ISP Throttling VPN?

[airvpnclient 13]

Currently connected to Sirius over SSL and still getting no better than 2 Mbps.  It seems that the traffic shaping on SSL is a little different, since the upload speeds are better than straight VPN without SSH or SSL.

[airvpnclient 13]

Summary of analysis:

 

AirVPN speeds over Bell Canada infrastructure are apparently limited by technical ITMPs to under 2Mbps.  Neither SSH, not SSL successfully evades ITMPs. 

 

The ITMPs for controlling SSL appear to be less unfavourable to uploading than the general VPN throttle.  This finding is independent of hardware on which OpenVPN is running and can be replicated.  The gateway router is a constant, and so could be the source of speed reductions, but this seems unlikely.

 

Bell Canada is not publicizing it's ITMPs and it remains to be seen if they are compliant with CRTC guidance in this regard.  I am assuming that my member-owned 3rd party service provider is not the source of IMTPs, but this is to be confirmed.

[GoXRUTD7k8FfZp5jmvNK4uE6 7]

After seeing this topic a bookmark came to mind that may aid you in determining if shaping is taking place.

[profits@nym.hush.com 0]

After seeing this topic a bookmark came to mind that may aid you in determining if shaping is taking place.

 

Running this application may be a security risk

 

I got the above warning....  Is it safe do proceed?

[GoXRUTD7k8FfZp5jmvNK4uE6 7]

 

After seeing this topic a bookmark came to mind that may aid you in determining if shaping is taking place.

 

Running this application may be a security risk

 

I got the above warning....  Is it safe do proceed?

 

I haven't used the site in a couple years. I assume that's a warning produced by a Java applet? If so, I'd accept it as that site's intent is not malicious I believe, however it's your choice.

[thar3tra 0]

I guess the security message is related to certain network-related commands contained in the javascript code. Jittertest (by speedtest.net) produces the same security message.

 

The test itself tests different patterns on different ports, including a test for SSH connections, so it would probably be useless for testing with your VPN active.

Quote from their research paper on how the test operates:

 

Our current implementation of Glasnost detects traf-

fic differentiation that is triggered by transport protocol

headers (e.g., ports) or packet payload. These triggers

are more common than IP headers

[felixthakat 1]

I was almost ready to give up but I gave the SSL tunnel option a try (https://airvpn.org/ssl/) over Bell fibe. It seems to work very well. Enjoy!

[AirBreathingNetzien 0]

Comcast in my area started throttling VPN big time (straight 57MB downstream, 54MB through Air, down to 1-3MB downstream ... they didn't touch the up stream).   Switched to SSL using OpenVPN and STunnel and I now get around 45MB downstream (still less than than the 54MB AirVPN downstream I started with, but acceptable considering the alternative).

[zz0i0XLpHPXbPOO4hv4GCumR-8 0]

Bell Fibe 50/10 here.  With OpenVPN running on my DD-WRT router I can only get 10/10.  It's throttled for sure.  I get exactly 10mbps download every time I test.  I suppose there's no way to run SSL through the VPN when running on a router?  If not this sucks.

[Staff 10014]

Bell Fibe 50/10 here.  With OpenVPN running on my DD-WRT router I can only get 10/10.  It's throttled for sure.  I get exactly 10mbps download every time I test.  I suppose there's no way to run SSL through the VPN when running on a router?  If not this sucks.

 

Hello,

 

many consumers' routers CPUs can't process more than 10 Mbit/s AES-256 throughput due to their processing power. Our Data Channel cipher is AES-256-CBC. Probably bottleneck is the router CPU. Try to connect directly from your computer to make a comparison.

 

Kind regards

[hashswag 11]

There needs to be a choice:  "Yes, my ISP heavily throttles all VPN Protocols/Options, requiring SSL encryption, slowdown is consistent with double encryption."

[mlp 10]

Comcast in my area started throttling VPN big time (straight 57MB downstream, 54MB through Air, down to 1-3MB downstream ... they didn't touch the up stream).   Switched to SSL using OpenVPN and STunnel and I now get around 45MB downstream (still less than than the 54MB AirVPN downstream I started with, but acceptable considering the alternative).

 

Similar on comcast in my area, but not as severe.  30 down to about 15 down.  Recent numbers here - https://airvpn.org/topic/12406-slow-connection-using-vpn/?p=21954 .  Again, another data point for best rates are over SSL.

[zz0i0XLpHPXbPOO4hv4GCumR-8 0]

Hello,

 

many consumers' routers CPUs can't process more than 10 Mbit/s AES-256 throughput due to their processing power. Our Data Channel cipher is AES-256-CBC. Probably bottleneck is the router CPU. Try to connect directly from your computer to make a comparison.

 

Kind regards

 

Thank you for sharing your expertise.  I installed the AirVPN software and get far higher speeds this way, which is good.  I thought my Netgear WNDR3700 v4 router with DD-WRT would be faster than most, but I guess not.  One day I'll buy a faster router because I prefer that solution over this software solution.  Thanks again.

[elumiknotty 0]

Comcast Xfinity customer here. I get about 270 Mbps down and 45 up when not using a VPN. I'd used PIA successfully for about a year, with about 150 Mbps down and 30 up, until about 2 weeks ago when my VPN connections dropped to 45 down and 2 up, no matter what ports I connected through. I did a ton of troubleshooting, including installing PIA on other devices with different operating systems.  Same exact capped speeds.  I had a suspicion my VPN traffic was being shaped by the ISP. After researching I thought SSL or SSH would work, so I gave AirVPN a try, and it worked great for a day on the 30000-some SSH port, and then suddenly it dropped to the same 2 Mbps upload (and between 10 and 30 down).  I've been trying to connect to tens of different 5-star servers through various SSL and SSH ports, and they are all capping at 2 Mbps up as well.

Anyone have any ideas, or any info as to what this could be if you don't think it really is Comcast?

[madrat 19]

I'm seriously not very knowledgeable about this stuff, however, using a Vancouver server and several different protocol settings, the best I could do was 3 Mbit/s. After switching to a California server, it immediately went up to 10 Mbit/s. Of course I have no idea at all why that is. Latency for both was around 40 ms. The final test I did on Vancouver was "automatic" protocol setting. I left it on that setting for the California server test.

[madrat 19]

I'm seriously not very knowledgeable about this stuff, however, using a Vancouver server and several different protocol settings, the best I could do was 3 Mbit/s. After switching to a California server, it immediately went up to 10 Mbit/s. Of course I have no idea at all why that is. Latency for both was around 40 ms. The final test I did on Vancouver was "automatic" protocol setting. I left it on that setting for the California server test.

I'm finding all this speed testing very interesting. I assume a lot changes from day to day. Today, the USA server I got 10 Mbit's on yesterday, today I get 3 Mbit's. Switching back to a Vancouver server, today I got 11 Mbit's whereas yesterday I got 3 Mbit's. The latency for both was about the same today as yesterday - between 35 & 45 ms. Does this happen to most users or have I possibly set up Eddie improperly?