VPN makes my upload speed faster?

[barbary]

Hello All,

I have a strange tail to tell, although you may go duh of course. I only work from home, Big Foot has been seen more times than me in the office. So my link between myself and the office and myself and the data center is something I keely monitor.

When I started 2Mbs adsl was the best i could get. These days I have 100Mbs Virgin Media which was just upgraded for free to 120Mbs. Nice but slightly hollow offering as I've atually been getting about 115Mbs since it was installed.

What dawned on me today was I'd never tested what speed I get when using the VPN. If I'm pulling actual real customer data I use a VPN to the office and get the data that way, just to be safe. So I tried it and got about 32Mbs, a fair drop but it makes sense to me, the speed of the VPN box at the office is probably a factor and many other things.

Occasionaly I need to upload data now this used to take all night with adsl but with 10Mbs upload on the Virgin Media its not such a chore. So I uploaded the same data I just got, just to try it, expecting to get 10Mbs but I didn't I got 32Mbs. So I tried it again and got 32Mbs.

So I tried speedtest over the VPN and got 23.5Mbs down and 23.5Mbs up.

I tried uploading some jpgs to a website and got 32Mbs up.

Does this make any sense?

After all my upload cap is only controled by software is there some bug at Virgin Media that a VPN connection is getting around the cap?

Does anyone else have Virgin Media 120Mbs service and a VPN to a fast link and can try the same thing?

[+PeterUK MVC]

Its called VPN compression so your not getting around any cap as your real speed that the cap is limiting you is limiting you to 10Mb just that VPN compression can give you more much like compressing a file but in like real time.

[barbary]

Its called VPN compression so your not getting around any cap as your real speed that the cap is limiting you is limiting you to 10Mb just that VPN compression can give you more much like compressing a file but in like real time.

Jpgs and zips are already compressed. If you were right then we should all be doing this all the time.

[+BudMan MVC]

"Nice but slightly hollow offering as I've atually been getting about 115Mbs since it was installed."

And how do you know that is not an issue with your router being able to actually do 120?

As to your upload question, what do you get without using vpn to speedtest for your upload? Your seeing 115 both up and down? Is there any sort of say virus scanner that might be scanning the files when not using the vpn?

What are you using to measure this upload? And actual ftp client that reports speed? Of each transfer, are you sure some files are not just being skipped and showing better speed. Depending on your measurement methods, I wouldn't always trust what some dialog says for speed.

[barbary]

"Nice but slightly hollow offering as I've atually been getting about 115Mbs since it was installed."

And how do you know that is not an issue with your router being able to actually do 120?

Let me clarify, I've just re read my post and it is slightly less than clear. I was supposed to be on 100 and got 115 now I'm supposed to get 120 and get 120.

As to your upload question, what do you get without using vpn to speedtest for your upload? Your seeing 115 both up and down? Is there any sort of say virus scanner that might be scanning the files when not using the vpn?

What are you using to measure this upload? And actual ftp client that reports speed? Of each transfer, are you sure some files are not just being skipped and showing better speed. Depending on your measurement methods, I wouldn't always trust what some dialog says for speed.

I get 10Mbs up if not using the VPN which is what I should get. That was my point. I don't follow your question about the virus scanner at all.

[+BudMan MVC]

so your connection is 120 down and 10 up?

And your saying when you upload file using vpn you get 32 up? Well thats just impossible to be honest, so I would assume its a issue with measurement. If you connection is only 10mbps up.

You say your seeing faster than even your speed test shows for upload on the vpn of 23mbps?

Again how are you measuring this upload?

[barbary]

so your connection is 120 down and 10 up?

And your saying when you upload file using vpn you get 32 up? Well thats just impossible to be honest, so I would assume its a issue with measurement. If you connection is only 10mbps up.

You say your seeing faster than even your speed test shows for upload on the vpn of 23mbps?

Again how are you measuring this upload?

1) Speedtest as I put in my original post

2) Just doing a plain windows copy and just seeing what speed it sits at

3) ftp client that gives you an average speed

4) A clock.

[ObiWanToby]

I'd say VPN compression. It has to be. That is what I notice when I use a VPN as well. Latency is higher, compression is lower.

[+BudMan MVC]

"Just doing a plain windows copy and just seeing what speed it sits at"

So your doing a SMB copy to a webserver on the internet? Really? Or are you uploading to a file at your work location, ie the vpn endpoint in one test. But in your other test your hitting this server via a public facing connection?

So how many files, what size of files. You sure its not some kind of cache? You can see some nice performance increases with compression sure, I wouldn't think an already compressed format would see much benefit though.

What type of vpn are you using? What is the compression method? LZO? Can we see this speed test?

"just seeing what speed it sits at"

Yeah thats not a very valid form of testing ;)

"3) ftp client that gives you an average speed

4) A clock."

And if some files are not being copied because they already exist it could like your getting 100x faster depending.

As to speed test.. What servers are you hitting? Quite possible that speedtest server is showing you different speeds based upon which one your hitting and from what source.

[barbary]

I'd say VPN compression. It has to be. That is what I notice when I use a VPN as well. Latency is higher, compression is lower.

So lets get this right windows is taking the 600Mb zip file compressing it on the fly to 200Mb and sending that. Some questions..

