$30 off During Our Annual Pro Sale. View Details »

User Focused Security at Netflix: Stethoscope

Jesse Kriss
January 14, 2017

User Focused Security at Netflix: Stethoscope

Presented by Andrew White and Jesse Kriss at ShmooCon 2017.
Netflix Tech Blog post: http://techblog.netflix.com/2017/02/introducing-netflix-stethoscope.html

GitHub repo: https://github.com/Netflix/Stethoscope

User Focused Security is an approach we are using to address employee information security at Netflix. If we provide employees with the right information and low-friction tools, we believe they can get their devices into a more secure state without heavy-handed policy enforcement.

Letting people retain control over their devices means that they can maintain flexibility and productivity and address security recommendations as appropriate to their levels of access. This approach will only be successful, though, if we can provide clear and specific action, and make it easy to do the right thing.

Stethoscope is a web-based tool that gives Netflix employees a view into the security state of their devices, with specific recommendations regarding disk encryption, firewalls, and other device settings. The website, in conjunction with email alerts, gives Netflix employees a straightforward way to see what actions they should take to remain safe.

Andrew White and Jesse Kriss are both members of the Information Security team at Netflix, where they work on designing and building software tools that help people make good decisions around corporate security.

Andrew holds a PhD in Computer Science from the University of North Carolina at Chapel Hill and a B.S. in Computer Science and B.A. in Mathematics from the University of Richmond.

Jesse (@jkriss) holds a Master’s in Human-Computer Interaction from Carnegie Mellon University and B.A. in Music from Carleton College. Prior to Netflix, he worked at NASA/JPL, Obama 2012, Figure 53, and IBM Research.

Jesse Kriss

January 14, 2017

More Decks by Jesse Kriss

Other Decks in Technology


  1. User Focused Security at Netflix: Stethoscope SHMOOCON 2017 JAN 14

  2. • PhD from UNC in Fall 2015 • Researched side

    channels in encrypted network traffic • Software engineer at Netflix Andrew White
  3. • Masters in HCI from Carnegie Mellon • User experience

    • Web development • Information visualization • Formerly: IBM Research, Figure 53, Obama 2012, NASA/JPL Jesse Kriss
  4. None
  5. None
  6. ...but no security background.


  8. None
  9. Infosec at Netflix

  10. Keep Netflix employees and information safe Thousands of employees. Even

    more devices. Lots of people with access. Worldwide offices.
  11. BYOD 3,000 users 8,000 devices

  12. All cloud everything Streaming infrastructure is 100% cloud > 100,000

    EC2 instances > 700 internal cloud applications
  13. None
  14. Responsible people thrive on freedom, and are worthy of freedom.”

  15. Bad processes creep in. We try to get rid of

    rules when we can, to reinforce the point.” “
  16. Screenshot by Chris Gansen

  17. Values are embedded in and communicated by systems, tools, and

    procedures, not just people.
  18. Only at Netflix?

  19. 1. Education, not just automatic enforcement

  20. Photo by #WOCinTech Chat

  21. None
  22. Work with your colleagues, not against them. 2.

  23. None
  24. The timing seems right for a renewal of interest in

    synthesizing usability and security.” Mary Ellen Zurko “ , 1996
  25. BY HUMANS FOR HUMANS User Focused Security


  27. • Education • Self service • Personalized • One place

    to go • Actionable • Complete the feedback loop The approach.
  28. • Forced updates • Company-wide emails • Information overload •

    “This probably doesn’t apply to me...” And avoiding...
  29. None
  30. None
  31. None
  32. None
  33. None
  34. None
  35. None
  36. None
  37. • Stickers! How do we get people to see it?

  38. None
  39. • Stickers! • New employee “training” • Targeted email campaigns

    How do we get people to see it?
  40. One place to go What about other security alerts?

  41. None
  42. None
  43. HOW THE THING IS BUILT Technical architecture

  44. • Back-end ◦ Python using Twisted + Klein ◦ Plugin

    architecture • Front-end: React • Nginx ◦ Serves static files ◦ Proxies requests to API server • No persistence layer required Technology stack
  45. • Windows: LANDESK • Mac: JAMF • Linux: OSquery (coming

    soon) • Mobile: Google MDM Device data sources
  46. • Authentication logs (BYOD) ◦ Wireless ◦ VPN • bitFit

    (owned devices) Ownership attribution
  47. Device data retrieval

  48. Security practices • Disk encryption • Firewall • Automatic updates

    • Up-to-date OS/software • Screen lock • Not jailbroken/rooted • Security software stack (e.g., Carbon Black)
  49. Status determination

  50. • Events ◦ Google, Duo auth logs ◦ Import from

    Elasticsearch ◦ Augment with, e.g., geolocation data • Accounts: Google • Alerts/feedback: Elasticsearch/REST Other information
  51. • Logging ◦ Accesses: to Elasticsearch ◦ Errors: to Atlas

    • Auth: OpenID Connect • Batch: to Elasticsearch/REST Utilities
  52. SHARING IS CARING Open-source

  53. • Giving back to the community • Knowledge sharing •

    Collaboration Why open-source?
  54. • Front-end source ◦ React-scripts for simple setup, builds, test,

    etc. ◦ Static resources • Back-end source ◦ Plugins previously mentioned ◦ Tests, example configuration, etc. • Nginx configuration • Docker development configuration What’s included
  55. • Primary device data source • [Ownership attribution] • Authentication

    provider What do you need?
  56. THE BIG PICTURE Aggregated data

  57. • Visualization at manager, organization level • Identifies groups for

    targeted efforts Individuals to organizations
  58. • Nightly batch retrieval allows tracking trends over time •

    Identifies practices which need particular attention Are we making progress?
  59. LESSONS SO FAR What we’ve learned

  60. • Inventory needs to be up-to-date and accurate • Data

    sources can have different representations for identifiers • Don’t always get a unique identifier for a device Data quality
  61. • Different users need/want different levels of context • “Make

    it turn green” works well for many people Context
  62. • Additional notification channels • Continuing user research (interviews, surveys)

    • Measure long-term effectiveness Future work
  63. • Open sourcing very soon • We are hiring! Want

    to help us?
  64. COME SAY HI GET IN TOUCH Thank you! netflix.github.io techblog.netflix.com

    @NetflixOSS Andrew White andreww@netflix.com Jesse Kriss jkriss@netflix.com Brooks Evans brookse@netflix.com