Sunday, August 30, 2020

Group Instant Messaging: Why Blaming Developers Is Not Fair But Enhancing The Protocols Would Be Appropriate

After presenting our work at Real World Crypto 2018 [1] and seeing the enormous press coverage, we want to get two things straight: 1. Most described weaknesses are only exploitable by the malicious server or by knowing a large secret number and thereby the protocols are still very secure (what we wrote in the paper but some newspapers did not adopt) and 2. we see ways to enhance the WhatsApp protocol without breaking its features.


We are of course very happy that our research reached so many people and even though IT security and cryptography are often hard to understand for outsiders, Andy Greenberg [2], Patrick Beuth [3] and other journalists [4,5,6,7,8] wrote articles that were understandable on the one hand and very accurate and precise on the other hand. In contrast to this, we also saw some inaccurate articles [9,10] that fanned fear and greatly diverged in their description from what we wrote in our paper. We expected this from the boulevard press in Germany and therefore asked them to stick to the facts when they were contacting us. But none of the worst two articles' [9,10] authors contacted us in advance. Since our aim was never to blame any application or protocol but rather we wanted to encourage the developers to enhance the protocols, it contradicts our aim that WhatsApp and Signal are partially declared attackable by "anyone" "easily" [9,10].

Against this background, we understand Moxie's vexation about certain headlines that were on the Internet in the last days [11]. However, we believe that the ones who understand the weaknesses, comprehend that only the malicious server can detectably make use of them (in WhatsApp) or the secret group ID needs to be obtained from a member (in Signal). As such, we want to make clear that our paper does not primarily focus on the description of weaknesses but presents a new approach for analyzing and evaluating the security of group instant messaging protocols. Further we propose measures to enhance the analyzed protocols. The description of the protocols' weaknesses is only one part of the evaluation of our analysis approach and thereby of the investigation of real world protocols. This is the scientific contribution of our paper. The practical contribution of the analyzed messengers, which is the communication confidentiality for billion users (in most cases), is great and should be noted. Therefore we believe that being Signal, WhatsApp, or Threema by applying encryption to all messages and consequently risking research with negative results is much better than being a messenger that does not encrypt group messages end-to-end at all. We do not want to blame messengers that are far less secure (read Moxie's post [11] if you are interested).

Finally we want note that applying security measures according to the ticket approach (as we call it in the paper [12]) to the invitation links would solve the issues that Facebook's security head mentioned in his reply [13] on our findings. To our knowledge, adding authenticity to group update messages would not affect invitation links: If no invitation link was generated for a group, group members should only accept joining users if they were added by an authentic group update message. As soon as a group invitation link was generated, all joining users would need to be accepted as new group members with the current design. However there are plenty ways how WhatsApp could use invitation links without endowing the server with the power to manage groups without the group admins' permission:
One approach would be generating the invitation links secretly and sharing them without the knowledge of the server. An invitation link could then contain a secret ticket for the group and the ID of the group. As soon as a user, who received the link, wants to join the group, she can request the server with the group ID to obtain all current group members. The secret ticket can now be sent to all existing group members encrypted such that the legitimate join can be verified.

Of course this would require engineering but the capability of WhatsApp, shipping drastic protocol updates, can be assumed since they applied end-to-end encryption in the first place.

[1] https://www.youtube.com/watch?v=i5i38WlHfds
[2] https://www.wired.com/story/whatsapp-security-flaws-encryption-group-chats/
[3] http://www.spiegel.de/netzwelt/apps/whatsapp-gruppenchats-schwachstelle-im-verschluesselungs-protokoll-a-1187338.html
[4] http://www.sueddeutsche.de/digital/it-sicherheit-wie-fremde-sich-in-whatsapp-gruppenchats-einladen-koennen-1.3821656
[5] https://techcrunch.com/2018/01/10/security-researchers-flag-invite-bug-in-whatsapp-group-chats/
[6] http://www.telegraph.co.uk/technology/2018/01/10/whatsapp-bug-raises-questions-group-message-privacy/
[7] http://www.handelsblatt.com/technik/it-internet/verschluesselung-umgangen-forscher-finden-sicherheitsluecke-bei-whatsapp/20836518.html
[8] https://www.heise.de/security/meldung/WhatsApp-und-Signal-Forscher-beschreiben-Schwaechen-verschluesselter-Gruppenchats-3942046.html
[9] https://www.theinquirer.net/inquirer/news/3024215/whatsapp-bug-lets-anyone-easily-infiltrate-private-group-chats
[10] http://www.dailymail.co.uk/sciencetech/article-5257713/WhatsApp-security-flaw-lets-spy-private-chats.html
[11] https://news.ycombinator.com/item?id=16117487
[12] https://eprint.iacr.org/2017/713.pdf
[13] https://twitter.com/alexstamos/status/951169036947107840

