Slashdot Mirror


Facebook Messenger To Get End-To-End Encryption

Reader wiredmikey writes: Facebook announced Friday it would roll out optional "end to end encryption" for its Messenger application, following a trend aimed at stronger security and protection against snooping. The new feature will be known as "secret conversations" which can be read only by the sender and recipient. Facebook shared technical details about its implementation of the security in a technical white paper (PDF). Facebook earlier this year began implementing this end-to-end encryption on its WhatsApp messaging service.ZDNet's Zack Whittaker, however, warns about a catch in Facebook's effort. He writes: But already the company has faced some criticism for not encrypting messages by default, instead making the service opt-in, like Apple's iMessage, or even Facebook's other chat app, WhatsApp, which recently switched on default end-to-end encryption earlier this year. Cryptographer and Johns Hopkins professor Matthew Green, who reviewed an early version of the system, said in a tweet that though you "have to turn on encryption per thread," he added that providing encryption to almost a billion people makes it hard to "put that genie back in the bottle."

20 of 99 comments (clear)

  1. To my knowledge messenger unlike whatsapp by MarkH · · Score: 3, Insightful

    Keeps a copy on fb servers. So this change is cosmetic

    1. Re:To my knowledge messenger unlike whatsapp by cmseagle · · Score: 2

      Diffie-Hellman key exchange. Exchanging keys over a public channel is a solvable problem. Presumably Facebook will follow basic crypto protocol if they're at all serious about end-to-end encryption.

  2. Re:Why opt-in? by bytestorm · · Score: 2

    It breaks fewer people's shit at once if there's a bug they didn't catch. It's like beta testing a new feature with a small group before deploying it to everyone. It's prudent.

  3. This would imply by JustAnotherOldGuy · · Score: 3, Funny

    This would imply that there is information of value being exchanged on Facebook; a proposition I find difficult to believe.

    --
    Just cruising through this digital world at 33 1/3 rpm...
    1. Re:This would imply by Opportunist · · Score: 3, Insightful

      If only important messages were encrypted, every bad person would instantly know which ones he should decrypt.

      --
      We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
    2. Re:This would imply by gsslay · · Score: 4, Insightful

      You are absolutely right. What you put on Facebook is of no value whatsoever. You have nothing to regret giving it to us. We just like collecting meaningless chatter and none of our client advertisers have the slightest interest in it. Nothing to worry your little heads over, nothing to see here.

      - Mark Zuckerberg

    3. Re:This would imply by JustAnotherOldGuy · · Score: 2

      Actually, this move makes sense considering Facebook is currently trying to get people to use Messenger to interact with other parts of their life including _banking_:

      Lol, if I had a facebook page, the last thing I'd ever do is let it "interact" with my bank account in any way, shape, or form.

      --
      Just cruising through this digital world at 33 1/3 rpm...
  4. Zack mistyped iMessage & WhatsApp auto-encrypt by sasparillascott · · Score: 2

    Just to point out, Zack Whittacker who wrote the ZDNet article mis-typed, as iMessage and WhatsApp are encrypted by default. His following sentence appears to show he actually meant they were automatically encrypted. The opt-in encryption that Facebook and Google are providing will also be the preferred option of the govts / 3 letter agencies that want to keep everything for future use. Its crazy to have Facebook's app on your smartphone anyways...and tracking bracelet with a microphone and camera.

  5. Re:Why use we keep secrets? by Opportunist · · Score: 2

    ...said the AC.

    --
    We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
  6. Re:Why use we keep secrets? by OzPeter · · Score: 2

    Why do we need encryption to keep secrets? The Bible teaches us that evil is done in the shadows and in darkness, but bringing it can't operate in the open when light is shined upon it.

    - Pastor Mitch

    Ok .. whats your full name, DOB, address, SSN and bank account details?

    Shine some light on them and you can be sure nothing bad will happen.

    --
    I am Slashdot. Are you Slashdot as well?
  7. Re:Translation: by cryptizard · · Score: 3, Informative

    End-to-end specifically means that Facebook can't read it, if it is implemented as they say. The ends in question are both users.

  8. Re:Biggest technical flaw: MITM checks are manual? by Solandri · · Score: 2

    (This is unlike SSL/TLS/HTTPS where clients usually automatically verify the ID of the server, and servers often automatically verify the ID of the client.)

    SSL et al don't verify the ID of the server/client. They ask a certificate authority (CA) to verify those IDs. That's why those protocols are vulnerable to MitM attack due to a bad or compromised certificate authority (which for example is exactly what Lenovo did - inserting their own CA into the list of trusted CAs).

