On Mon, Mar 20, 2017 at 12:00 AM, iry <[email protected]> wrote:
> Jean-Philippe Ouellet:
>> The same problems would still apply, but would instead apply to
>> obtaining and bootstrapping trust in your copy of Qubes. If this is
>> a problem you really wish to solve, simply providing trust via
>> signed qubes templates is not sufficient.
>
> I am not sure why "simply providing trust via signed qubes templates
> is not sufficient". Do you mean if one do no verify the Lantern
> software, he/she will probably not do verify for Qubes OS installation
> file neither? Could you please explain more about this to me? Thank
> you very much!

Yes, this is what I mean. Bootstrapping guarantees of authenticity and
integrity is not solved by simply wrapping the intended payload
(Lantern) in another thing (Qubes) which must itself also be
bootstrapped, unless that thing (Qubes) is already somehow trusted. In
this case, I sincerely doubt the intended users already have a trusted
copy of Qubes.

Rather, the net effect of such wrapping is that an adversary who
wishes to subvert the delivery of Lantern may instead target Qubes.

I'm not saying don't do it, only pointing out (as I think you already
understand) that meaningful secure delivery is not solved simply by
signing a template with a key distributed with Qubes, and as such the
download sources & verification section of your proposal was IMO
perhaps somewhat misleading.

>>> B. Privacy ... By placing Lantern in an independent qube, it will
>>> not be able to collect the local user behaviors that happen in
>>> other qubes.
>>
>> I think claiming this is somewhat misleading unless you also also
>> enforce that the traffic within is somehow also protected
>> independently.
> Thank you for your feedback! I did find it misleading. The following
> is my correction to the previous statement. How do you like this one?
>
> By placing Lantern in an independent qube, it will not be able to
> collect the user behaviors that happen in other qubes.

...except for all the traffic analysis and active man-in-the-middleing
it could do even on networks which might otherwise not be attacker
controlled.

> Additionally, a whonix-gateway can be set up between Lantern-Gateway
> and an AppVM, making all the traffic through the Lantern client
> encrypted by Tor client already.

I think this (or equivalent) should be very much encouraged, otherwise
your above statement may not hold true.

I view your proposed ProxyVM as providing only availability,
potentially at degraded integrity and confidentiality (in the event
that the circumvention tool is somehow vulnerable, which you indeed
recognize may be the case). Someone wanting all three properties may
need to take additional measures besides just using a given
censorship-circumvention tool, and should be well informed that that
is the case.

This is a minor detail though which can be addressed in documentation
much later on. Don't worry about it, your proposal is still very
strong IMO.

>> Running the server-side infrastructure to forward traffic has
>> associated costs, and perhaps those running them somehow rely on
>> income from Lantern Pro in order to cover those costs? I think it
>> would be wise to at least start a discussion with them about this
>> rather than taking an adversarial approach towards Brave New
>> Software right from the start.
>
> Thank you very much for your advice! I agree with you that I should
> start a discussion with the Lantern community. I did not even realize
> this can be a problem. Your advice let me realize that there is a
> difference between being a DIY user and being a responsible developer.

Heh. Reminds me of my game-hacking days ;)

There's a difference between privately using a cheat you wrote, and
distributing it and ruining the game for everyone...

> I will start a discussion with them soon! And I will report the result
> to the mail-list.

Sounds good.

> Please let me say thank you again for your feedback, Jean!

You're very welcome! This kind of detailed and engaged communication
is a model to all GSoC applicants :)

-- 
You received this message because you are subscribed to the Google Groups 
"qubes-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/qubes-devel/CABQWM_AeafOm8JFLPDLrx%2BYSj%2BQtbbXgpmhe43VcCVWu02GcYA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to