Maybe I’m completely wrong about everything I’m going to say and in that case we can laugh about this theory I guess but here it goes…
Most people are only worried about if the VPN provider is keeping logs or not. But even if they don’t keep logs you could still be tracked by anyone who can see incoming and outgoing connections to the VPN server.
This would be easier to explain if I drew some images but I hope you understand anyway with just text. What it looks like for these adversaries is:
- they know your IP and who you are.
- They see you connect to a VPN server.
- They see VPN server connecting to many different servers and they don’t know which one is you.
But when it comes to number 3, they could actually figure out which one is you.
Obviously, if you are the only person connected to the VPN server they will see that there is no one else besides you using it and then any outgoing connection from the VPN server must be you.
If there are just a few users. Maybe three users are just connected to the VPN server but not doing anything, just idle. Another user is spending time reading reddit. Then you connect to the vpn server and within a minute a new outgoing connection from the vpn server starts and goes to lemmy. Pretty good guess that is you from their perspective. And to make the guess even better, when the connection to lemmy ends, you decide to immediately end your connection to the VPN server. I’m confident this would be enough evidence in a court and then it’s definitely enough for data harvesting and mass surveillance.
All this analysis can be done automatically with AI, even if there are hundreds users on a VPN server, the AI will over a larger amount of time (not just hours but days/weeks/months) collect enough data to be able to profile users and make good guesses which domains you are visiting even if the VPN prpvider doesn’t have logs.
What is the solution to avoid this type of tracking? Tor baby, tor. Leeegggoooo Whonix!
You must log in or # to comment.
I feel like ISP-level timing attacks are a similar issue with tor, like we know the Germans can deanonymize you if they’re willing to expend the effort. Not saying tor isn’t better, all VPNs regardless of protocol can be forced to send in the clear if an upstream actor breaks the traffic correctly afaik as of now.
Some VPNs allow multi-hopping, similar to Tor. I couldn’t give you an exhaustive list but most popular ones support this. Mullvad and Proton do, for example. There are also strategies to add noise into VPN traffic.
This is not a silver bullet, of course. Tor has similar problems as you describe if an adversary has visibility into enough nodes. As always, this comes down to your threat model.
On the one hand, I find the advertising of VPNs outright dishonest. On the other hand, I would trust any reputable VPN provider much more than I trust my ISP or cell carrier.
I find the advertising of VPNs outright dishonest.
Can you explain? I’m curious
They make it sound like it’s going to protect you from identity theft
Most particularly they generally pretend that nothing on the web is encrypted whereas in practice HTTPS is nearly universal at this point.
Sure. I’m referring to the ones that run big ad campaigns, like Nord and Mullvad. They tend to overstate how a VPN can protect you, sometimes in ways that barely make sense. There is no epidemic of criminals stealing personal credit card information over insecure wi-fi, for example. The ads play into ignorance and fear.
That said, yeah, I’d rather be on a VPN when on a public wi-fi network. But I’m not really worried about someone sniffing my encrypted HTTPS traffic (which is pretty much everything nowadays; Firefox by default won’t even load unencrypted web sites).
The main defense against VPN timing attacks is to ensure your VPN exit node isn’t somewhere that the same person would have access to as your connecting IP.
That said, if someone runs a website or service where you have a unique login or custom token and they have access to your ISP’s connection logs… a standard VPN will once again give you away. This is why TOR exists.
I generally argue that an exit VPN doesn’t really provide much privacy; the only real services it provides are georelocation and protection against low effort bulk filtering (eg, identifying torrenters or bulk metadata collection).
For everything else, either encryption and third party DNS is enough, or the exit VPN isn’t enough to stop targeted surveillance.
I hope DAITA gets implemented in all VPN services.
Also, if you have a limited RAM smartphone and your VPN is operating in userspace, then all it takes is for one really large image to grace your smartphone screen for your OS to go into out-of-memory kill mode. What’s it going to kill? The foreground app you’re trying to use, or the background VPN app.
In my experience, the VPN goes down before the browser does. Mounting a swap on your phone is not the worse solution against this, but the UI starts to get really unresponsive.
In settings on my android, non root.
Always on VPN. Block unless active
If you are who I think you are, we’ve probably had this discussion before. Even with an always on VPN, if the system runs out of memory it will kill the VPN first before the browser. In a perfect world the traffic would still be routed into a dead tunnel. From what I’ve seen, once the VPN is killed, the tunnel device is gone and the default route snaps back to wlan
I am not. And I’ve never had this discussion.
Always on vs the additional option of blocking internet until the VPN connects.
The second option is more system level?
Using shizuku (rish) in termux I checked the active links with VPN on and then force stopped / killed the VPN in terminal and checked again. The VPN tunnel disappeared but the dummy kill switch tunnels remained. I could not access any network connection.
*The routing table also maintains the dummy kill switch
(oh sorry, but) I’ve heard this argument before. All I can say is that in my experience, when the system is out of memory, it kills some process (e.g. the UI) which upon restarting resets the networking
You’ll be happy to know I just force killed :
Android system
Google services framework
Network
System UI
System WiFi Resources
Wi-Fi
Settings
System connectivity resources
Secure UI serviceThe results are the same
VPN kill switch prevents network access.
Yet another argument for root. Then you can exclude the VPN app from OOM. Or even move it into /system.
I understand why this isn’t done (moving such apps to system), since mobile uses immutable OS concept. But we still need a way to manage such apps appropriately.
Then you can exclude the VPN app from OOM.
I didn’t know about this feature. what’s the config?
OOM_DISABLE on $PID or echo -17 > /proc/$PID/oom_adj
