Does Certificate Transparency Break the Web? Measuring Adoption and - - PowerPoint PPT Presentation

does certificate transparency break the web
SMART_READER_LITE
LIVE PREVIEW

Does Certificate Transparency Break the Web? Measuring Adoption and - - PowerPoint PPT Presentation

Does Certificate Transparency Break the Web? Measuring Adoption and Error Rate Emily Stark , Ryan Sleevi, Rijad Muminovic, Devon OBrien, Eran Messeri, Adrienne Porter Felt, Brendan McMillion, Parisa Tabriz estark@chromium.org How


slide-1
SLIDE 1

Does Certificate Transparency Break the Web?

Measuring Adoption and Error Rate

Emily Stark, Ryan Sleevi, Rijad Muminovic, Devon O’Brien, Eran Messeri, Adrienne Porter Felt, Brendan McMillion, Parisa Tabriz estark@chromium.org

slide-2
SLIDE 2
slide-3
SLIDE 3
slide-4
SLIDE 4

How successfully has CT been deployed?

Adoption and compliance User impact Outcomes of various design and deployment decisions

slide-5
SLIDE 5

Outline

  • Background and data sources
  • Analyzing CT compliance
  • Deployment challenges
slide-6
SLIDE 6

Outline

  • Background and data sources
  • Analyzing CT compliance
  • Deployment challenges
slide-7
SLIDE 7

Web server Root certificate authority

cert

CT log: a public, auditable, append-only ledger

signed certificate timestamp

slide-8
SLIDE 8
slide-9
SLIDE 9

Data sources

  • Telemetry from Chrome
  • Active scans of popular websites
  • Qualitative analysis of Chrome help forum posts

(from various points in 2015-2018)

slide-10
SLIDE 10

Outline

  • Background and data sources
  • Analyzing CT compliance
  • Deployment challenges
slide-11
SLIDE 11

CT was supported on

71%

  • f HTTPS requests in Chrome

(February 2018)

slide-12
SLIDE 12

CT compliance

When Chrome requires a site to support CT, how often does the site comply?

slide-13
SLIDE 13

CT compliance

When Chrome requires a site to support CT, how often does the site comply?

99.7%

  • f CT-required HTTPS requests were compliant

(September 2018)

slide-14
SLIDE 14

Outline

  • Background and data sources
  • Analyzing CT compliance
  • Deployment challenges
slide-15
SLIDE 15

Outline

  • Background and data sources
  • Analyzing CT compliance

○ Low compliance would be bad ○ Compliance shouldn’t be taken for granted ○ Contributing factors to high compliance

  • Deployment challenges
slide-16
SLIDE 16

Outline

  • Background and data sources
  • Analyzing CT compliance

○ Low compliance would be bad ○ Compliance shouldn’t be taken for granted ○ Contributing factors to high compliance

  • Deployment challenges
slide-17
SLIDE 17

Users proceeded ~2x more often than certificate errors overall (September 2018)

slide-18
SLIDE 18

60% of help forum threads have an incorrect solution or explanation e.g., “I have tried resetting to default settings (so disabling all extensions).”

slide-19
SLIDE 19

Outline

  • Background and data sources
  • Analyzing CT compliance

○ Low compliance would be bad ○ Compliance shouldn’t be taken for granted ○ Contributing factors to high compliance

  • Deployment challenges
slide-20
SLIDE 20

Outline

  • Background and data sources
  • Analyzing CT compliance

○ Low compliance would be bad ○ Compliance shouldn’t be taken for granted ○ Contributing factors to high compliance

  • Deployment challenges
slide-21
SLIDE 21

Malformed SCT designed to hide domain name from CT logs

slide-22
SLIDE 22

Top 10 websites causing CT errors

(July/September 2018) Name stripping Buggy CA implementation CA lacking CT support Chrome 67 8 2 Chrome 68 10

slide-23
SLIDE 23

Outline

  • Background and data sources
  • Analyzing CT compliance

○ Low compliance would be bad ○ Compliance shouldn’t be taken for granted ○ Contributing factors to high compliance

  • Deployment challenges
slide-24
SLIDE 24
slide-25
SLIDE 25

EV UI requires CT

<= 4% of connections with EV certificates lost EV UI due to CT

slide-26
SLIDE 26

Issuing organization EV certificates w/o SCTs Total EV certificates % w/o SCTs Verizon Cybertrust Security 8550 8556 99.9% Symantec Corporation 1923 495528 3.9% SwissSign AG 1719 1908 90.1% Certplus 1391 1391 100.0% Cybertrust Japan Co., Ltd 1373 24748 5.5%

slide-27
SLIDE 27
slide-28
SLIDE 28

Outline

  • Background and data sources
  • Analyzing CT compliance

○ Low compliance would be bad ○ Compliance shouldn’t be taken for granted ○ Contributing factors to high compliance

  • Deployment challenges
slide-29
SLIDE 29

In 19% of help forum threads, users circumvented error by switching browsers e.g., “I had to download another browser, which im starting to like.”

slide-30
SLIDE 30

Concluding tidbits

How has CT adoption/compliance changed over time? Why have popular websites adopted CT? What is the client-side performance cost of CT? Open problems

slide-31
SLIDE 31

Does Certificate Transparency Break the Web?

Measuring Adoption and Error Rate

Emily Stark, Ryan Sleevi, Rijad Muminovic, Devon O’Brien, Eran Messeri, Adrienne Porter Felt, Brendan McMillion, Parisa Tabriz estark@chromium.org