SLIDE 1
Measurement and Analysis
- f Private Key Sharing in
Measurement and Analysis of Private Key Sharing in the HTTPS - - PowerPoint PPT Presentation
Measurement and Analysis of Private Key Sharing in the HTTPS - - PowerPoint PPT Presentation
Measurement and Analysis of Private Key Sharing in the HTTPS Ecosystem Frank Cangialosi, Taejoong Chung, David Choffnes, Dave Levin, Bruce M. Maggs, Alan Mislove, Christo Wilson How do we know with whom we are communicating? How do we know
SLIDE 2
SLIDE 3
How do we know with whom we are communicating?
SLIDE 4
Certificate
SLIDE 5
Public Private
Certificate
SLIDE 6
Certificate Authorities Public Private
Certificate
SLIDE 7
Certificate Authorities Public Private
Certificate
SLIDE 8
Certificate
How do we know with whom we are communicating?
SLIDE 9
Certificate Certificate
How do we know with whom we are communicating?
SLIDE 10
Certificate Certificate
How do we know with whom we are communicating?
SLIDE 11
Certificate Certificate
TLS Handshake
How do we know with whom we are communicating?
SLIDE 12
Certificate Certificate
How do we know with whom we are communicating?
SLIDE 13
Certificate Certificate
How do we know with whom we are communicating?
SLIDE 14
Certificate Certificate
Authentication fundamentally assumes: Only knows
How do we know with whom we are communicating?
SLIDE 15
The PKI in today’s web
SLIDE 16
Rare!
The PKI in today’s web
SLIDE 17
The PKI in today’s web
SLIDE 18
The PKI in today’s web
SLIDE 19
The PKI in today’s web
SLIDE 20
The PKI in today’s web
SLIDE 21
Third-party Hosting Providers
- Content delivery networks
- Web hosting services
- Cloud providers
Varying levels of involvement But all trusted to deliver content
SLIDE 22
The PKI in today’s web
SLIDE 23
Certificate
The PKI in today’s web
SLIDE 24
Certificate
The PKI in today’s web
SLIDE 25
Certificate
The PKI in today’s web
SLIDE 26
Certificate
The PKI in today’s web
SLIDE 27
Certificate
The PKI in today’s web
SLIDE 28
Certificate
The PKI in today’s web
SLIDE 29
Third-party hosting providers know their customers’ private keys
SLIDE 30
Third-party hosting providers know their customers’ private keys Authentication fundamentally assumes: Only knows
SLIDE 31
Example of key sharing
SLIDE 32
What’s wrong with sharing?
- 1. Complicates the trust model,
users don’t know who they’re really trusting
- 2. Potential to create
centralization of trust
- 3. Potential to create
single point of failure (in terms of management)
SLIDE 33
This study
SLIDE 34
This study
How many websites share their private keys?
SLIDE 35
This study
How many keys have 3rd parties obtained? How many websites share their private keys?
SLIDE 36
This study
How has this affected key management? How many keys have 3rd parties obtained? How many websites share their private keys?
SLIDE 37
How do we detect sharing at scale?
Rapid7 weekly port 443 scans 2013-2015
DATA
Domain Domain Domain Domain Certificate IP Addr Domain Domain Domain Domain Certificate IP Addr
SLIDE 38
How do we detect sharing at scale?
Rapid7 weekly port 443 scans 2013-2015
DATA
IPv4 Scan Domain Domain Domain Domain Certificate IP Addr Domain Domain Domain Domain Certificate IP Addr
SLIDE 39
How do we detect sharing at scale?
Rapid7 weekly port 443 scans 2013-2015
DATA
Domain Domain Domain Domain Certificate IP Addr Domain Domain Domain Domain Certificate IP Addr
SLIDE 40
How do we detect sharing at scale?
Rapid7 weekly port 443 scans 2013-2015
DATA
Domain Domain Domain Domain Certificate Domain Domain Domain Domain Certificate IP Addr IP Addr
SLIDE 41
How do we detect sharing at scale?
Rapid7 weekly port 443 scans 2013-2015
DATA
IPv4 Scan Domain Domain Domain Domain Certificate Domain Domain Domain Domain Certificate IP Addr IP Addr
SLIDE 42
How do we detect sharing at scale?