Further articles:
- Matthew Green's blog post: https://blog.cryptographyengineering.com/2018/01/10/attack-of-the-week-group-messaging-in-whatsapp-and-signal/
- Schneier on Security: https://www.schneier.com/blog/archives/2018/01/whatsapp_vulner.html
- Bild: http://www.bild.de/digital/smartphone-und-tablet/whatsapp/whatsapp-sicherheitsluecke-in-gruppenchats-54452080.bild.html
- Sun: https://www.thesun.co.uk/tech/5316110/new-whatsapp-bug-how-to-stay-safe/

Related posts


  1. Pentest Tools Website Vulnerability
  2. Hacker Tools For Mac
  3. Pentest Tools Framework
  4. Android Hack Tools Github
  5. Hack Tools
  6. Pentest Tools Subdomain
  7. Pentest Automation Tools
  8. Easy Hack Tools
  9. Hacker Tools Github
  10. Pentest Tools Website
  11. New Hacker Tools
  12. Hack Tools For Windows
  13. Hacking Tools For Windows Free Download
  14. Beginner Hacker Tools
  15. Tools 4 Hack
  16. Hacker Tool Kit
  17. Hack Tools Online
  18. New Hacker Tools
  19. Hacking Tools For Windows Free Download
  20. Pentest Tools Free
  21. Pentest Tools Open Source
  22. Hack App
  23. Hack Tools
  24. How To Hack
  25. Pentest Tools Review
  26. Hacker Tools For Windows
  27. Hacking Tools Windows
  28. Hacker Tools Online
  29. Hacker Tools Apk Download
  30. Easy Hack Tools
  31. Hacker Tools 2020
  32. Hacker Tools Mac
  33. Hack Tools
  34. Hack Tool Apk No Root
  35. Hacker Tools Software
  36. Hackrf Tools
  37. Physical Pentest Tools
  38. Ethical Hacker Tools
  39. Best Pentesting Tools 2018
  40. Pentest Reporting Tools
  41. Pentest Tools Apk
  42. Hack Tools 2019
  43. Pentest Tools Find Subdomains
  44. Hack Tool Apk No Root
  45. Hacking Tools Usb
  46. Top Pentest Tools
  47. Tools Used For Hacking
  48. Hack Rom Tools
  49. Hack And Tools
  50. Hacker Tools Free
  51. Hacker Tools Mac
  52. How To Hack
  53. Hack Tools For Windows
  54. Hack Website Online Tool
  55. Hacking Tools Github
  56. Physical Pentest Tools
  57. Hacker
  58. Hack Rom Tools
  59. Pentest Tools Apk
  60. Pentest Tools For Mac
  61. Pentest Tools Linux
  62. Pentest Tools Find Subdomains
  63. Hacking Tools Software
  64. Hack Tools For Windows
  65. Hacking Tools For Kali Linux
  66. Pentest Tools Windows
  67. Pentest Tools For Windows
  68. Pentest Tools Url Fuzzer
  69. Hacking Tools Pc
  70. Hacker Tools For Ios
  71. Pentest Tools For Ubuntu
  72. Termux Hacking Tools 2019
  73. Pentest Box Tools Download
  74. Hacking Tools Windows 10
  75. Tools 4 Hack
  76. Hacking Tools For Windows
  77. Pentest Tools Online
  78. Pentest Tools Url Fuzzer

No comments: