ROS2 Security Working Group Online Meeting - Apr 17th, 2019 between 2PM and 3PM PST

Hi everyone,

We are planning to hold the next ROS 2 Security Working group meeting on Apr 17th, 2019 between 2PM and 3PM PST.

Tentative agenda and meeting details are below.

Suggestions for other topics are very welcome!

Agenda

  1. ROS 2 threat model comments and feedback
  2. Next steps for the security work:
    1. AWS RoboMaker team planned work for Q2
    2. How to collaborate and tackle security risks as a community?
    3. Alias Robotics update about threat modeling for industrial robots.
    4. Alias Robotics update with ROSIN security analysis for ROS 2 based on our tools.
  3. Safety & Security (Acutronic Robotics to discuss some ongoing work in the intersection of both)
  4. Database of robot vulnerabilities (discussion, disclosures and a new security article maybe?)

Meeting Details

You have been invited to an online meeting, powered by Amazon Chime.

Chime meeting ID: 7596499512

Join via Chime clients (manually): Select ā€˜Meetings > Join a Meetingā€™, and enter 7596499512

Join via Chime clients (auto-call): If you invite auto dial-in as attendee, Chime will call you when the meeting starts, select ā€˜Answerā€™

Join via browser screen share: https://chime.aws/7596499512

Join via phone (US): +1-929-432-4463,7596499512#

Join via phone (US toll-free): +1-855-552-4463,7596499512#

International dial-in: https://chime.aws/dialinnumbers/

5 Likes

FYI - @vmayoral @gavanderhoorn @gbiggs @ruffsl @jacob @kyrofa

1 Like

Thanks for organizing Thomas! Weā€™ll be there. How can we su suggest items for the agenda?

Thank you Thomas for leading the organization once again! Eager to participate from Alias Robotics on this. We have some work to share with you guys. Can you please add the following Items to the agenda:

  • 2.3 Alias Robotics update about threat modeling for industrial robots.

  • 2.4 Alias Robotics update with ROSIN security analysis for ROS 2 based on our tools.

1 Like

@Odei - done!
@vmayoral feel free to reply to this thread and Iā€™ll edit my post.

1 Like

Great, thanks a lot. Hereā€™s my wishlist then, editing your agenda above:

@Dejan_Pangercic and @gbiggs, you may be interested on this.

1 Like

As @Odei pointed out, in collaboration with Acutronics Robotics, we have been extending the threat model developed by Amazon Robotics to include an industrial robotic Arm. In this, case, the MARA modular robot arm. We have submitted the PR to ROS 2 design. Feedback is very much appreciated!

2 Likes

Hi @Thomas_Moulard, sorry for the late ping.

If there is some time left today Iā€™d like to ask whether this group here would be interested to perform such analysis on such a car robot/use case in Autoware?

D.

Thank you everyone for your time today!

Todayā€™s meeting recording is available here.

The next meeting is already planned in two weeks as we decided today and the time is hopefully more Asia-friendly. Details are in this thread.

@Dejan_Pangercic - did we spend enough time on your questions today? If not, letā€™s add your questions to the next meeting agenda.

@Thomas_Moulard thanks, I think that we are fine. To recap what you said:

  1. We can use this document as a reference and apply it to our car and use case
  2. We can create a PR like this https://github.com/ros2/design/pull/228 once we have done our own threat analysis
  3. The difference between your threat model and https://www.aliasrobotics.com/research/rsf.htm is in that the latter are guidelines on how to make a secure robot and the threat model checks whether implementation of the guidelines is correct (that is there is not attack surface exposed)
  4. You try hard to focus on ROS 2 framework only (as oppose to e.g. RTOS)

Is that correct?

Otherwise we in Autoware currently do not have anyone that has security background, so we were looking for someone that could help us get started. But I guess we can try on our own first and ask for help here if needed.

Yes! Iā€™d suggest to also to read the section ā€œIncluding a new robot into the threat modelā€ in https://github.com/ros2/design/pull/228

For 4. weā€™d be open to add pointers to URLs about how to secure RTOS but we just donā€™t want the doc to become a guide to secure RTOS as this is out of the scope of the doc. You can see, for instance, what we wrote around NTP attacks. We point to ā€œgood practicesā€ but they are not directly described in the doc.

Apologies I couldnā€™t participate yesterday @Thomas_Moulard and the rest of the group :frowning:. Terribly overloaded these last few weeks trying to fix some internal matters. Iā€™ve noted down in my calendar the next meeting and Iā€™ll put together some slides to kick off the next meeting with our contributions so far while discussing in a bit more detail our disclosure with MARA.