Slashdot Mirror
Cable Packet Shaping Causing Slowdowns
knorthern knight writes "To counter P2P programs that encrypt their traffic to evade detection, Rogers Cable in Canada has apparently started degrading all encrypted IP traffic, according to a post on Michael Geist's blog. How many of you log in to work over a VPN or ssh-tunnel? How many get usenet news or email over an encrypted connection? This could be a problem for Rogers Cable customers. Geist, who teaches at U of Ottawa, has 'been advised that the University computer help desk has received a steady stream of complaints from Rogers customers about off-campus email service.'"
Shaw cable on the western side of Canada also mangles packets. Check with Vonage to find out how Shaw is trying to cripple their business by dropping calls, packets, or just dropping the network connection for people using Vonage VoIP.
Support NYCountryLawyer RIAA vs People
This is somewhat "broken". If you can't use https or ssh with an internet connection, then that particular internet provider is little more than a glorified TV. If anything, ssh and https should be the highest priority.
There are reasons why p2p systems have started encrypting their traffic. Due to popular discontent with bandwidth throttling, they are trying to classify their traffic with a group of services that cannot be removed without breaking the functionality of the internet for that service provider. So their ideal solution to that is to break the functionality of their internet connection?
Okay, I can see (from their perspective) how you wouldn't want someone who is paying the same as your other customers using 500x the bandwidth that they use. After all, you're paying for the bandwidth.
So why not simply SEGMENT your network and put those heavy users on their own block? If you're that worried about P2P crap, they're probably sharing amongst themselves anyway. This would make it easier for you.
So why not offer GRADUATED pricing levels? 2 GB/month for $x. 5 GB/month for $2x. 10 GB/month for $10x. You could even break it down to traffic that stays on your own network and traffic that reaches the Internet.
The whole thing about the opposition to "Net Neutrality" is about extracting the MAXIMUM profit from the existing infrastructure with the minimum of technological advancement. Fuck that. We have the technology right now to make this a non-issue in almost every case. They just don't want to use it because there is a chance they can make more money by crippling the system.
I am a telecommuter and I have certainly noticed the bandwidth decrease for encrypted traffic; at any given time, for my job, I absolutely have to have roughly 15 citrix-application windows open at any given time, and the only way to access the metaframe server is via a VPN connection (as per corporate security policy). I have noticed major, major slowdowns; it's unfortunate that I cannot do my work properly as a telecommuter due to this new procedure of Rogers. Don't get me wrong, everything still works properly, the only thing is that with this slow down of my Citrix sessions (due to the traffic being encrypted), I have learned to live with a "Click now, work later" style application behaviour; it reminds me of using a 486 PC.
No, not at all. The net neutrality debate is about whether ISPs can throttle content based on the content's particular source, not on the content type.
Throttling based on content type is called packet shaping, and it's been done in the US and elsewhere for many years. Nothing about the net neutrality legislation would affect that, and anyone who says otherwise is confused or trying to deliberately mislead.
Throttling based on source, where content of the same type from different sources receives different priorities, is what the net neutrality legislation is about. In other words, any ISP can choose to tone down streaming video traffic so that all their customers can use basic web and email services. No ISP should be able to block video streaming from Google but allow video to stream from Microsoft, just because Microsoft paid them money. (Unless that was clearly advertised to the ISP's customers before they signed up, that is.)
In this case, it sounds like the ISP is throttling all encrypted content, regardless of its source or destination, so the net neutrality concept doesn't apply at all.
It doesn't hurt to be nice.
No. They want to ADVERTISE an "unlimited" package so that people will leave their graduated plans and come over to the "unlimited" provider.
Whereupon the "unlimited" provider throttles encrypted communications. And whatever else for someone going over the maximum of the "unlimited" plan.
[i]Most people want to have unlimited traffic even if they have no concept of the amount of traffic they need.[/i]
Not really. Most people would rather save a bit of money. So the companies use deceptive advertising.
I'm saying that we need to force them to get rid of the deceptive advertising. There's no TECHNOLOGICAL reason for it.
They can sell "unlimited standard usage" packages that throttle connections after 2GB/month.
They can sell "unlimited gamer" packages that throttle connections after 5GB/month.
They can sell "unlimited pro" packages that throttle connections after 10GB/month.
The reason that they don't is that they can save MONEY by being STUPID and selling a single "unlimited" package and fucking with the connections so that things such as encrypted sessions are dead slow. It's about them being lazy. That is it.
wget http://autocast.ca/test.dat
Length: 10,485,760 (10M) [text/plain]
18:52:39 (539.62 KB/s) - `test.dat' saved [10485760/10485760]
wget https://autocast.ca/test.dat
Length: 10,485,760 (10M) [text/plain]
18:53:03 (560.59 KB/s) - `test.dat.1' saved [10485760/10485760]
No slowdown on https downloads at this moment from this location.
scp test.dat odin.canadacast.ca:/root/
test.dat 100% 10MB 97.5KB/s 01:45
scp odin.canadacast.ca:/root/test.dat .
test.dat 100% 10MB 602.4KB/s 00:17
No slowdown on that either.
Upstream rate is 97.5% of this cable modem's capability (800kbps)
This is on a saturday, at 7:10pm local time.
Not quite peak usage time of day but not 3am either.
This does not prove anything of course.
I've only failed to prove that there is traffic shaping, I have not proven that there is no traffic shaping.
Maybe I'll try again at a known peak traffic time.