New in OS X: Get MacRumors Push Notifications on your Mac

Resubscribe Now Close

Cryptography Experts Recommend Apple Replace its iMessage Encryption

IMessage_IconApple has implemented a series of short- and long-term defenses to its iMessage protocol after several issues were discovered by a team of researchers at Johns Hopkins University, according to a report published today (via PatentlyApple).

This attack is different to the one Johns Hopkins researchers discovered in March, which allowed an attacker to decrypt photos and videos sent over iMessage.

The technical paper details how another method known as a "ciphertext attack" allowed them to retrospectively decrypt certain types of payloads and attachments when either the sender or receiver is still online.

The scenario requires that the attacker intercepts messages using stolen TLS certificates or by gaining access to Apple's servers. While the attack takes a high level of technical expertise to be successful, the researchers note that it would be well within the means of state-sponsored actors.
Overall, our determination is that while iMessage’s end-to-end encryption protocol is an improvement over systems that use encryption on network traffic only (e.g., Google Hangouts), messages sent through iMessage may not be secure against sophisticated adversaries.
The team also discovered that Apple doesn't rotate encryption keys at regular intervals, in the way that modern encryption protocols such as OTR and Signal do. This means that the same attack can be used on iMessage historical data, which is often backed up inside iCloud. In theory, law enforcement could issue a court order forcing Apple to provide access to their servers and then use the attack to decrypt the data.

The researchers believe the attack could also be used on other protocols that use the same encryption format, such as Apple's Handoff feature, which transfers data between devices via Bluetooth. OpenPGP encryption (as implemented by GnuPGP) may be vulnerable to similar attacks when used in instant messaging applications, the paper noted.

Apple was notified of the issue as early as November 2015 and patched the iMessage protocol in iOS 9.3 and OS X 10.11.4 as a result. Since that time, the company has been pushing out further mitigations recommended by the researchers through monthly updates to several of its products.

However, the team's long-term recommendation is that Apple should replace the iMessage encryption mechanism with one that eliminates weaknesses in the protocol's core distribution mechanism.

The paper detailing the security issue is called Dancing on the Lip of the Volcano: Chosen Ciphertext Attacks on Apple iMessage, and was published as part of the USENIX Security Symposium, which took place in Austin, Texas. You can read the full paper here.



Top Rated Comments

(View all)

38 months ago

John Hopkins is a renowned medical school in Baltimore. What makes them the experts on cryptography?


It's more than just a medical school.

Jesus ****ing christ on a stick we're less than three comments in and 2/3 of them are dismissing this out of hand because it's not a 100% positive Apple story but a constructive criticism of how they can improve weaknesses in their cryptography.
Rating: 40 Votes
38 months ago

John Hopkins is a renowned medical school in Baltimore. What makes them the experts on cryptography?


They have an Information Security Institute. Professor Matthew Green was part of the research team.

Green is part of the group which developed Zerocoin ('https://en.wikipedia.org/wiki/Zerocoin'), an anonymous cryptocurrency ('https://en.wikipedia.org/wiki/Cryptocurrency'). His research team has exposed flaws in more than one third of SSL/TLS ('https://en.wikipedia.org/wiki/Transport_Layer_Security') encrypted web sites as well as vulnerabilities in encryption technologies, including RSA BSAFE ('https://en.wikipedia.org/wiki/RSA_BSAFE'), Exxon/Mobil Speedpass ('https://en.wikipedia.org/wiki/Speedpass'), E-ZPass ('https://en.wikipedia.org/wiki/E-ZPass'), and automotive security systems. In 2015, Green was a member of the research team that identified the Logjam ('https://en.wikipedia.org/wiki/Logjam_(computer_security)') vulnerability in the TLS protocol.

Green is a member of the technical advisory board for the Linux Foundation Core Infrastructure Initiative, formed to address critical Internet security concerns in the wake of the Heartbleed ('https://en.wikipedia.org/wiki/Heartbleed') security bug disclosed in April 2014 in the OpenSSL ('https://en.wikipedia.org/wiki/OpenSSL') cryptography library.

