Can Anyone Help Me Determine If My Isp Is Shaping Traffic?

[poopypants]

i have run through the glasnost test and it says that there is no shaping happening. i dont know if thats true or not. my problems started cropping up when i switched to my new computer (so im not sure if its related or just a coincidence). my download speed is pretty solid (a bit slower than it has been but still hitting over 2MB/s on lots of torrents) however my upload speed is shit. i have been dealing with some of the guys over at the utorrent site but so far we got nothing. there is a tech coming out to inspect my connection on sunday. so my initial question is, can anyone here help me determine if there is some sort of shaping happening? i am willing to go through the motions of troubleshooting it (hell, i have been for the last week or so and i have come up with nothing) and what not if someone can help me figure this out.

edit: in my utorrent log i am constantly getting disconnect:peer offline error on the ubuntu and open office torrents, the same happens with the backtrack torrent

Edited April 23, 2010 by poopypants

[Sparda]

The answer is quite simply yes, every (consumer) ISP has to do it other wise there would not be enough bandwidth for every one.

I'm not sure the question you asked is quite in-line with what is happening. Connections been dropped is different to traffic been throttled.

it is entity possible that it is to do with your new computer. There are some network cards that hate the way particular types of traffic of applications use them. For example a friend of mine had to 'downgrade' from a netgear gigabit NIC to a generic 100Mbps card because when ever the WoW bittorent updater ran no other network-ing applications would work. If you know (as far as it is possible to tell) there is nothing wrong at the software layer, try the hardware layer. Admittedly I haven't seen this type of problem recently. The WoW problem only became a problem when WoW changed to the bittorent update client.

[VaKo]

When you say "shit" exactly how shit do you mean? Unless you have an SDSL connection, you are on an asynchronous connection which for most consumer level ISP's means you will have an upload speed of 256 - 512KBits/s, and on business level connections this might be higher (mine is 1MBit/s as I pay extra for a business account). You also have to factor in contention ratio (usually 50:1 for consumer and 20:1 for business) as you will be sharing a DSLAM card and its virtual path with multiple users, so if you happen to be connected along side a bunch of other people running bittorrent, VOD or similar services your connection will be slower than if your DSLAM neighbours are people who just facebook and email.

This is where the network neutrality arguments fail. In order to provide a decent service ISP's will have to put QOS measures in place, including DPI techniques. The only problem I have with this is making it a service to profit from rather than viewing it as a network maintenance issue.

As for testing it yourself:

http://broadband.mpi-sws.org/transparency/bttest.php

http://www.cc.gatech.edu/~partha/diffprobe/shaperprobe.html

[poopypants]

When you say "shit" exactly how shit do you mean? Unless you have an SDSL connection, you are on an asynchronous connection which for most consumer level ISP's means you will have an upload speed of 256 - 512KBits/s, and on business level connections this might be higher (mine is 1MBit/s as I pay extra for a business account). You also have to factor in contention ratio (usually 50:1 for consumer and 20:1 for business) as you will be sharing a DSLAM card and its virtual path with multiple users, so if you happen to be connected along side a bunch of other people running bittorrent, VOD or similar services your connection will be slower than if your DSLAM neighbours are people who just facebook and email.

This is where the network neutrality arguments fail. In order to provide a decent service ISP's will have to put QOS measures in place, including DPI techniques. The only problem I have with this is making it a service to profit from rather than viewing it as a network maintenance issue.

As for testing it yourself:

http://broadband.mpi-sws.org/transparency/bttest.php

http://www.cc.gatech.edu/~partha/diffprobe/shaperprobe.html

here is what i mean shit.

this is what the connection speed WAS

then i started getting this

granted that was from much farther away but it was the closest. recently a server in tokyo has opened back up and now i get about 54 up and 54 down. i am using a fiber optic 1gigabit connection. japanese internet is always fast. what i mean by shit is that i no longer get any upload speed. a lot of the same rules that apply in the states do not really apply in japan. they are pretty much in their own world when it comes to the internets. so anyhow, i will use those test sites that you listed when i get home today and see what happens.

[Infiltrator]

When you say "shit" exactly how shit do you mean? Unless you have an SDSL connection, you are on an asynchronous connection which for most consumer level ISP's means you will have an upload speed of 256 - 512KBits/s, and on business level connections this might be higher (mine is 1MBit/s as I pay extra for a business account). You also have to factor in contention ratio (usually 50:1 for consumer and 20:1 for business) as you will be sharing a DSLAM card and its virtual path with multiple users, so if you happen to be connected along side a bunch of other people running bittorrent, VOD or similar services your connection will be slower than if your DSLAM neighbours are people who just facebook and email.