Rapid7 weekly port 443 scans 2013-2015
DATA
Domain Domain Domain Domain Certificate Domain Domain Domain Domain Certificate IP Addr IP Addr
SLIDE 43
How do we detect sharing at scale?
Domain Domain Domain Domain Certificate Domain Domain Domain Domain Certificate IP Addr IP Addr
Rapid7 weekly port 443 scans 2013-2015
DATA
SLIDE 44
How do we detect sharing at scale?
Domain Domain Domain Domain Certificate Domain Domain Domain Domain Certificate IP Addr IP Addr
5.1 million valid leaf certificates Rapid7 weekly port 443 scans 2013-2015
DATA
SLIDE 45
How do we detect sharing at scale?
Rapid7 weekly port 443 scans 2013-2015
DATA
Domain Domain Domain Domain Domain IP Addr IP Addr
5.1 million valid leaf certificates
SLIDE 46
How do we detect sharing at scale?
Domain Domain Domain Domain Domain IP Addr IP Addr
Does the same entity that owns the domain own and operate the server at that IP address?
SLIDE 47
How do we detect sharing at scale?
Domain Domain Domain Domain Domain IP Addr IP Addr
Reverse DNS
Does the same entity that owns the domain own and operate the server at that IP address?
SLIDE 48
Domain equivalence?
google.com google.co.uk google.com youtube.com nestle.com friskies.com whitehouse.gov whitehouse.com
Domain names alone are not enough
SLIDE 49
Incorporating whois
Emails in whois records reflect administrative domain
SLIDE 50
google.com google.co.uk google.de zagat.com golang.org
Incorporating whois
Emails in whois records reflect administrative domain
SLIDE 51
google.com google.co.uk google.de zagat.com golang.org whois Registrant Email: Admin Email: Tech Email: dns-admin@google.com dns-admin@google.com dns-admin@google.com
Incorporating whois
Emails in whois records reflect administrative domain
SLIDE 52
google.com google.co.uk google.de zagat.com golang.org dns-admin@google.com dns-admin@google.com dns-admin@google.com
Incorporating whois
Emails in whois records reflect administrative domain
SLIDE 53
google.com google.co.uk google.de zagat.com golang.org dns-admin@google.com dns-admin@google.com dns-admin@google.com
Incorporating whois
Emails in whois records reflect administrative domain
SLIDE 54
google.com google.co.uk google.de zagat.com golang.org Registrant Email: Admin Email: Tech Email: dns-admin@google.com dns-admin@google.com dns-admin@google.com whois whois dns-admin@google.com dns-admin@google.com dns-admin@google.com
Incorporating whois
Emails in whois records reflect administrative domain
SLIDE 55
google.com google.co.uk google.de zagat.com golang.org
Emails in whois records reflect administrative domain
dns-admin@google.com dns-admin@google.com dns-admin@google.com
Incorporating whois
SLIDE 56
google.com google.co.uk google.de zagat.com golang.org
Emails in whois records reflect administrative domain
whois whois dns-admin@google.com dns-admin@google.com dns-admin@google.com
Incorporating whois
SLIDE 57
google.com google.co.uk google.de zagat.com golang.org
Emails in whois records reflect administrative domain
Registrant Email: Admin Email: Tech Email: dns-admin@google.com dns-admin@google.com dns-admin@google.com whois whois dns-admin@google.com dns-admin@google.com dns-admin@google.com
Incorporating whois
SLIDE 58
google.com google.co.uk google.de zagat.com golang.org dns-admin@google.com dns-admin@google.com dns-admin@google.com
Incorporating whois
Emails in whois records reflect administrative domain
SLIDE 59
google.com google.co.uk google.de zagat.com golang.org
Incorporating whois
Domain Organization
Emails in whois records reflect administrative domain
SLIDE 60
How do we detect sharing at scale?
Domain Domain Domain Domain Domain IP Addr IP Addr
Does the same entity that owns the domain own and operate the server at that IP address?
SLIDE 61
Domain Domain Domain
How do we detect sharing at scale?
Domain Domain Org Domain Org Domain Domain Org IP Addr IP Addr
Does the same entity that owns the domain own and operate the server at that IP address?
SLIDE 62
Domain Domain Domain
How do we detect sharing at scale?
Domain Domain Org Domain Org Domain Domain Org Host Host
Does the same entity that owns the domain own and operate the server at that IP address? Key sharing: domain org ≠ host org
SLIDE 63
Domain Domain Domain
How do we detect sharing at scale?
Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org
Does the same entity that owns the domain own and operate the server at that IP address? Key sharing: domain org ≠ host org
SLIDE 64
Outline
How many keys have providers aggregated? How does sharing impact key management? How prevalent is key sharing?
SLIDE 65
How prevalent is key sharing?
Domain Domain Domain Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org
SLIDE 66
How prevalent is key sharing?
Domain Domain Domain Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org Host Org Domain Org
SLIDE 67
How prevalent is key sharing?
Domain Domain Domain Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org Host Org Domain Org
Self-hosted Key sharing
SLIDE 68
How prevalent is key sharing?
Domain Domain Domain Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org Host Org Domain Org
Self-hosted Key sharing
SLIDE 69
How prevalent is key sharing?
Domain Domain Domain Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org Host Org Domain Org
Self-hosted Key sharing
SLIDE 70
How prevalent is key sharing?
0.2 0.4 0.6 0.8 1 1 10 102 103 104 105 CDF Number of Third-Party Hosting Providers Used Organizations
SLIDE 71
How prevalent is key sharing?
0.2 0.4 0.6 0.8 1 1 10 102 103 104 105 CDF Number of Third-Party Hosting Providers Used Organizations
SLIDE 72
How prevalent is key sharing?
0.2 0.4 0.6 0.8 1 1 10 102 103 104 105 CDF Number of Third-Party Hosting Providers Used Organizations
23.5% Self-hosted
SLIDE 73
How prevalent is key sharing?
0.2 0.4 0.6 0.8 1 1 10 102 103 104 105 CDF Number of Third-Party Hosting Providers Used Organizations
23.5% Self-hosted
76.5% share at least 1 key
SLIDE 74
How prevalent is key sharing?
0.2 0.4 0.6 0.8 1 1 10 102 103 104 105 CDF Number of Third-Party Hosting Providers Used Organizations
23.5% Self-hosted
76.5% share at least 1 key Who?
SLIDE 75
Who shares?
0.2 0.4 0.6 0.8 1 200k 400k 600k 800k 1M Fraction of Domains Hosted
- n Third-party Providers
Alexa Site Rank (bins of 10,000) At least one key shared All keys shared
SLIDE 76
Who shares?
0.2 0.4 0.6 0.8 1 200k 400k 600k 800k 1M Fraction of Domains Hosted
- n Third-party Providers
Alexa Site Rank (bins of 10,000) At least one key shared All keys shared
SLIDE 77
Who shares?
0.2 0.4 0.6 0.8 1 200k 400k 600k 800k 1M Fraction of Domains Hosted
- n Third-party Providers
Alexa Site Rank (bins of 10,000) At least one key shared All keys shared
43.2% (of Top 10k) share at least one
SLIDE 78
Who shares?
0.2 0.4 0.6 0.8 1 200k 400k 600k 800k 1M Fraction of Domains Hosted
- n Third-party Providers
Alexa Site Rank (bins of 10,000) At least one key shared All keys shared
43.2% (of Top 10k) share at least one 22.4% share all
SLIDE 79
Who shares?
Key sharing is common across the Internet
0.2 0.4 0.6 0.8 1 200k 400k 600k 800k 1M Fraction of Domains Hosted
- n Third-party Providers
Alexa Site Rank (bins of 10,000) At least one key shared All keys shared
43.2% (of Top 10k) share at least one 22.4% share all
SLIDE 80
Outline
How many keys have providers aggregated? How does sharing impact key management? How prevalent is key sharing?
SLIDE 81
Outline
How many keys have providers aggregated? How does sharing impact key management? How prevalent is key sharing?
- 76.5% share with ≥ 1 provider
- Common even among most
popular websites
SLIDE 82
Domain Domain Domain Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org Host Org Domain Org
How many keys have providers aggregated?
SLIDE 83
Domain Domain Domain Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org Host Org Domain Org
How many keys have providers aggregated?
SLIDE 84
100 101 102 103 104 105 106 100 101 102 103 104 105 106 Number of Distinct Customers Served Rank-Order Third-Party Hosting Providers
How have keys been aggregated?
