back to article Meet Riffle, the next-gen anonymity network that hopes to trounce Tor

Next week, top eggheads will unveil a new anonymizing internet tool that they claim is snoop-proof and faster and more reliable against attack than Tor. Dubbed Riffle, the system was developed by MIT and the École Polytechnique Fédérale de Lausanne in Switzerland. It uses the same onion-encryption system as Tor, which wraps …

  1. Anonymous Coward
    Anonymous Coward

    blockchain ..

    sounds like a combination of blockchain and onion routing ... packets (transactions) are thrown into the network, and the node(s) that get to process are randomised by the nature of the contention protocol. So you can't tell which machines were involved in the propagation.

  2. Oengus

    Munitions

    I hope they are developing this outside the clutches of the US Government so that the US Government can't classify it as a munition and ban its export without a license.

    1. Anonymous Coward
      Anonymous Coward

      Re: Munitions

      That "banning" worked really well for the US since Netscape Navigator, didn't it ;-P

    2. Dave 126 Silver badge

      Re: Munitions

      >I hope they are developing this outside the clutches of the US Government

      Second paragraph, first line: :)

      Dubbed Riffle, the system was developed by MIT and the École Polytechnique Fédérale de Lausanne in Switzerland.

      >so that the US Government can't classify it as a munition and ban its export without a license.

      That's been circumvented before, by publishing code in a book, a la Phil Zimmerman's PGP:

      The claimed principle was simple: export of munitions—guns, bombs, planes, and software—was (and remains) restricted; but the export of books is protected by the First Amendment. The question was never tested in court with respect to PGP. In cases addressing other encryption software, however, two federal appeals courts have established the rule that cryptographic software source code is speech protected by the First Amendment

      - https://en.wikipedia.org/wiki/Pretty_Good_Privacy

      1. Khaptain Silver badge

        Re: Munitions

        >Second paragraph, first line: :)

        >Dubbed Riffle, the system was developed by MIT and the École Polytechnique Fédérale de Lausanne in "Switzerland."

        The problem here is the "MIT" part of the equation. The EPFL are obviously in Switzerland but Swiss laws don't have any holding the USA.

        MIT are not a Swiss based university and I am pretty much sure that the Feds are capable of requesting whatever they want on the grounds of "Anti-Terrorism" laws, or one of those George Bush "Weapons of mass destruction" declarations..

        1. Dave 126 Silver badge

          Re: Munitions

          >Swiss laws don't have any holding the USA.

          They don't need to.

          As long as Riffle only requires software (which in hardcopy is protected as free speech) and not specialist hardware (a 'munition'), there is nothing stopping the code escaping the US. Since part of the dev team is in Switzerland, one can make a fair assumption that the code is already out of the US.

        2. Anonymous Coward
          Anonymous Coward

          Re: Munitions

          Well if the code is Switzerland, good luck to the FBI getting it blocked. The Swiss courst are not known for bending over to government requests, even more so US ones.

    3. phuzz Silver badge

      Re: Munitions

      You know a lot of the early work on Tor was funded by the US government?

      Check out their Sponsors page.

  3. Anonymous Coward
    Anonymous Coward

    Given That TOR Was Cracked Years Ago.....

    This is a good thing.

    https://en.wikipedia.org/wiki/Tor_(anonymity_network)

    1. Anonymous Coward
      Anonymous Coward

      Re: Given That TOR Was Cracked Years Ago.....

      Practically all the attacks listed on the wikipedia page have been patched out....

      1. Anonymous Coward
        Anonymous Coward

        Re: Given That TOR Was Cracked Years Ago.....

        Practically all the attacks listed...

        Well that's OK then :|

      2. Anonymous Coward
        Anonymous Coward

        Re: Given That TOR Was Cracked Years Ago.....

        "Practically all the attacks listed on the wikipedia page have been patched out..."

        Yeah.... you just keep believing in that.

    2. Thatguyfromthatforum

      Re: Given That TOR Was Cracked Years Ago.....

      Except the weaknesses have been patched so efficiently that the TLAs are more interested in browser exploits as opposed to trying to beat tor itself.

