Slow VPN Speeds - What we are really asking about

BlueCat57

Distinguished
Apr 7, 2009
430
4
18,815
I recently tested my VPN speeds with SpeedTest .net and found that the speeds I get when using an overseas exit server were way below what I get when using domestic exit servers.

That led me to search for ways to speed up my VPN connections. Based on the questions and answers I've found, either I am the only person who doesn't give a cr@p about domestic exit servers, or more likely, most of the people complaining about slow VPN speeds don't realize that they and those that are answering their question are talking about two different things. The askers are asking about overseas exit servers and the answerers are answering about domestic servers.

Most of the article I have found touting the "fastest VPN services" are only talking about domestic exit servers and don't even give hard comparisons charts showing the actual numbers from the tests they supposedly did to determine the winners.

I know that in my case the question I am asking is specifically about overseas exit servers and if there is a way to increase my speed when using them.

I use VPN Unlimited, to which, for reference, you can get a lifetime subscription for under $40 if you shop around for deals. Note that lifetime service is different from infinity service. Infinity is for more devices that can use your subscription.

I have no complaints about VPN Unlimited's service. They have exit servers in the countries I want them in. And on the two or three occasions that I contacted them via phone and email their responses were fast and helpful. Most of those occasions were regarding purchases, not their service.

So, here is what I really want to know. Is there a way to increase the connection speed when using overseas exit servers?

I use my VPN service primarily to get around geographical restrictions. I'm guessing that many of those complaining about slow VPN speeds are actually complaining about overseas exit servers and not a domestic connection. Based on my very limited testing (as in one test), my domestic speed dropped from 30 Mbps to 22 Mbps. A big drop but still fairly fast. When using a WiFi connection the speed drop was even less. It is entirely possible that I have something set up wrong in the PC I am using. I'll be troubleshooting that possibility and report back.

I understand how exit servers work and that the server's traffic volume affects my speed. Most of the servers I use normally have under a 40% workload. Maybe that is high enough to cut my speeds to 10 Mbps or lower, but even the exit server in Australia with a 5% workload doesn't provide any more speed.

Maybe it is my VPN service, but I can't guess at that since all of the articles and forum posts focus on domestic exit server speeds. And those don't appear to take as big a hit.

Ping times on those overseas exits bounce all over the place. I did multiple tests on my main system and at first though a longer ping time meant slower speeds, but that turned out to be false.

So, I'm turning to the amazing crowdsourcing solution providers of Tom's Hardware for the answer.
 
Solution
A large part of the issue is the concept of "exit server" is not valid. It is very rare to find a vpn service where you would connect to a local data center, be carried over the vpn provider network to another country and they exit out a different data center.

It would be better to think of what you are calling a exit server as a simple home NAT router. Your traffic goes over the normal internet to the VPN router. get a new IP, then uses the normal internet as it leaves the router.

Lets say the server you want to access is in the same data center as the VPN provider and both use the same ISP. In that case you should get about the same performance other than the 10% overhead you lose because of the encryption. Now lets say the...
Of course an overseas exit point returns slower speeds and terrible ping. If you're in the US, exiting in Australia and connecting to lets say Google. That means your going across the entire pacific ocean, lord knows how many exchange points to get to Australia then up to Singapore to get Google then back to Australia and finally back across the Pacific to the US.

There is huge overhead doing this. Even without a VPN you'll see a noticeable slowdown in connection speed when contacting a server overseas.

If you want the best speed (other than trying every VPN one by one to see if one has a better overseas connection) the only thing you can do is only connect to exit servers that you know are in the same country as the actual servers you are trying to access in the first place (hint don't use Australia, its at the end of a dirt road in Internet terms).
 
Well, certainly there are several factors, when using oversea connections you go through several connections in between and thus you're bound to the slowest one in them.

Lets say you connect to Australia, between you and Australia you could hop through 8 different connections and if the slowest one among them is of only say 15mbps, that will be your max speed even if all the other connections are 1gbps.

Something at your side that could be affecting is your router, "tunnelizing" and ciphering all the data you send and receive through the vpn connection is quite hard on a common router since most of them do it at a software level, a more expensive router designed specifically for VPN traffic will do it at a hardware level working it much faster without the need of any sacrifice.

Besides that a bad configured MTU on your router could result in a loss of bandwidth, usually the default value is 1500 which automatically gets reduced to 1492 when using a VPN. Some routers have the "Jumbo frame" option which ups this value to 200000 and more, causing a contradictory work on the router (this jumbo frame is also used on Windows since Windows 8.1 I think, but don't quote me on that) so that's something to look at as well.
 
