Verizon, Comcast Say They Are P2P Friendly

An anonymous reader writes "Verizon and Comcast announced they will not 'block or throttle Internet traffic delivered via peer-to-peer networks' — essentially proclaiming that they are now P2P friendly. The decision came as a result of a test conducted with Verizon and Pando Networks, testing the benefits of a P2P/ISP partnership. During the test, the amount of P2P content delivered to Verizon subscribers from inside its network grew from 2 percent to 50 percent. This shows ISPs need to work with P2P companies to improve content delivery and manage traffic. Verizon also announced it will be looking at ways to use P2P technology to deploy new features on FiOS TV." Just the same, read on for one approach to mitigating likely tightening restrictions on P2P network use. Another anonymous reader writes "RIAA/MPAA have recently been targeting torrent aggregators like PirateBay, because the aggregators are the vulnerable components of the BitTorrent protocol. A new open-source project to thwart such attacks was announced on p2p-hackers and released yesterday:

Cubit, a new open-source p2p overlay, enables the Azureus BitTorrent client to look up torrents via approximate keyword search... Cubit completely decentralizes the lookup process through an efficient, light-weight peer-to-peer overlay that can perform approximate matches. It performs searches without relying on any centralized components, and therefore is immune to legal and technical attacks targeting torrent aggregators."

Comments

Re:I'll believe it when...

[frooddude]

If you're on Comcast that speed change could just be the effects of Speedboost. They give a short term bump in throughput for each new transaction.

I rarely get good torrent speeds unless I'm dealing with a highly transacted image. Like a new release of ubuntu.

Re:Even 100% is not good enough...

[CyprusBlue113]

I don't think you quite understand what is being talked about. A truly perfect P2P system would only need 1 copy period to come in to an ISP. Now you will never see much anything close to this, but it can definately be a *much* better situation than it is now.

Re:Right...

[kernelphr34k]

I'm a FIOS customer, and have yet to see any issues with my torrents and disconnects, or any speed or BW issues. It does help to have a 15mb/15mb connection, but still. Curious to see how these companies will handle the P2P load. . .

The article meshes with my experience

[dave562]

I've been a Verizon DSL subscriber since the late 1990s, back when they were still GTE. They have had constantly good service and great uptimes. I started using torrents about a year ago and have never had any problems. I have one running at home right now. On my 1.5/384 line I'm getting about 170k down and 40k up, constantly.

It has been my experience that in some ways DSL is superior to cable. I remember when cable first came out everyone who got it thought it was great. Then their neighbor got it, and their other neighbor got it, and suddenly it became obvious that the entire neighborhood was on one shared pipe and a single bandwidth hog could ruin it for everyone. It doesn't seem like much has changed in the last decade. With DSL you can count on getting the bandwidth that you pay for but the peak available bandwidth isn't as high as cable. On cable you might get some really high peak speeds, but the cable networks haven't been designed to sustain high transfer rates for long periods of time.

Re:The article meshes with my experience

[Shakrai]

At some point, all your DSL connections are aggregated somewhere and that aggregation point becomes the bottleneck.

It seems like every time we have this discussion that someone repeats this half-truth and gets a +5 out of it. Yes, DSL connections are aggregated somewhere. But that's not the whole story.

There's nothing technical stopping a telco from having a 1:1 contention ratio if they deem it in their best interests. Contrast that to cable -- the only way to attain a 1:1 ratio on cable is to segment the network into insanely small slices or devote more channels on the coax plant to HSI services. DOCSIS 2.0 only offers ~42Mbits of downstream -- assuming 5Mbit connections (the standard for Roadrunner around here and actually quite low compared to other areas) it only takes nine people to completely saturate the downstream pipe.

Even without a 1:1 contention ratio it's going to take a lot more than nine customers to peg the backhaul connection from your local DSLAM.

I used to have DSL and I found my connection would degrade noticeably in the late afternoon and evening simply because we had a lot of people in the area connected with lots of kids.

As with anything, YMMV. I've never seen a slowdown in six years of working with Verizon and Frontier (a smaller telco based out of Rochester). I have seen them occur on Roadrunner -- in some neighborhoods around here it's downright painful when the college kids are in town.

Re:The article meshes with my experience

[mrchaotica]

I moved mine to channel 11, a non-interfering channel, and wah-lah, performance problem solved.

It's voilà, damnit!

I call "Bullshit" on Comcast

[Duncan Blackthorne]

..and here's the proof (as of just a few minutes ago), courtesy of the Glasnost test:

Is BitTorrent traffic on a well-known BitTorrent port (6881) throttled?

* 2 out of 2 BitTorrent transfers were interrupted while uploading (seeding) using forged TCP RST packets. It seems like your ISP hinders you from uploading BitTorrent traffic to our test server.

* The BitTorrent download worked. Our tool was successful in downloading data using the BitTorrent protocol.

* There's no indication that your ISP rate limits your BitTorrent downloads. In our tests a TCP download achieved minimal 713 Kbps while a BitTorrent download achieved maximal 720 Kbps.

Is BitTorrent traffic on a non-standard BitTorrent port (4711) throttled?

* 2 out of 2 BitTorrent transfers were interrupted while uploading (seeding) using forged TCP RST packets. It seems like your ISP hinders you from uploading BitTorrent traffic to our test server.

* The BitTorrent download worked. Our tool was successful in downloading data using the BitTorrent protocol.

* There's no indication that your ISP rate limits your BitTorrent downloads. In our tests a TCP download achieved minimal 661 Kbps while a BitTorrent download achieved maximal 741 Kbps.

Is TCP traffic on a well-known BitTorrent port (6881) throttled?

* There's no indication that your ISP rate limits all downloads at port 6881. In our test, a TCP download on a BitTorrent port achieved at least 713 Kbps while a TCP download on a non-BitTorrent port achieved at least 661 Kbps.

* There's no indication that your ISP rate limits all uploads at port 6881. In our test, a TCP upload on a BitTorrent port achieved at least 1353 Kbps while a TCP upload on a non-BitTorrent port achieved at least 1403 Kbps.

Re:Even 100% is not good enough...

[bconway]

In practice, many (most?) ISPs use transparent HTTP caches, so having 50% of the data stay internal is still no good, as on popular files (eg, a big youtube video), 99% of the traffic stays internal for HTTP. No they don't. Start here.

Confirmed today: Comcast, Verizon (DSL + FiOS), Time Warner, and Speakeasy.