      1. Anonymous Coward
        Anonymous Coward

        Re: Given That TOR Was Cracked Years Ago.....

        that's why the Browser gets patched

        1. Anonymous Coward
          Anonymous Coward

          Re: Given That TOR Was Cracked Years Ago.....

          Browsers are like Flash: constant patching, yet endless holes remain.

          The WWW is not an acceptable basis for private/anonymous communication. We need simple protocols, designed for a decentralized network from the start.

  4. yossarianuk

    Useful for ms May

    Good timing !

  5. Fitz_

    If it's sending messages to all nodes, and an attacker controls some of those nodes in known geographic locations with precise clocks, could the attacker analyse traffic to determine geographic location of the client based on packet timing and latency?

    1. Thatguyfromthatforum

      Yes and no I imagine. Similar attacks (like tagging attacks) won't work (allegedly) due to the integrity checks, but also if it works like other mix nets, it won't send individual traffic to nodes, it'll send packages of multiple packets, they may also use fake delays like i2p and tor.

  6. Matt Siddall

    Not sure i see how this works

    In TOR, user anonymity is partly provided by the fact that the user only directly communicates with the first node in the chain, and each of the subsequent nodes has no idea where the data is coming from. With this, it sounds like you broadcast your initial message to all nodes, so each of them know where the message came from?

    I'm sure there's more to it than is being said, but I'm not sure how this makes sense.

    1. diodesign (Written by Reg staff) Silver badge

      Re: Not sure i see how this works

      You send out an initial message to start the process of mathematically proving that future messages aren't tampered with. You don't route all messages through all nodes.

      Basically, you can't hide the fact you're using Tor or Riffle - a snooper can see you interacting with an entry node. However, the snooper isn't supposed to see where you're connecting to. Networks like Tor are supposed to disconnect you from your online activity.

      So, for example, snoopers can see you using Tor to reach some kind of system but they don't know which system you're visiting - well, not without some sophisticated timing attacks.

      With that in mind, you might as well ping all entry nodes so they can collectively ensure future messages from you aren't tampered with.

      C.

      1. Matt Siddall

        Re: Not sure i see how this works

        My understanding is likely not particularly accurate, but as I understood things:

        With TOR

        (i) Only the entry node knows who I am, and if I'm operating a passthrough node even they can't be sure that my address is where the traffic originates.

        (ii) Only the exit node knows my destination, and it has no idea who I am

        (iii) Unless the same entity controls the whole network of nodes that I use, noone (including those nodes) can know both who I am and what my destination is.

        In other words, I don't have to even trust the nodes in TOR - they can't compromise me directly even if they wanted to.

        It sounds like that first point is not at all the case on Riffle - and if the exit node is one of the possible entry nodes, then whoever controls that single node can see who I am and my destination.

        Thus (although I'm probably missing something), I can see Riffle being useful for critical communication, but not anonymity.

  7. Thatguyfromthatforum

    This sounds like a mix of tor and i2p which distributes authentication based on distributed hash tables, it also means that all users of the network must also function as nodes. I'll stick to tor thanks, but sounds promising.

    I'd like to see el reg accessible via tor or at least use https at some point, but that's a side note.

  8. Dick Knuckle

    No mention

    Of whether it will be Open Source.

  9. Aodhhan

    Okay so...

    If you think the US government doesn't already have the code, then you're not much of a thinker.

    The more something bounces around the slower the responses and more likely there will be collisions. Which means, follow the error messages back home. It is using TCP after all.

    I'm sure this will be secure for a while. However, you're mistaken if you think most nation states won't have this cracked within a few years.

  10. Anonymous Coward
    Anonymous Coward

    This

    "Ultimately, if just one of the computers routing a Riffle connection remains uncompromised, that one machine will detect when the mathematical proof has been broken, signaling that someone has tried to tamper with the traffic. At that point, the alarm can be raised to stop people from being identified."

    So if one node is compromised it can signal that all the other nodes are tampering and fake the alarm.. Good work

  11. Brian Miller

    Reference implementation?

    I can't find a reference implementation. Where did they hide the code for this?

  12. Anonymous Coward
    Anonymous Coward

    All fun and games

    Until a nonce is using your connection and the rozzers kick your back door in

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like