1) OK so like I said above why doesn't all the connections do this all the time? I don't need the security part just this awsome compression part.

2) Seems to take my computer a fair time to zip the file to 600Mb. Why doesn't it just zip it down to 200Mb in the first place that would be really handy?

3) It also uses a good deal of processor power to zip the file to 600Mb but it seems to take none to get from 600Mb to 200Mb whats up with that?

4) Is there a way of closing or deleting threads?

[barbary]

"Just doing a plain windows copy and just seeing what speed it sits at"

So your doing a SMB copy to a webserver on the internet? Really? Or are you uploading to a file at your work location, ie the vpn endpoint in one test. But in your other test your hitting this server via a public facing connection?

So how many files, what size of files. You sure its not some kind of cache? You can see some nice performance increases with compression sure, I wouldn't think an already compressed format would see much benefit though.

What type of vpn are you using? What is the compression method? LZO? Can we see this speed test?

"just seeing what speed it sits at"

Yeah thats not a very valid form of testing ;)

"3) ftp client that gives you an average speed

4) A clock."

And if some files are not being copied because they already exist it could like your getting 100x faster depending.

As to speed test.. What servers are you hitting? Quite possible that speedtest server is showing you different speeds based upon which one your hitting and from what source.

It was a single zip file and then bunch of jpgs none of which existed on there before. Yes it was to the office file server first then an ftp in the office. Simply because I knew the ftp client would give me an average.

I agree that just seeing the numbers on the screen is the least acurate of the methods and if it was 1Mbs I was talking about then I wouldnt have posted but its 3 times the speed and clocks dont lie.

I wish I had kept the speed test thing I just tried it again and now I'm getting 15 down and 7 up. Either this was some weird glitch in the matrix or its now 3.30pm and everyone at work is using the connection it was 6.30am when I did it before . I'll try it again later and see if I can repeat the weird results. I just tried copying the file and only get about 3.5Mbs now.

It's PPTP if that means anything to you.

[+PeterUK MVC]

How many Upstream Channels do you have listed in the hub?

[+BudMan MVC]

Yeah PPTP tells me a lot, for starters your using OLD tech ;) I would assume a just default windows client connection from that, so by default it would be using compression - you can turn that off in properties. But I doubt it has anything to do with what your seeing, which I would call more a false positive until we have better info.

So this was one large file - this is a better test then lots of smaller ones I agree. But testing to a server on the same network as your vpn endpoint compared to what your seeing in a speedtest to some public server? Not sure I would call that apples to apples.

As to using just normal windows file copy - do you have Remote Differential Compression in play? There are many many things that could come into play with doing a windows copy of some file. As to our apples to apples thing - how could you test this without using the vpn?

As to a ftp test, again your not apples to applies test if your hitting this server via 2 different paths, one is a vpn endpointed inside your network I would assume. Next it ftp interface to the public net.. Which could have all kinds of limitations on it - you can set limits on speeds, for all I know its 2 different interfaces on the public side one that is for vpn one for other connections, ftp interface could be at a different speed than vpn interface, etc. etc..

I really do applaud you for trying to track down such oddness for sure - I would be the same way! Makes no sense I agree, so lets rule out all the obvious things that could be confusing and your going to need to get valid tests showing this if we want to track it down. Seems like your not seeing it now, so when you can reproduce it. Lets get some actual numbers to work with, and something more than just screen shot of windows file copy dialog ;)

[barbary]

Yeah PPTP tells me a lot, for starters your using OLD tech ;) I would assume a just default windows client connection from that, so by default it would be using compression - you can turn that off in properties. But I doubt it has anything to do with what your seeing, which I would call more a false positive until we have better info.

So this was one large file - this is a better test then lots of smaller ones I agree. But testing to a server on the same network as your vpn endpoint compared to what your seeing in a speedtest to some public server? Not sure I would call that apples to apples.

As to using just normal windows file copy - do you have Remote Differential Compression in play? There are many many things that could come into play with doing a windows copy of some file. As to our apples to apples thing - how could you test this without using the vpn?

As to a ftp test, again your not apples to applies test if your hitting this server via 2 different paths, one is a vpn endpointed inside your network I would assume. Next it ftp interface to the public net.. Which could have all kinds of limitations on it - you can set limits on speeds, for all I know its 2 different interfaces on the public side one that is for vpn one for other connections, ftp interface could be at a different speed than vpn interface, etc. etc..

I really do applaud you for trying to track down such oddness for sure - I would be the same way! Makes no sense I agree, so lets rule out all the obvious things that could be confusing and your going to need to get valid tests showing this if we want to track it down. Seems like your not seeing it now, so when you can reproduce it. Lets get some actual numbers to work with, and something more than just screen shot of windows file copy dialog ;)

It makes no difference what route its taking or where the data goes or what settings are on the ftp server. If I was debating getting a slower speed then it would but I'm not I'm just debating the upload speed from the house. Both routes leave the house and go across the internet thats all that matters.

Without VPN I get 10Mbs exactly what I should get and what its caped to.

With VPN I got 32Mbs over 3 times what I should get.

It could be 2 totaly different ftp servers. What setting could I set on the public ftp server that would make it go faster than the cap?