If I had to put the types of compartmentalization in some order:
Multi-account containers
Browser Profiles
Desktop Users
VM
Separate devices
I bolded the first three, because the last two have more obvious use cases. The difference between the first three are muddier.
Plus, there are two wild cards that don’t quite fit: containers and PWA.
So: When to use which type of compartmentalization? What setup do YOU have?
Feel free to answer either way or both ways, specify which! Share how your setup differs from your ideal setup, the practical advantages and annoyances of your method, etc.
Additional questions:
What are the technical differences between those methods?
What are the differences in usage?
How might mileage differ between different threat models/use cases?
When you log into invasive services you don’t want to use anymore for the tedious and sometimes long-term task of cleanup, getting affairs in order, and exporting your data, which would you use? A disposable VM? A disposable desktop user? A disposable browser profile? A disposable container?
Sorry to give an expected answer but it really does depend on your threat model. If I just need a different Reddit account for the sake of separating my post history from onlookers but don’t care if Reddit could connect those accounts, I’ll probably just use different browser profiles or set my browser data to wipe on exit and simply log into whichever one I want on the same browser. If security or anonymity was a more serious concern, I’d use VMs and/or separate devices depending on my budget.
I see! Thank you for the answer. Where do you think different desktop users/profiles fall, here?
I’m kind of new to desktops. If you’re okay with answering, what’s the technical breakdown on how these compartmentalization methods differ that I might not know yet? (Since I can’t tell how it fits into my threat model if I don’t I have an idea of that.) For example, I could infer from your response that browser profiles don’t really prevent services from linking your accounts… maybe because they don’t hide hardware info? I don’t really know the inner workings like that, so I can’t tell how it plays into my threat model yet.
One issue with using the same browser is sharing the same browser fingerprint. Some browsers counter fingerprinting better than others. Different browser profiles can have different configurations and therefore different fingerprints, but you might still share a lot of similarly unique characteristics. Depending on how anonymous you need to be, you might want to stick exclusively with the Tor Browser.
But of course you can take it further. Tails and Whonix provide much better anonymity than simply just using the Tor Browser on your host OS, you can read about their benefits on their respective websites. Whonix can be a bit advanced, so if you’re new to this I’d stick with trying Tails unless you really need the unique benefits of Whonix such as security by compartmentalization (since it’s contained within a virtual machine).
Thank you for those answers! I have some cursory knowledge of browser fingerprinting. I was also wondering jf that kind of fingerprinting can also extend to separate users/logins on the desktop? Many people use this to have a user for work and a user for fun. I think they don’t share home directories, but I don’t know.
I assume there’s something fingerprintable there, too, related to OS and hardware, but desktop users are so little discussed that I’m not sure what the opinion is on them. I would love to know where you think they lie/their use case is in relation to all the others.