SLIDE 85
100 101 102 103 104 105 106 100 101 102 103 104 105 106 Number of Distinct Customers Served Rank-Order Third-Party Hosting Providers
How have keys been aggregated?
SLIDE 86
100 101 102 103 104 105 106 100 101 102 103 104 105 106 Number of Distinct Customers Served Rank-Order Third-Party Hosting Providers
How have keys been aggregated?
secureserver.net unifiedlayer.com amazonaws.com Cloud Flare Inc. Rackspace Hosting akamaitechnologies.com
266,110 151,628 117.229 78,369 54,158 15,440 … … #Organizations Hosting provider 277,891 175,089 122,158 87,077 63,418 22,671 … #Domains
SLIDE 87
100 101 102 103 104 105 106 100 101 102 103 104 105 106 Number of Distinct Customers Served Rank-Order Third-Party Hosting Providers
How have keys been aggregated?
secureserver.net unifiedlayer.com amazonaws.com Cloud Flare Inc. Rackspace Hosting akamaitechnologies.com
266,110 151,628 117.229 78,369 54,158 15,440 … … #Organizations Hosting provider 277,891 175,089 122,158 87,077 63,418 22,671 … #Domains
SLIDE 88
100 101 102 103 104 105 106 100 101 102 103 104 105 106 Number of Distinct Customers Served Rank-Order Third-Party Hosting Providers
How have keys been aggregated?
secureserver.net unifiedlayer.com amazonaws.com Cloud Flare Inc. Rackspace Hosting akamaitechnologies.com
266,110 151,628 117.229 78,369 54,158 15,440 … … #Organizations Hosting provider 277,891 175,089 122,158 87,077 63,418 22,671 … #Domains
Top 1% of providers hold keys for 86% of all organizations
SLIDE 89
Domain Domain Domain Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org Host Org Domain Org
Does key sharing make enticing attack targets?
SLIDE 90
Domain Domain Domain Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org Host Org Domain Org
Does key sharing make enticing attack targets?
SLIDE 91
Domain Domain Domain Domain Domain Org Domain Org Domain Domain Org Host Host Host Org Host Org Host Org Domain Org
Does key sharing make enticing attack targets?
SLIDE 92
Does key sharing make enticing attack targets?
0.2 0.4 0.6 0.8 1 100 101 102 103 104 105 106 Cumulative Fraction of Domains’ Keys Acquired Number of Hosting Providers Compromised Alexa Top 1k Alexa Top 1m All Domains
SLIDE 93
Does key sharing make enticing attack targets?
0.2 0.4 0.6 0.8 1 100 101 102 103 104 105 106 Cumulative Fraction of Domains’ Keys Acquired Number of Hosting Providers Compromised Alexa Top 1k Alexa Top 1m All Domains
SLIDE 94
Does key sharing make enticing attack targets?
0.2 0.4 0.6 0.8 1 100 101 102 103 104 105 106 Cumulative Fraction of Domains’ Keys Acquired Number of Hosting Providers Compromised Alexa Top 1k Alexa Top 1m All Domains 60% of Top 1K, same provider
SLIDE 95
Does key sharing make enticing attack targets?
0.2 0.4 0.6 0.8 1 100 101 102 103 104 105 106 Cumulative Fraction of Domains’ Keys Acquired Number of Hosting Providers Compromised Alexa Top 1k Alexa Top 1m All Domains 60% of Top 1K, same provider
SLIDE 96
Does key sharing make enticing attack targets?
0.2 0.4 0.6 0.8 1 100 101 102 103 104 105 106 Cumulative Fraction of Domains’ Keys Acquired Number of Hosting Providers Compromised Alexa Top 1k Alexa Top 1m All Domains >40% of all sites, 10 providers 60% of Top 1K, same provider
SLIDE 97
Does key sharing make enticing attack targets?
0.2 0.4 0.6 0.8 1 100 101 102 103 104 105 106 Cumulative Fraction of Domains’ Keys Acquired Number of Hosting Providers Compromised Alexa Top 1k Alexa Top 1m All Domains
Popular hosting services are prime targets for attack
>40% of all sites, 10 providers 60% of Top 1K, same provider
SLIDE 98
Outline
How many keys have providers aggregated? How does sharing impact key management? How prevalent is key sharing?
