselves forward to have someone else take control of the accelerator in their car. Light users might not object so much but heavier users – the ones who are presumably causing the problems – aren’t going to put themselves forward for this. So what’s the point? Rival ISP iiNet believes that the trials have an ulterior motive.

“I often describe it as boiling the frog – you just do it by gently turning up the heat,” said John Lindsay, CTO at iiNet.

“They are starting to take their customers on a journey. They know what the end point looks like, but they don’t want to describe it too clearly just yet for fear of scaring off the profitable customers. Calling it a trial is a nicer way of introducing the concept and the technology into the business.”

At the moment Telstra are limiting the availability of this trial to just a few hundred keen volunteers, so if you fancy having your torrent traffic temporarily meddled with – perhaps with a view to those changes being implemented long-term – hurry over to Telstra now and get your name down.

“I would reconfirm that this trial does not involve looking at or recording the content of what customers chose to consume on the internet,” Telstra’s John Chambers concludes.

Reassuring.

Source: TorrentFreak, for the latest info on copyright, file-sharing and anonymous VPN services.

Previously the researchers published data up until 2010, and now the results have been updated to include the first quarter of 2012. This allows us to give an overview of trends and changes that have emerged in recent years.

United States

BitTorrent throttling in the US is not as prevalent as it used to be. The main reason for this is the Comcast BitTorrent blocking controversy which started in 2007. The FCC eventually ruled that Comcast had to stop its targeted interference of customers’ BitTorrent traffic.

As a result of this ruling, the throttling percentage took a dive from nearly 50 percent to only 3 percent in 2010. In the first quarter of 2012, Comcast’s throttling level was still at 3 percent, which puts the provider among the best behaving ISPs.

Throttling was also greatly reduced at Charter over the last year, from 11 percent down to 4 percent. Cox is exposed as the most heavy throttler among the major ISPs, but with 6 percent this is still rather acceptable.

Worst: Cox (6%)

Best: Comcast, Verizon, AT&T and others (3%)

UK

In the UK BitTorrent throttling in on the rise, at least among some providers. BT is the worst offender by limiting 65 percent of all BitTorrent transfers during the first months of 2012. This is up from 57 percent last year and almost twice as much as the 35 percent in 2010.

At O2 and BSkyB BitTorrent users are better off with throttling percentages of 2 and 4 percent respectively. For O2 this is a significant decline compared to their 13 percent last year.

Virgin Media sits somewhere in the middle after it interfered with 22 percent of all BitTorrent transfers in the first quarter of 2012. This is down from 33 percent during the same period last year.

Worst: BT (65%)

Best: O2 (2%)

Canada

Canada is not the most friendly country for BitTorrent users. Nearly all of the major Internet providers are heavy throttlers and Rogers tops them all. For more than half a decade Rogers has continuously throttled more than three-quarters of all BitTorrent traffic.

During the first quarter of 2012 the provider interfered with 80 percent of all BitTorrent transfers, showing that there has been no improvement. Bell is a good second with 77 percent, up from 56 percent last year.

BitTorrent users in Canada are best off at Cogeco and Telus with 3 and 0 percent respectively.

Worst: Rogers (80%)

Best: Telus (0%)

Australia

BitTorrent is extremely popular in Australia, but according to the Measurement Lab data ISPs are throttling less than last year.

iiNet is the worst offender interfering with 11 percent of BitTorrent downloads during the first months of 2012, down from 32 percent last year. Dodo hasn’t been caught throttling at all in 2012, a healthy improvement compared to the 14 percent score last year.

The rest of the Aussie ISPs are also throttling considerably less than in 2011.

Worst: iiNet (11%)

Best: Dodo (0%)

The Rest

A quick look at some other countries shows that in Sweden none of the large ISPs have been throttling BitTorrent traffic heavily so far in 2012. The same is true for other European countries such as France, Italy, Spain, Poland and the Netherlands.

In Germany, Kabel Deutschland limited more than one-third (37%) of all BitTorrent transfers, a major improvement on the 69 percent from a year ago.