A large part of the issue is the concept of "exit server" is not valid. It is very rare to find a vpn service where you would connect to a local data center, be carried over the vpn provider network to another country and they exit out a different data center.

It would be better to think of what you are calling a exit server as a simple home NAT router. Your traffic goes over the normal internet to the VPN router. get a new IP, then uses the normal internet as it leaves the router.

Lets say the server you want to access is in the same data center as the VPN provider and both use the same ISP. In that case you should get about the same performance other than the 10% overhead you lose because of the encryption. Now lets say the server uses a different ISP. You might have a poor path via the internet to get to the server but a good path to get to the vpn data server because it is 2 different ISP. In that case you would get better performance going the vpn service hopping off and going over to the target server....assuming there is good interconnect between the ISP in the building which is not always the case. The reverse can also happen where you have worse path to the vpn data center than directly to the target server.

This is extremely hard to figure out, you would have to dig around and find out what ISP(s) each server you want to use and there relationship to your actual ISP.

Now there are a few VPN providers that do actually provide a network. These are mainly targeted and optimized for gamers. In this case the VPN provider has purchased a optimum path between their data centers. This more or less runs over the same fibers paths but because it is private connection they can get guaranteed service levels. These are some of the most expensive vpn providers to use because you must pay more so the provider can recover the costs of these data connections between their data centers. I forget exactly which ones still have their own network. There are a bunch of fake ones that try to claim they have a private network but do not really. These companies that tell lies like this is what makes finding good vpn services hard.

Still the bottom line is you get what you pay for. The cheap/free services will be very much inferior to ones you must pay a monthly fee for.
 
Solution

BlueCat57

Distinguished
Apr 7, 2009
430
4
18,815
I did see something about MTU but haven't had a chance to read about it yet.

I'm certainly a believer in "you get what you pay for." I just expected a bit more of a "miracle" with a run of the mill VPN service.

I just got a new WiFi router and all the advertising talks about speed and how fast they are. But after doing some testing I figured out that what most people really need is better signal penetration. The "g" standard is fast enough for many people, and "n" certainly is faster than most people's "advertised" speed. But what people really want is to put ONE router in the most CONVENIENT place in their home and get a signal everywhere.

Articles about VPN services tend to focus on speed rather than explaining that "10% overhead for encryption" or the subtleties of going through multiple connections to get to your final overseas destination. Those things aren't as sexy as "the fastest" or "best" but they are more relevant to what people actually want. And quite frankly testing 20 or more "exit servers" from a dozen or so VPN services just isn't going to pay the bills for most magazines.

Thanks for your replies.

I'm going to do a little more reading on some of the issues I'm having with my "test" computer. I need to set up my VPN service on another system in my house so I can see if I get the same results from that one. My "test" system may have some issues with the PowerLine connection that I'm using. Spent a couple of hours trying to test things only to find out that the brand new ethernet cable I was using was defective.
 

BlueCat57

Distinguished
Apr 7, 2009
430
4
18,815
Thanks for all your answers. Each one provides a bit of worthwhile information. I picked bill001g's answer as the "solution" since it explained a lot that I didn't know about VPNs.

I did find a "solution" to improve my VPN speeds without getting into the obscure settings on my computer. I switched from the PowerLine adapter I was using to connect that computer to a WiFi "ac" dongle. Not only did my regular internet speeds increase to nearly the ISP's "advertised" speeds but for 3 out of 4 "exit servers" the speeds doubled or tripled. Of course, Australia didn't change a bit, but I only did a couple of tests so who knows what that speed will be when final series of The Doctor Blake Mysteries starts "airing" later in the year.

PS - Of course right after I posted this I thought I'd do a speed test because Tom's seemed to be running a bit slow. I blamed the advertising but it turns out at this moment in time I am down to less than a third of "advertised" speed. Not sure what's up with that, but I'll just have to keep testing to see if it stays down. Oh well, that's computers for you. Gremlins. It's got to be Gremlins.

PPS - Well good news. It isn't just the one computer. I ran tests on all my systems and devices and they are all slow right now. Not sure why. Is it still true that everyone on a "switch" shares the bandwidth? It is around 5:45 pm right now so everyone is getting home and connecting to their home networks. If sharing is true, then maybe that's it. Or of course, my ISP could just be throttling me, or maybe the NSA is eavesdropping again.
 

laternet

Prominent
Jul 27, 2017
2
0
510


I don't know if you realize it, but when someone gives a really short reply that says, "But now I use _______, and it's much better," everyone just assumes that you're a shill.
 

laternet

Prominent
Jul 27, 2017
2
0
510


Whoa. A moderator who actually moderates! Did NOT expect that. I'm impressed!
 

TRENDING THREADS