- Top 1% of providers hold
keys for 86% of orgs
- Attractive targets for attack
- 76.5% share with ≥ 1 provider
- Common even among most
popular websites
SLIDE 99
Key Management
Request certificates Renew expiring certificates Revoke and reissue compromised certificates
SLIDE 100
Who manages private keys?
CAs
Website acquires Third-party acquires
SLIDE 101
Who manages private keys?
CAs
Website acquires Third-party acquires
SLIDE 102
Who manages private keys?
CAs
Website acquires Third-party acquires
SLIDE 103
Who manages private keys?
Website acquires Third-party acquires
SLIDE 104
Who manages private keys?
Website acquires Third-party acquires
Diverse
“Self-managed”
SLIDE 105
Who manages private keys?
Website acquires Third-party acquires
Diverse Heavily skewed
“Self-managed” “Outsourced”
SLIDE 106
Who manages private keys?
Website acquires Third-party acquires
Diverse Heavily skewed
“Self-managed” “Outsourced”
58.4% of Alexa Top 10K 33.0% of all domains
SLIDE 107
How does sharing impact key management?
0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced
Natural experiment: Heartbleed (4/7/2014)
SLIDE 108
How does sharing impact key management?
0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced
Natural experiment: Heartbleed (4/7/2014)
SLIDE 109
How does sharing impact key management?
0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced CloudFlare revocations
Natural experiment: Heartbleed (4/7/2014)
SLIDE 110
How does sharing impact key management?
0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced 0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced Outsourced (w/o CF) CloudFlare revocations
Natural experiment: Heartbleed (4/7/2014)
SLIDE 111
How does sharing impact key management?
0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced 0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced Outsourced (w/o CF) CloudFlare revocations Slightly more thorough
Natural experiment: Heartbleed (4/7/2014)
SLIDE 112
How does sharing impact key management?
0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced 0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced Outsourced (w/o CF) CloudFlare revocations 10 days to react! Slightly more thorough
Natural experiment: Heartbleed (4/7/2014)
SLIDE 113
How does sharing impact key management?
0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced 0.75 0.8 0.85 0.9 0.95 1 04/07 04/11 04/15 04/19 04/23 04/27 05/01 05/05 Fraction of Certificates Not Revoked Date Self-managed Outsourced Outsourced (w/o CF) CloudFlare revocations 10 days to react! Slightly more thorough
Natural experiment: Heartbleed (4/7/2014) A few revoked thoroughly, but many did not!
SLIDE 114
How does sharing impact key management?
0.2 0.4 0.6 0.8 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 CDF of Hosting Providers Fraction of Heartbleed-vulnerable Certificates Revoked Self-managed Outsourced
Natural experiment: Heartbleed (4/7/2014)
(One year after Heartbleed)
SLIDE 115
How does sharing impact key management?
0.2 0.4 0.6 0.8 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 CDF of Hosting Providers Fraction of Heartbleed-vulnerable Certificates Revoked Self-managed Outsourced
Natural experiment: Heartbleed (4/7/2014)
(One year after Heartbleed)
SLIDE 116
How does sharing impact key management?
0.2 0.4 0.6 0.8 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 CDF of Hosting Providers Fraction of Heartbleed-vulnerable Certificates Revoked Self-managed Outsourced
66% of providers did not revoke a single vulnerable certificate!
Natural experiment: Heartbleed (4/7/2014)
(One year after Heartbleed)
SLIDE 117
How does sharing impact key management?
0.2 0.4 0.6 0.8 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 CDF of Hosting Providers Fraction of Heartbleed-vulnerable Certificates Revoked Self-managed Outsourced
66% of providers did not revoke a single vulnerable certificate! A small minority of providers revoked most of their vulnerable certificates, but none revoked all
Natural experiment: Heartbleed (4/7/2014)
(One year after Heartbleed)
SLIDE 118
Outline
How many keys have providers aggregated? How does sharing impact key management? How prevalent is key sharing?
- Creates single point of failure
- Most third-parties did poor
job of revoking
- Top 1% of providers hold
keys for 86% of orgs
- Attractive targets for attack
- 76.5% share with ≥ 1 provider
- Common even among most
popular websites
SLIDE 119