Tor onion hardening will be tear-inducing for feds

Re: How does making the network layer more anonymous help...

The malware would have to install and run locally on your PC. Obviously there are a number of security issues they could take advantage of which have been patched but anyone running an up to date tor browser should be protected against known malware.

Therefore the government would have to use their super-secret zero day stuff. This is quite risky as a blanket campaign (as opposed to targeting a specific PC) as it wouldn't be long before it was detected and decompiled and the zero-day would be closed down, costing the TLA a lot of money to find a new one and reducing their cyber-espionage stockpile.

Re: How does making the network layer more anonymous help...

It turns out they're actually talking about an enhancement for Tor Browser, not Tor itself. The article seems to lump them together, but towards the end it's clear what they meant.

Re: Good.

> US Department of State Bureau of Democracy, Human Rights, and Labor,

Nice try but everyone knows those don't exist in the US government. The State Dept exists to pretend like we will ever solve any of the world's problems (instead of the Dept. of Defense making more). The Labor department exists to issue some pro labor ruling about once a decade that only gets overturned and gets more Republicans elected in an off POTUS election year.

Re: don't get it

Exactly my point above.

Re: don't get it

If not you're familiar with how Tor works, which I am not completely yet either, it is my understanding that you have some client software that acts like a VPN to make your TCP/IP connections route into the Tor system. It routes your traffic 'round and 'round, hence the onion references, then your traffic exits Tor to reach the site you wish to visit. I should look this up, but I'll continue. The FBI was somehow able to "wire" so many peering and other major Internet access points that they could detect client connections to/from Tor and somehow be able to correlate that to traffic exiting Tor to the target site and the client trying to reach it. Basically bypassing the onion in favor of the input and output points surrounding Tor. This is all from reading about it in the news, so might be off by 1, or more. Anyway, at that point they can issue malware into the client to "tattletale" on the target site and some info on the user, including a real IP address, not an anonymized one.

The work being done looks like a beefing up to that weakness; the entry and exit points around Tor and how they are accessed, and possibly within Tor as well. It's an arms race, to be sure.

Re: don't get it

There are really two ways to break Tor-based anonymity. One is to break the Tor anonymity itself (which seems like it may be possible for nation-state-level players, although it may be expensive and/or they may be reluctant to share the data with other players like law enforcement). The other is to break the anonymity above the Tor level: get the user (or, in this case their browser) to tell you who they are without ever having to break Tor!

It is the latter which is addressed in this article. This is about making the "Tor Browser" (not really anything to do with Tor itself but a browser with Tor access conveniently built in) more secure. Like any other browser, Tor Browser has bugs which could be exploited (and have been) to run code on the user's system. That code can make the system report its real name and IP address to the adversary -- allowing law enforcement to know who has accessed what pages, for example. This article is about helping to make the Tor Browser more secure by making those bugs harder to exploit.

Re: Hopeless...

I looked up the definition of "Turing machine". Security isn't mentioned.

_{Translation: I think this means Alan Turing is responsible for all computer security problems because, as we all know, you can't bolt security onto a working project; it has to be built in from the start.}

Re: Jacob should fork Tor

Unfortunately it is very hard to make useful security tradeoffs. We all know that there is no perfect security and we are used to the idea of a need to tradeoff between security and cost (how valuable is the item you are trying to secure? No point on spending more money than that on securing it).

What we very often forget about is usability. If you increase security by reducing usability (ease-of-use, performance, etc) then you are reducing the number of people who will use that security. So, your choices here will depend on whether you are aiming at committed, hardcore, tinfoil-hat-wearing security geeks or Facebook-loving grandmothers or where in between.

Some things (like the move to https: instead of http:) have so little impact on usability that they are no-brainers. However, the decisions made by the Tor project, including the controversial ones (like whether or not to enable Javascript in the Browser, whether to support UDP, whether to add background traffic, etc) are really hard as they have considerable impact on usability and hence real-world takeup.

@AC may have preferred different choices. And I think that some Tor developers and researchers are moving towards some changes, as the threat environment and usage has changed. But I think the Tor developers have generally made pretty good choices and I certainly acknowledge that these are hard decisions with no right answers.

@AC can go ahead and fork Tor/TBB with different choices, and then try to build up enough usage to get useful levels of anonymity. But I think the better choice is to work within the Tor environment, discuss potential changes, conduct (or sponsor) research and development and operate (or fund) relays.

Ranting on El Reg is not likely to help (yes, I know I am guilty of it as well!).

There's also Whonix, which if I understand TOR_Middlebox correctly, does basically the same thing.

Re: Run it as middlebox!

Thanks, Zakhar. <snip> don't run W$ or Android in your VM, they will "call home" with your license key/gmail creds <snip> An excellent point, and one not everyone would think of.

Your post is still here, and the link is still alive and well.