Hacker News Reader

Hacker News

Comment by coolspot | Hacker News Reader

Full thread

coolspot

Maybe feds don’t want to upgrade their vast cloud of Tor nodes just yet [1][2]

1 - https://blog.torproject.org/tor-security-advisory-relay-earl...

2 - https://arstechnica.com/information-technology/2014/07/activ...

View on HN

Is there anyone that still believes TOR will protect him from state actors?

It's nice that people still run exit nodes, but it's so expensive and dangerous that you cannot expect a satisfactory number of nodes to be run for moral reasons and not for monetary/surveillance ones.

Also since most of the nodes are run on VM instances, a few companies could potentially recover all the routing and keys from the servers RAM if they feel like it. They're not going to do it for the average joe, but things like the Patriot act and it's European equivalents make sure intelligence agencies can do it if they want.

I'm just throwing an idea out here. It is often stated that running Tor exit nodes is both costly and dangerous. For understandable reasons. Yet the network needs more such nodes, to defeat the scenario of one powerful entity operating hundreds of exit nodes. It would be nice if even ordinary Tor users can run transient exit nodes. Now as it stands, nobody will do it. So why not change the functionality of Tor such that it becomes possible for a user to operate an exit node for a whitelisted list of sites (either based on IP or domain names) that they are willing to funnel traffic for? It's still better than giving up and letting a majority of exit nodes be controlled by state actors and other powerful entities right? Serving traffic only for a set of white-listed endpoints can considerably reduce the risk of running afoul of local law enforcement.

I realise this effectively cripples Tor, but normal "uncensored" exit nodes can still operate, by those willing to run them.

I've seen this idea thrown around in Tor mailing lists as well.

The developers basically replied that while they're not happy about how Tor is used by some people, if they started censoring some things, they would eventually be required to censor others. At that point, what is really the point of Tor - just chain some VPNs together.

The point would be the stronger anonymity offered by Tor, whereas with VPNs, you're basically at the mercy of the service providers not to keep logs. Also the developers needn't censor anything but merely build in a mechanism that individual node operators can use or not as they see fit. Like adblock lists, the lists themselves would presumably be made and circulated by interested individuals, and nothing to do with Tor devs.

Anyway, thinking further, I realised even a whitelist can be bypassed in various ways to do illegal stuff, so you'd still need to be quite prepared and legally informed before running an Exit node. :-(

You need all nodes in the path, not just exit nodes, to identify source. And traffic already HTTPSed, so controlling exit node won't help. So far I've never heard of state actors beating TOR. They usually exploit things like TOR browser vulnerabilities, etc.

TOR was developed to protect american spies, so it's exactly its purpose: to defend against state actors.

no. if you control the entry and exit nodes, you can identify the source using traffic correlation analysis

Yes I believe that TOR will protect against state actors in the very similar way that leaks and investigative journalism protect against corruption and abuse. It increase costs where its too costly for the average case and for the super important and critical cases then the cheaper and more available methods are likely to bring results way before such costly attacks are even considered.

For example, if they want to know who owns a specific tor hidden service they could get the lawyers and higher ups to pressure amazon and microsoft to dump rams of all their customers (and manage all the secrecy so to not create bad PR), or they could just make a search on their vulnerability database and attack the website directly. They can also just put a well trained and experienced detective to find traces from the operators that they might accidentally make on facebook, stackoverflow and so on. They can have an agent go under cover or try recruit people close to the target. Each of those have costs and risk that is far cheaper than attacking the tor network, and having read a lot of stories in regard to the silk road, wikileaks and other users of tor it seems that this is indeed the primary way state actors operates.

Tor project should implement subscription model to pay for running thousands of tor nodes.

Users pay its yearly fee via Bitcoin. Once a year.

Tor client authenticates to entry node with wallet ID. Node checks in blockchain that fee's been paid and forwards traffic.

Plus - clean nodes operated by trusted Tor project team.

Minus - experts need review that wallet checking is not a deal breaker. I have not really thought this through, just throwing ideas :))

i think this is a good idea, but it would require a lot of thought to ensure that the wallet linking requirement does not deanonymize paid users. i assume youre aware that BTC is not actually anonymous without taking some nontrivial steps.

A cryptocurrency implementing a zero-knowledge proof, like ZCash, may be a better choice here.

blog.cryptographyengineering.com/2014/11/27/zero-knowledge-proofs-illustrated-primer

... time machines and Part 2. I Love this guy :)

Thank you :D I'm far from coins, but heard that monero may be more anonymous.

But it does not kills the idea, imho. Buying and paying coins anonymously is a different, independent area.

Came to my mind that this subscription model may spawn several Tor networks, trust whichever you want, like with VPN providers. Just keep the Tor code maintained by the public.

And several networks may use each other's nodes ... Cyberpunk in its best

> I'm far from coins, but heard that monero may be more anonymous

i'm in a similar position. i find crypto interesting but i don't understand it well enough to be confident about the anonymity provided. a concern i have is that people are already looking for ways to deanonymize crypto wallets and tor users separately. to link tor to crypto wallets would seem to increase the attack surface to deanonymize tor users. i'll leave it to those more knowledgeable than me to say how much of a risk this really is.

again, i do think it is very important for the long term viability of tor to create more incentive for users to support it financially.

> very important for the long term viability of tor to create more incentive for users to support it financially

Yeah, that was the idea, nothing grows without money. Trust me, people will pay money for Tor once model is proven to be secure and anonymous.

There is private networking, Tor, but it needs private money to run.

There is private money, coins, but it needs private networking to run.

Lets merge them together, somehow :) Onion Coin need to be established in .onion domain only, dark web (onion) exchanges converting Onions to 'meat' (a.k.a IPV4) internet coins, to pay for Tor nodes.

Got me googling stuff (yeah, shut up), found deeponion.org - cryptocurrency sent through the TOR :)

Problem is, wallet IDs are public, so I can use anyone's who paid the fee ... Etherium contracts?

You could do something like provide a public key in the transaction, and then sign the handshake with the tor node.

Reply on news.ycombinator.com