ANSWERED First time user

[scgt1 3]

Just signed up last night and in haste did a year sub. Looking like a completely terrible mistake from speed.

 

The client for some reason thinks I need to connect to Canada while there is a server 20 minutes from me in Dallas, TX. So that aside I'm manually connected to my local server for 75% less latency and just a little more speed. My real problem is the actual speed I have. I'm not even getting half what I pay for through Airvpn and if this is normal then yea time for a refund. If there are settings I can change to get my speed to what I'm paying for that would be great but I have no idea what I'm doing with these settings so everything is on default in the application.

 

With AirVPN

 

Without AirVPN on my normal connection

 

 

[zhang888 1066]

This is a good result, and puts you in the top 10 list of all Air users. This is expected on regular hardware

and the current single core OpenVPN implementation.

 

Not sure there are many providers that can guarantee and actually provide such speeds.

[Guest]

It highly depends on your settings, I have had a few speed limits happen to me too but I changed the settings, and buffer sizes. But keep in mind like zhang said it can only use one core and since the encryption is rather heavy depending on the speed of your CPU per core you can get limited, after I fixed all my stuff I hit 136.26 Mbit / sec easy

[OpenSourcerer 1441]

Another thing to consider is, although the next server geographically might be 20 minutes away from you, it doesn't need to be the technical case. BGP might route you through some other servers thousands of kilometers away, and every hop decreases your speed. I reach full speeds just because I know which servers have got a direct BGP route to my ISP. This habit of choosing my servers never surprised me. Full speeds, no latency loss.

[Staff 10014]

Hello!

 

The performance you get is remarkably good. Our infrastructure is not sized to provide much more than that: consider the price and the fact that each server is connected to a 1 Gbit/s line. Out of curiosity, what are your bandwidth allocation per client expectations for less than 5 EUR per month? See again the Terms of Service you explicitly declared to have read and accepted to see the current guaranteed bandwidth allocation.

 

Please note that to the best of our knowledge no consumers' VPN provider in the world is currently able to offer the same, not even for a higher price.

 

That said, since there is still plenty of free bandwidth in the infrastructure, you can probably get higher performance (only if you have a powerful CPU and your ISP does not shape traffic) by enlarging OpenVPN sockets buffers sizes. Set them to 256 KB. Additionally, do not trust too much speed tests. They just measure the peak performance in a short interval of time in a single connection.

 

To pick a specific server of your choice, use the "Servers" tab. You can also define a white list of servers. Eddie will connect only to a server included in a white list, when one is defined.

 

Kind regards

[scgt1 3]

That said, since there is still plenty of free bandwidth in the infrastructure, you can probably get higher performance (only if you have a powerful CPU and your ISP does not shape traffic) by enlarging OpenVPN sockets buffers sizes. Set them to 256 KB. Additionally, do not trust too much speed tests. They just measure the peak performance in a short interval of time in a single connection.

 

To pick a specific server of your choice, use the "Servers" tab. You can also define a white list of servers. Eddie will connect only to a server included in a white list, when one is defined.

 

At the time of testing it showed 4**/1000 currently it's 6% 56/1000 for Auva. It may be getting dated but I wouldn't declare it an unpowerful cpu (3770K) I have no idea if my ISP shapes traffic as you say. I have never gotten degrated speed in the 5 years I've been with them. The enlarging the buffers to 256 is under prefs/advanced and is the send/receiver buffer size at the bottom correct? I have the Auva set as white list so it will connect to it now instead of somewhere in Canada.

 

Should any port settings be ticked or anything? As my OP stated I'm not network savy. Go figure I've been building computers since the early 90s but when it comes to networking stuff I'm a lost cat.

[go558a83nk 364]

To the OP: For USA users if the star rating column is set to weight towards speed Canadian servers are put up top.  If weighted towards latency then you'll see what you want to see.  Staff will have to say why it is this way, I don't know.

 

It's possible to squeeze more speed, assuming you have a modern CPU running in the background. I often see users in the top 10 breaking 100mbit/s and have rarely seen over 200mbit/s.   Fiddle with ports, protocols, buffers, and servers.  Giganerd is right, closer (as the crow files) isn't always better.

 

I would start with TCP 443.  I'm guessing your testing was done with UDP 443.

[Revelead 16]

This is really fast to be honest. Im maxing my speed 50/10 which is incredible good. Try out a provider which costs double the price of airvpn and tell us if you get better speeds

[scgt1 3]

To the OP: For USA users if the star rating column is set to weight towards speed Canadian servers are put up top.  If weighted towards latency then you'll see what you want to see.  Staff will have to say why it is this way, I don't know.

 

It's possible to squeeze more speed, assuming you have a modern CPU running in the background. I often see users in the top 10 breaking 100mbit/s and have rarely seen over 200mbit/s.   Fiddle with ports, protocols, buffers, and servers.  Giganerd is right, closer (as the crow files) isn't always better.

 

I would start with TCP 443.  I'm guessing your testing was done with UDP 443.

It was on the default automatic.

[scgt1 3]

Connected to Chicago 120ms lat UGH. Went back to Dallas just a fluke?

 

This is with the suggested forced TCP 443 instead of auto.

[go558a83nk 364]

Connected to Chicago 120ms lat UGH. Went back to Dallas just a fluke?

 

This is with the suggested forced TCP 443 instead of auto.

 

auto uses UDP.  So if this is with TCP then speed difference is likely not a fluke.  I find TCP gives faster speed when the server is "close" to me.