He sits on the technical advisory boards for CipherCloud ('https://en.wikipedia.org/wiki/CipherCloud'), Overnest and Mozilla Cybersecurity Delphi. Green co-founded and serves on the Board for Directors of the Open Crypto Audit Project (OCAP), which undertook a security audit ('https://en.wikipedia.org/wiki/Security_audit') of the TrueCrypt ('https://en.wikipedia.org/wiki/TrueCrypt') software.

https://en.wikipedia.org/wiki/Matthew_D._Green
Rating: 35 Votes
38 months ago

John Hopkins is a renowned medical school in Baltimore. What makes them the experts on cryptography?


Comments like these annoy me quite a bit (unless I'm missing some type of sarcasm). As an Apple user and someone with a background in cryptography who has actually read the entire paper, you don't need to have a MIT or Stanford paper to make a cryptanalysis. In cryptography papers are heavily peer reviewed and skepticism is part of the process the whole time. At the end of the day it boils down to mathematics and computer science and these are provable things, so it's not hypothesis. The paper includes examples of how the attacks can be carried out and under specific conditions. It explains the protocols and the exact mechanisms used to extract the payloads in their settings. All the caveats are stated. Also, it does state that Apple implemented a lot of their recommendations in later versions of iOS and OS X/macOS (their paper references iOS 9.3 and OS X 10.11.4 or later)
Rating: 31 Votes
38 months ago

I think I read this on news.google.com.au.... sounds like a beat up to me. Next....


You have to read more than just the title before you can make an informed comment.
Rating: 19 Votes
38 months ago

I think I read this on news.google.com.au.... sounds like a beat up to me. Next....


Doesn't sound like a beat up to me. Sounds like good advice and it seems Apple has been favorable at receiving advice in the past. Hopefully, they address the concerns for all our sakes.
Rating: 13 Votes
38 months ago

John Hopkins is a renowned medical school in Baltimore. What makes them the experts on cryptography?


A school cannot be great at more than one field?
Rating: 11 Votes
38 months ago

I'm glad they're fixing the issue its refreshing


Technically, they're just patching and adding workarounds to it. The proper fix is to overhaul the entire encryption protocol to avoid these weaknesses and that means a good chance it may not be work on older macOS and iOS versions, which I suspect is what Apple is concerned about.
Rating: 9 Votes
38 months ago
OK, it's somewhat disconcerting that they were able to retrospectively decrypt certain content. But iMessage certainly isn't a "secure messenger" (like Threema or Signal). Yes, it offers end-to-end encryption, but you're not even able to verify that you're actually using your conversation partner's public key to encrypt your messages (hi there, man in the middle). It surely doesn't come as a surprise that you probably shouldn't use iMessage (or WhatsApp, for that matter) if you care about security.
Rating: 5 Votes
38 months ago

Comments like these annoy me quite a bit (unless I'm missing some type of sarcasm). As an Apple user and someone with a background in cryptography who has actually read the entire paper, you don't need to have a MIT or Stanford paper to make a cryptanalysis. In cryptography papers are heavily peer reviewed and skepticism is part of the process the whole time. At the end of the day it boils down to mathematics and computer science and these are provable things, so it's not hypothesis. The paper includes examples of how the attacks can be carried out and under specific conditions. It explains the protocols and the exact mechanisms used to extract the payloads in their settings. All the caveats are stated. Also, it does state that Apple implemented a lot of their recommendations in later versions of iOS and OS X/macOS (their paper references iOS 9.3 and OS X 10.11.4 or later)

It was a genuine question...and I appreciate the answers of everyone who replied.
[doublepost=1471185176][/doublepost]

It's more than just a medical school.

Jesus ****ing christ on a stick we're less than three comments in and 2/3 of them are dismissing this out of hand because it's not a 100% positive Apple story but a constructive criticism of how they can improve weaknesses in their cryptography.


I was not hostile at all in my post, you on the other hand decided it was better to criticize rather than directly answer my question like others did.
Rating: 5 Votes
38 months ago

I think I read this on news.google.com.au.... sounds like a beat up to me. Next....

What? The school praises Apple, says they've been working with suggestions in order to make shorter term patches but the team recommends replacing the whole encryption method. It's very useful constructive criticism. If I were to guess, Apple is probably testing a more permanent solution but, since the messaging platform is used across millions of devices, it takes some time.
Rating: 4 Votes

[ Read All Comments ]