For Brazilian BitTorrent users Net Servicos appears to be the worst option with a throttling percentage of 13, and in Japan Infoweb Fujitsu and J:COM should be avoided with 60 and 83 percent respectively.

If we haven’t mentioned your Internet Provider here, a full overview of all the data is available here. If you want to know whether your BitTorrent traffic is being limited you can test your connection with the Glasnost tool.

For those BitTorrent users who have a choice, the overview may help to make the right decision when signing up with an Internet provider.

Source: TorrentFreak, for the latest info on copyright, file-sharing and anonymous VPN services.

The report was initiated in 2011 following requests from the European Commission addressed to BEREC, the Body of European Regulators of Electronic Communications. The aim was to gather information relating to the net neutrality debate, specifically regarding actions taken by ISPs to hinder, throttle, block or degrade Internet traffic.

A questionnaire was sent to out to ISPs across Europe and in the end responses from a total of 381 ISPs were considered, 266 ‘fixed line’ ISPs (140 million subscribers) and 115 mobile operators (200 million subscribers).

Among other things they were asked about their traffic management practices in respect of a range of objectives including preserving quality of service, contractual agreements with customers, network security, and other measures required by legal order.

Perhaps unsurprisingly the most frequently reported restrictions outside simple contractual data-capping concerned the blocking and/or throttling of peer-to-peer and VoIP traffic. A total of 49 fixed-line ISPs reported interfering with P2P, as did 41 mobile operators.

While the restrictions affect at least 20% of all subscribers, the report notes that the impact on end users can vary. Some ISPs restrict all of their users, some just a percentage. Others restrict traffic during certain times of the day, while some have contractual restrictions that are not necessarily enforced technically.

Overall, in the fixed-line broadband market at least 21% of users are affected by P2P restrictions, either technically or by clauses in their contracts. In the mobile market that figure is 36%. When it comes to VoIP related restrictions, at least 21% of the mobile market faces restrictions with the possibility of an additional 18% depending on conditions.

The findings of the study will be used in the ongoing Open Internet debate in Europe. Earlier this year the Netherlands was the first country to sign Net Neutrality principles into law, which would restrict the use of P2P-throttling by ISPs. In the European Parliament the Greens are proposing to guarantee Net Neutrality in the rest of Europe too.

The full report titled BEREC findings on traffic management practices in Europe can be downloaded here (pdf)

Source: TorrentFreak, for the latest info on copyright, file-sharing and anonymous VPN services.

The Health Status of Net Neutrality – The Operators’ Impact on Internet Traffic is their new report which looks at how suppliers of fixed and mobile Internet services in Sweden, traditionally some of the fastest in the world, regulate the flow of traffic in their networks.

The plan was to discover if throttling activities exist among operators, and if so to assess how feasable it is to accurately measure it, and then decide whether it was worth continuing with and expanding upon the project.

Tests were conducted by .SE on the services of a dozen ISPs and measurements were taken for three different types of traffic – standard web browsing, file-sharing and video (such as YouTube).

“What is evident from the measurement results is that some mobile operators systematically downgrade user traffic such as the file-sharing protocol BitTorrent,” says Jörgen Eriksson.

Eriksson, who had responsibility for conducting the tests, says at least one ISP blocked all incoming connections to torrent clients.

The report notes that interfering with BitTorrent is a bad idea, since much open software distribution relies on it. Furthermore, messing with P2P protocols in general is problematic since other services such as Skype, Spotify and Voddler use them.

“If an operator attempts to limit these protocols and the operator’s customers know that their Internet connection does not give them full access to this type of service the operator will lose customers,” says the report.

In Sweden, service providers are free to restrict traffic providing they comply with certain conditions, but the report criticizes ISPs for their lack of transparency.

“The most interesting conclusion is that it is very difficult, if at all possible, to find information among operators about what they block or prioritize,” says Eriksson.

“We know that mobile market players see it as an advantage to NOT be compared with others. There is thus a risk that even if the technical information is presented, it will be useless for those who do not have a deep understanding of how the Internet is built.”

For now and until .SE’s next and more comprehensive report, the tested ISPs will retain their anonymity giving them time to reconsider their strategy, which given wider considerations might not be a bad idea.