This is where the network neutrality arguments fail. In order to provide a decent service ISP's will have to put QOS measures in place, including DPI techniques. The only problem I have with this is making it a service to profit from rather than viewing it as a network maintenance issue.

As for testing it yourself:

http://broadband.mpi-sws.org/transparency/bttest.php

http://www.cc.gatech.edu/~partha/diffprobe/shaperprobe.html

Say that the ISP is shaping his traffic, is there a software that can be used to bypass the ISP shaping technique?

[poopypants]

When you say "shit" exactly how shit do you mean? Unless you have an SDSL connection, you are on an asynchronous connection which for most consumer level ISP's means you will have an upload speed of 256 - 512KBits/s, and on business level connections this might be higher (mine is 1MBit/s as I pay extra for a business account). You also have to factor in contention ratio (usually 50:1 for consumer and 20:1 for business) as you will be sharing a DSLAM card and its virtual path with multiple users, so if you happen to be connected along side a bunch of other people running bittorrent, VOD or similar services your connection will be slower than if your DSLAM neighbours are people who just facebook and email.

This is where the network neutrality arguments fail. In order to provide a decent service ISP's will have to put QOS measures in place, including DPI techniques. The only problem I have with this is making it a service to profit from rather than viewing it as a network maintenance issue.

As for testing it yourself:

http://broadband.mpi-sws.org/transparency/bttest.php

http://www.cc.gatech.edu/~partha/diffprobe/shaperprobe.html

just got home. as i said previously, the glasnost (your first link) test doesnt report anything wrong. the speed is incredibly slow but it doesnt say that there is any throttling. the second shaper also shows no throttling but the speed is ridiculously low for my connection.

DiffProbe beta release. October 2009. Build 1002.

Shaper Detection Module.

Connected to server 38.102.0.87.

Estimating capacity:

Upstream: 54359 Kbps.

Downstream: 82356 Kbps.

The measurement will take upto 2.5 minutes. Please wait.

Checking for traffic shapers:

Upstream: No shaper detected.

Median received rate: 53274 Kbps.

Downstream: No shaper detected.

Median received rate: 76645 Kbps.

For more information, visit: http://www.cc.gatech.edu/~partha/diffprobe

[VaKo]

I love your definition of slow lol.

What router are you using btw? What CPU does it run?

[Charles]

Wow, if 20Mbps is slow, I wonder what fast would be considered.

I think VaKo is on the right track, what type of router do you have?

I know there are some that cannot handle super high speed links.

Edited April 24, 2010 by Charles

[poopypants]

I love your definition of slow lol.

What router are you using btw? What CPU does it run?

lol when you go from 93 to 20, you notice :P.

my router isnt the problem, but for those wondering its a buffalo wzr-hp-g300nh.

like i said, this started to occur when i got my new computer so im wondering if its something with the actual hardware of this machine. even though it shouldnt be as its an alienware so its designed for doing crazy shit as is.

[VaKo]

I dunno, I started to notice through put issues with a WRT54G on a 20MBit connection. Not saying this is the issue but it was one I definitely noticed. Poor things CPU just couldn't handle QOS on that much data. I now use a 500Mhz AMD Geode based ALIX board.

As for your PC, what changed? XP to Vista/7? Hardware etc? Did anything else change at that time? What you need to do is be forensic about this, work through every single change and eliminate it. For example, install Ubuntu and XP to a seperate HDD and test. Get a seperate system to test with. make sure TCP offload is enabled etc. Was it just the PC or did your connection change?

[poopypants]

I dunno, I started to notice through put issues with a WRT54G on a 20MBit connection. Not saying this is the issue but it was one I definitely noticed. Poor things CPU just couldn't handle QOS on that much data. I now use a 500Mhz AMD Geode based ALIX board.

As for your PC, what changed? XP to Vista/7? Hardware etc? Did anything else change at that time? What you need to do is be forensic about this, work through every single change and eliminate it. For example, install Ubuntu and XP to a seperate HDD and test. Get a seperate system to test with. make sure TCP offload is enabled etc. Was it just the PC or did your connection change?

well, i went from an xps 420 (2.4ghz quad core processor) running windows 7 home premium x86 to an alienware m15x with an i7 720 running windows 7 ultimate x64. thats the only real change thats happened, granted its a big one. on the old computer from what i was noticing everything seemed to be working fine. i have been trying to be forensic about it, bypassing routers, etc. i was thinking that maybe ipv6 and ipv4 might be conflicting somehow on this new puter. i dunno if the old puter had the ipv6 settings or what. the tech from one of the isps will be here in about 30 min so we will see what he comes up with.

edit: so they just came and went and said there was no problem, but obviously there is. we called the ISP and asked if they were throttling because i am a heavy user and they said no. but its clear something is amiss here. any thoughts guys? this is driving me crazy.

Edited April 25, 2010 by poopypants