    You can think of CAs as a repository of public keys in the public/private key system. They link a claimed identity to a public key. You want to connect to the citibank.com website, you ask a CA what their website's public key is, the CA tells you, and use it to encrypt your traffic to the bank. But if you happen to accidentally type cifibank.com as the URL, and you have a compromised CA in your trusted list who gives you the public key for that bad website, that website can see everything you think you're sending to the real bank, and forward everything you're sending to the real citibank.com website and forward the response back to you to keep up the charade, while they're eavesdropping in on everything.

    If you have some other way to validate the public key of the individual you are messaging with, doing so without a CA is actually more secure. It just can't be done automagically - the end user has to do the work of validating it him or herself. (Although the whole thing remains dodgy as long as we're storing our private keys on the communication device itself, which could be hacked over the network and the private key copied. Ideally the private key would be stored in a separate self-contained processor. The device would send ciphertext to this processor, which would do the decrypting and return plaintext. Vice versa for encryption.)

    tl;dr - They're both vulnerable, just in different ways. CAs are more convenient for random encounters where you have no other secure means to communicate with a person, directly verified keys are more secure when you also have another secure channel with the person you are communicating with - like a one-time face to face meeting.

  9. Re:Translation: by cryptizard · · Score: 2

    No they haven't, read the description of their implementation.

  10. Re:Why use we keep secrets? by The-Ixian · · Score: 3, Insightful

    Do you close the door to the bathroom stall when you take a dump?

    Do you have passwords on any of your accounts?

    Do you make your SS or CC numbers known to the world?

    Privacy is a protection.

    --
    My eyes reflect the stars and a smile lights up my face.
  11. Re:Why should I trust it? by The-Ixian · · Score: 2

    Even if you rolled your own e2e encryption, you still have to trust hundreds or even thousands of strangers who built the hardware or are somewhere in the distribution chain.

    In addition, you would also need to get the other side of the conversation to use your encryption scheme which implies, among other things, sending them a key.

    It is really impossible to "trust no one"

    --
    My eyes reflect the stars and a smile lights up my face.
  12. Re:Breaks reading messages on phone and desktop by cryptizard · · Score: 3, Informative

    That's true, but you do need some "anchor" device for this to work or else there is nothing to bind together the many browser you may have across many devices. Without of course just giving Facebook the key like you said. In practice, most people have the phone's on and connected to cellular internet most of the time. I have used WhatsApp a lot and it really isn't an issue.

  13. Re:Why opt-in? by friedmud · · Score: 2

    My guess: advertising.

    Facebook probably mines the unencrypted messages to help form an "advertising profile" for you so they can better target ads at you when you're on Facebook.

  14. Re:Why should I trust it? by friedmud · · Score: 2

    You could definitely hand-build a small computer (think Raspberry PI) that is offline that you input the encrypted stream into (either via a camera that looks at your monitor or audio from your speakers or other means) that has a small printout on it that shows the decrypted conversation and allows you to answer back...

    There would definitely still be thousands of people involved in making the chips you select... but it would be pretty incredibly difficult to get a backdoor into that system!

  15. Re:Translation: by Fnord666 · · Score: 3, Informative

    No they haven't, read the description of their implementation.

    No thanks, I would rather read their actual implementation (ie open source). The only way you can even begin to trust such a communications system is if it is open source and you can build the client from the provided source. Insert oblig reference to Ken Thompson's "Reflections on Trusting Trust" here. At any rate, the description of the implementation is not the implementation itself.

    --
    'The tyrant will always find pretext for his tyranny.' - Aesop's Fables
  16. Re:Why should I trust it? by Opportunist · · Score: 2

    You don't even remotely need to trust the hardware you use as much as you'd have to trust Facebook in this scenario. They have FULL control over your message. You are using their channel AND their encryption. You can at no point verify that they do not decrypt the message, you can at no point verify that they do not alter the message and you will only be able to discover after the fact whether they actually delivered your message (if your partner does not reply or replies in a way that is inconsistent with the message he should have received). You can NEITHER assure confidentiality, nor integrity, nor availability.

    EVERYTHING else in the communication is less dramatic as a single point of failure in the trust chain. At every other point you can successfully audit and verify. Because you have various different parts that work together, and one of them failing the trust will show. Of course it is possible that EVERYONE, every single bit of hardware that you use, is out there to "get you". In that case, though, whether your communication is encrypted is the least of your concerns.

    --
    We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.