Restricting end-users’ access to peer-to-peer based services could have a knock-on effect to the wider Internet. Peer-to-peer protocols help to distribute traffic online, an improvement upon older and more bandwidth intensive models.

“If peer-to-peer protocols are blocked so the trend will go toward developing protocols according to the traditional server-client model, or data will be hidden in other traffic where it is difficult to discern,” the report adds.

“It will probably not be as effective and lead to an increase in traffic – rather than the decrease as ISPs seek when they block peer-to-peer protocols,” the report concludes.

Full report here (Swedish, pdf)

Source: TorrentFreak, for the latest info on copyright, file-sharing and anonymous VPN services.

The interactive data set published yesterday spans a two-year period and this initial release covers the period between April 2008 and May 2010. It includes BitTorrent throttling percentages of ISPs in dozens of countries, divided into three-month periods. Below we discuss a few trends and notable findings.

United States

The BitTorrent throttling practices of Comcast, exposed by Robb Topolski and TorrentFreak in 2007, were in part what led to the Measurement Lab research. After an FCC investigation Comcast was ordered to stop slowing down BitTorrent on a large-scale, and the data shows that the company has kept its word.

Early 2008 Comcast limited nearly half (49%) of all BitTorrent traffic but this was reduced to 3 percent by the first quarter of last year. Cox, another heavy throttler, went from 51 percent to 3 percent in the same time period. The data further shows that in 2010, Clearwire was the only U.S. Internet provider that limited more than 10 percent of all BitTorrent traffic, 17 percent to be precise.

Worst: Clearwire (17%)

Best: Comcast and others (3%)

Canada

In Canada, all large ISPs have admitted to slowing down BitTorrent traffic, and some do so to a great extent. Since the start of the measurements Rogers has continuously throttled more than three-quarter of all BitTorrent traffic, and there are no signs that this will stop.

During the first quarter of 2010 the two other large Canadian ISPs, Bell and Shaw, were throttling 16 and 14 percent respectively. Videotron on the other hand has never slowed down more than 7 percent, and only 3 percent during the last measurement year.

Worst: Rogers (78%)

Best: Videotron (3%)

Great Britain

In Great Britain, TalkTalk used to limit a third of all BitTorrent traffic, but this was reduced significantly by the end of 2009. They now only slow down BitTorrent during peak hours which resulted in a 12 percent throttling rate early 2010. Tiscali and BT Group are exposed as the most heavy throttlers while Virgin Media, O2 and BSkyB have had relatively low percentages throughout the measurement period.

Worst: Tiscali and BT Group (27%)

Best: BSkyB (5%)

Other

A quick look at some other countries shows that in Australia none of the large ISPs were throttling BitTorrent traffic heavily in 2010, and the same can be said for Sweden and France. In The Netherlands UPC used to throttle heavily, but this was no longer the case early 2010.

In Germany, Kabel Deutschland seems to be the poorest choice for BitTorrent users (36%), and in Poland UPC has to be avoided as they limit 87 percent of all BitTorrent traffic.

Those who are interested in seeing how their own ISP performs can take a look at the full dataset at deeppacket.info. The researchers promise to release more recent data in the future, and it will be interesting to see how the various throttling habits of ISPs develop.

For those who have a choice, which us unfortunately not always the case, the data can definitely help to make an informed decision when signing up at a new Internet Provider.

Source: TorrentFreak, for the latest info on copyright, file-sharing and anonymous VPN services.

As soon as the new service went live TalkTalk subscribers were noticing some strange behavior. During the day they were able to play games just fine, but after 6 in the evening, they were all of a sudden unable to connect. Then, after midnight the connection problems suddenly disappeared and games were loading just fine.

Initially some thought it could be that OnLive couldn’t keep up with the demand, but when the same pattern repeated during the following days it became apparent that something was seriously wrong. This was confirmed by OnLive’s support desk, who observed that the users with problems were facing excessive packet loss on their connections.

So what’s going on here?

