Spy nodes know what websites you are visiting?

Hi everyone, I just got introduced to the mechanics of the SAFE Network, just read http://blanshey.com and talked to @Melvin and @frabrunelle.

Quick question:

Say someone wants to access an illlegal website, will his closest nodes know that he is accessing that website? In other words, can you put spy nodes in the network that spy on their neighbour nodes and see what websites they are visiting?

potentially you could have a node than can detect that an address accessed a data element if you are connected and close (think if you were the tor node) but apart from that no, the user is not connected to the data retrieval mechanism. so it would boil down to knowing an IP:port got a data element and if you can get close (remember you need a lot of nodes) then you could tell an address linked to an IP asked for a data element. Then ISP’s leak info that can be used (for now)

I see this as short term as we can introduce further work on scatter gather to rid this last part, but this attack should not be underestimated. it is a country level attack or even more difficult (beyond 5 eyes) and gets harder as the network grows (like in nature as an ant colony grows, it gets stronger).

4 Likes

@dirvine If you don’t care about who it is that you are spying on, then presumably you could just wait until you were ranked high enough to be a manager of some kind, and then just report whatever traffic flowed through you. I think that is the concern that @Dirk83 is raising.

1 Like

Thanks, I am still new to the concepts, so I will do some more reading, what I am basically want to say is you are not trying to get close, because you are already as close as you can get to your closest neighbour nodes of which you know the ip address.

A scenario where you put huge amounts of spy nodes in the network to spy on everyone? And report/shut off/fine/jail anyone who accesses an illegal data (element)?

1 Like

I think another question is one the nature of public shares. How will they be different from normal encrypted traffic.

Yes there are ways around it if it ever becomes a concern. There is a secret sharing scheme that can create group addresses for specific addresses (crypto hard to create) that will allow us to combat such an attack, needless to say though this is potentially a non attack and harder over time. So we will see. I don’t see it as do-able right now though in terms of monitoring (knowing the sheer size of info). i.e. look at bitcoin, no encryption of network traffic at all and all transactions in the open, using tor becomes an issue as well in that case. We are significantly less vulnerable.

6 Likes

Say I want to access a data element which I don’t want my neighbour node to know and link it to my IP address.

Can’t I just encrypt this destination and send it to a node that can decrypt it? Then he retrieves the information I really need and sends it back to me?

You would basically need some system that you can target a node which can decrypt your package.

Yes this plus a lot of other ways, it gets tracked back via ISP calls though, group keys are a good answer as you encrypt to the group. In that way RSA or similar has to break (Lattice soon which is quantum secure at least as far as shor’s algorithm is concerned).

2 Likes

Good questions. Take a look here:

and here:

As far as I know, you’re closest nodes don’t know you ip-address [incorrect - see note below - @happybeing]. Think of it like some sort of proxy. So you connect to 4 ip-addresses. Ask for your personal file (datamap, data-atlas etc.) decrypt that file with your password, and then ask the 4 ip-addresses you’re connected to, to connect you to the 4 nodes based on XOR. That connection is encrypted. So the ip-addresses you’re connected to don’t see which chunks you ask from your closest nodes. They only see encrpyted data coming by. Your 4 closest nodes don’t know you ip-address. They only know your XOR-addres. So they know about address “FGKJGEKJGFEKJGFEJKGFGEJFGJEF” asking for chunk “JHWQJFHQWJFHQJFHQJHFQKJHFKJQHKJ”. And like @dirvine said, even that can be fixed.

1 Like

You mean, if I ask for a package to my closest nodes it can look like the package was requested by another user?

Edit: I think they do know your IP address; that’s how ISP’s route information from one computer to the next?

Yup, these are all ip-packets. So you go from 22.33.44.55.66 tot 77.66.44.33.11. That’s correct. On the Safenet, another layer of routing is used. So ip-address 22.33.44.55.66 might have a XOR-address like: “GyuHbkjhkJoipjKJHKJjpiJHGFUTfu” So on safe, a packet is sent not to ip-address 22.33.44.55.66 but to “GyuHbkjhkJoipjKJHKJjpiJHGFUTfu”. And nobody in the network knows that 22.33.44.55.66 has XOR-address “GyuHbkjhkJoipjKJHKJjpiJHGFUTfu” So he gets the packet, via some routing in the network. If you want to connect to node “hghYGhghkjgkHGYGYKGkgkghkhJK” you sent the message to the node that’s most close to him in your routing table. So it goes over different nodes( no ip-adress is known!), and finally you get to the node itself, because at the end of course you get to the ip-address that node uses. But the ip-address that got him the message doesn’t know that he’s having that XOR-address. He only sees a packet coming by and sent it to the right ip-address.

2 Likes

I think I made a mistake. Close nodes do know your IP-port and public key… Still confused. Pfff…

The nodes they list have IP:PORT and Public keys. The node will encrypt a message to one of these nodes requesting login (or connect). The bootstrap node gathers this info and returns it to the joining node (encrypted).

Hey @dirk83, I think I found the answer to your question

3 Likes