Based on the systematic problems between 6 PM and 12 AM, it appears that TalkTalk’s P2P throttling application is also cutting off OnLive. TalkTalk are very open about their traffic management practices, and they do indeed limit P2P traffic during the exact times OnLive users are facing problems.

OnLive uses UDP connections to transfer game data, and it seems that TalkTalk’s traffic shaping equipment mistakes this for P2P traffic.

TorrentFreak reached out to TalkTalk to find out more about the issue, as users in the ISP’s support forums were unable to find out more. Yesterday we got a response and TalkTalk acknowledged the problem, but a fix is yet to be found.

“This has been raised with our network team and we are currently investigating,” a TalkTalk spokesperson informed us. Meanwhile, TalkTalk subscribers are growing impatient as they feel that they are being ignored.

“I basically feel that my money spent on OnLive is being wasted. Working from 8:00 till 16.30 I have no use for this service as I would have wait until midnight to have access. Which is unrealistic, people need to have a good sleep, right?” OnLive user and TalkTalk subscriber Dawid told TorrentFreak.

And somehow we feel that OnLive is not very happy with the issue either, as they may lose customers who think that the cloud gaming platform is simply offering a bad service. A few users of other ISPs have been reporting issues as well, but only at TalkTalk are they still persisting more than a week after OnLive launched.

While we understand that problems can always occur, the above clearly illustrates the dangers of traffic shaping. There’s always going to be collateral damage. In the case of OnLive it is significant enough to be noticed and (hopefully) fixed, but what happens if a smaller company is affected? Hopefully TalkTalk will come forward with an official apology and more details on their UDP throttling, so this debacle can be prevented in the future.

Update (October 3, 2011): TalkTalk gave the following comment to TorrentFreak

“Unfortunately the recently launched OnLive gaming service was incorrectly identified as a peer-to-peer application. We’ve changed this and are currently testing with customers. We apologise for the inconvenience this has caused.”

Source: TorrentFreak, for the latest info on copyright, file-sharing and anonymous VPN services.

In one of the class action suits that were brought about from the long-running incident, Comcast has now agreed to settle, meaning those affected may be eligible for compensation.

About two and a half years ago, reports surfaced on what appeared to be the throttling of BitTorrent connections by Comcast. The throttling, first discovered by Rob Toplowski, was confirmed by TorrentFreak, and was first reported in August 2007. Other news agencies picked it up later, especially after tests by the EFF and Associated Press confirmed events and included reports that other network based activity was also affected.

Comcast long-denied any wrongdoing, but a leaked memo revealed that the company went as far as instructing its front-line staff to lie about the issue. Then the FCC got involved and things deteriorated. At a hearing at Harvard, Comcast packed the venue with people they bussed in, but it didn’t stop the BitTorrent throttling practice being termed a ‘hacker technique’.

Eventually, some assurances were made, and the FCC ordered Comcast to stop using Sandvine. Meanwhile lawsuits had been filed. One of these, Hart vs Comcast of Alameda, attained class action status, and there is now a proposed settlement.

Comcast has agreed to put $16M into a fund to pay BitTorrent users that were inconvenienced by the ‘network management’.

The downside is the size of the settlement. If you qualify, you can receive a maximum of $16, yet still Comcast refuses to accept it did anything wrong. The administrators of the settlement have set up a website to deal with questions about the case which can be found at www.p2pcongestionsettlement.com

It’s not hard to imagine that some customers will feel this doesn’t go far enough, and undoubtedly the discussion on this topic will continue. For the affected Comcast users there is still time to decide how to proceed – the deadline for claims is August 14th 2010. Meanwhile, network neutrality remains a pipe dream for most people.

Source: TorrentFreak, for the latest info on copyright, file-sharing and anonymous VPN services.

“The lesson of iTunes and Spotify is that what people want is ease of use and convenience and cheapness. And you only have to look at the decrease there has been in filesharing since the increase in popularity of Spotify.
“You only have to look at the number of people who came off illegal filesharing when iTunes came out to know that filesharing isn’t the answer, it’s not the future, it’s not valuable of itself – it’s a technology that currently is being used to circumvent the law.”

However, his words, as reported by the Birmingham Post, lack a certain ring of truth. For example, let’s take the claim that file-sharing decreased after Spotify gained popularity. While we agree that the service has the potential to convert many music pirates, its effect on the overall volume of file-sharing is simply not there.

The Pirate Bay informed TorrentFreak that traffic from the UK is still growing, at an average of around 1% per month. Mininova likewise has seen a 15% growth, of 7 million unique visitors a month, from 38.6 million in September 2008, to 45.6 million in August of 2009. Clearly Spotify hasn’t decreased much. Strike 1.

More worrying though is the claim that file-sharing technology is not valuable. For one, Spotify itself is based on file-sharing technology, with the brain behind the popular BitTorrent client uTorrent as one of its main developers. That aside, the state-funded BBC is involved in various BitTorrent-based projects, and the technology chiefs there believe that P2P TV has a future, and many independent artists are already putting it to use.

Over the past few years, we’ve brought you dozens of stories about people being enabled by the technology, from independent artists, to filmmakers (large and small) and even large corporations that can now effectively distribute data without incredible bandwidth outlay. File-sharing technology is very valuable to those people, Strike 2 for the MP from Birmingham.

That brings us to another statement Simon made, with Yahoo reporting him as saying that whilst it is illegal, there is currently no anti-piracy legislation. Those that have read our stories about Davenport Lyons, and ACS (and the thousands they have targeted) know there is indeed legislation. So too does Alan Ellis, and the Oink uploaders who were sentenced earlier this year. For Mr Simon, that’s strike 3.

It’s just lucky for him that no-one’s proposed a law where if an MP has gone on the record and made 3 basic factual errors, his parliamentary benefits should be cut off or throttled.

Mr Simon was contacted for comment, but did not reply at time of press

Source: TorrentFreak, for the latest info on copyright, file-sharing and anonymous VPN services.

Regardless, Comcast is not the only ISP that is throttling. As was revealed in the stats from Project Glasnost, Cox is also throttling heavily. So, while some are popping the champagne corks over this victory, others are still working hard to keep our ISP’s honest, and ensure that their customers are getting what they paid for.

The latest of these, is a project called Switzerland by the EFF. Still in very early alpha, it’s an attempt to not just detect sandvineing by an ISP, but other forms of throttling as well. Unlike Glasnost, which uses a central server and known torrent streams to detect activities from the ISP interfering, this will use a more decentralised method, where peers running Switzerland swap information about the packets they send to other Switzerland users, in encrypted data packets sent via a central server. In effect, it’s a checksum of torrent activity sent via a 3rd party. As Peter Eckersley, staff technologist for the EFF, and developer of Switzerland puts it “Alice and Bob are exchanging packets, they connect to a neutral server (Switzerland) to arbitrate between their different views of the packets”.

When asked why the EFF started this project, and why they believe a neutral network is important, he told TorrentFreak: “There were several reasons why we started the Test Your ISP project, and designed and built Switzerland. One reason was pragmatic: we were trying to run systematic tests of the interference that Comcast was deploying against P2P networks, and we decided that the only sensible way to do that was to build an automated sensor network. So we set about doing that.”

“The bigger picture, of course, is that without transparency the Internet won’t remain the amazing open and innovative thing that it has been,” Eckersley says. “And EFF’s mission is to make sure that the Internet stays open and innovative. We need to shine lights into the dark corners of the network, and make sure that ISPs aren’t setting themselves up in some control room and saying “protocol A okay, but protocol B doesn’t fit with our business plans, so let’s give it second-class treatment or stop it from working entirely.”

Some might worry that the client might open up people to being monitored by anti-p2p companies or other undesirables, using the system as another method of verification, but there is really no way around it. The simplest method to avoid that is, in Peters words, “avoid exchanging copyrighted files between Switzerland clients. The copyright risks are probably lower if you run your own Switzerland server, but it’s still going to keep logs.”

The question of what the FCC will do about these other ISPs and their traffic management is one to ponder. Our inquiries on this matter have been acknowledged, but not replied to at the time of publication.

Source: TorrentFreak, for the latest info on copyright, file-sharing and anonymous VPN services.