DNS Privacy dnsprivacy.org Sara Dickinson Sinodun (Salesforce, NLnet Foundation) sara@sinodun.com AFNIC JCSA Paris, France (July 2017)
Overview • The problem: Why Internet privacy and DNS Privacy are important (DNS leakage) • Recent Progress: Chart progress during last 3-4 years (DPRIVE) • Where are we now? Present current status and tools July 2017, Paris, France DNS Privacy @ AFNIC JCSA 2
Internet Privacy Slides from: Daniel Kahn Gillmor (ACLU) July 2017, Paris, France DNS Privacy @ AFNIC JCSA 3
Why does internet privacy matter? • Surveillance as social control • Machine learning at scale today means small number of people controlling network can perform mass surveillance July 2017, Paris, France DNS Privacy @ AFNIC JCSA 4
Behaviour changes (even when no-one is watching) July 2017, Paris, France DNS Privacy @ AFNIC JCSA 5
DNS is part of the leaky boat problem July 2017, Paris, France DNS Privacy @ AFNIC JCSA 6
DNS Privacy - A brief history July 2017, Paris, France DNS Privacy @ AFNIC JCSA 7
IETF Privacy activity March 2011 I-D: Privacy Considerations for Internet Protocols (IAB) Snowdon What timing! June 2013 revelations RFC6973: Privacy Considerations for Internet Protocols July 2013 RFC7258 : Pervasive Monitoring is an Attack: “ PM is an attack on the privacy of Internet users May 2014 and organisations .” July 2017, Paris, France DNS Privacy @ AFNIC JCSA 8
RFC 7258 “ PM is an attack on the privacy of Internet users and organisations .” “…that needs to be mitigated where possible, via the design of protocols that make PM significantly more expensive or infeasible . “ July 2017, Paris, France DNS Privacy @ AFNIC JCSA 9
DNS Privacy in 2013? • DNS is 30 year old! [RFC1034/5 (1987)] • Original design availability, redundancy and speed! • DNS is an ‘enabler’ • DNS standards: DNS sent in clear text -> NSA: ‘ MORECOWBELL ’ • UDP (99% of traffic to root) • TCP only for ‘fallback’ (pre 2010) • Perception: The DNS is public, right? It is not sensitive/personal information….it doesn’t need to be protected/encrypted 10 July 2017, Paris, France DNS Privacy @ AFNIC JCSA
DNS Disclosure Example 1 datatracker.ietf.org Root Rec datatracker.ietf.org Auth datatracker.ietf.org for .org Auth for ietf.org datatracker.ietf.org July 2017, Paris, France DNS Privacy @ AFNIC JCSA 11
DNS Disclosure Example 1 datatracker.ietf.org datatracker.ietf.org Leak information Root Rec datatracker.ietf.org Auth datatracker.ietf.org datatracker.ietf.org for .org Auth for ietf.org datatracker.ietf.org July 2017, Paris, France DNS Privacy @ AFNIC JCSA 11
EDNS0 problem • RFC6891 : Extension Mechanisms for DNS (EDNS0) Intended to enhance DNS protocol capabilities • But…. mechanism enabled addition of end-user data into DNS queries (non-standard options) 12 July 2017, Paris, France DNS Privacy @ AFNIC JCSA
EDNS0 problem • RFC6891 : Extension Mechanisms for DNS (EDNS0) Intended to enhance DNS protocol capabilities • But…. mechanism enabled addition of end-user data into DNS queries (non-standard options) ISP justification: Parental Filtering (per user) CDN justification: Faster content (geo location) 12 July 2017, Paris, France DNS Privacy @ AFNIC JCSA
DNS Disclosure Example 2 Parental Filtering ietf.org ? [00:00:53:00:53:00] Auth Rec Stub CPE [User src address] MAC address or id in DNS query July 2017, Paris, France DNS Privacy @ AFNIC JCSA 13
DNS Disclosure Example 2 Parental Filtering CDN Geo-location ietf.org ? ? ietf.org ? [00:00:53:00:53:00] [192.168.1] Auth Rec Stub CPE [User src address] Client Subnet (RFC7871) MAC address or id contains source subnet in DNS query in DNS query July 2017, Paris, France DNS Privacy @ AFNIC JCSA 13
DNS Disclosure Example 2 Auth Rec Stub CPE Even behind a NAT, Even behind a recursive do do not have not have anonymity! anonymity! DNS Privacy Tutorial @ IETF 97 Nov 2016, Seoul 14
DNS Disclosure Example 2 afnic.fr ? parisinfo.com ? dnsreactions.tumblr.com? Auth Rec Stub CPE Even behind a NAT, Even behind a recursive do do not have not have anonymity! anonymity! DNS Privacy Tutorial @ IETF 97 Nov 2016, Seoul 14
DNS Disclosure Example 2 afnic.fr ? afnic.fr ? parisinfo.com ? parisinfo.com ? dnsreactions.tumblr.com? dnsreactions.tumblr.com? Auth Rec Stub CPE Even behind a NAT, Even behind a recursive do do not have not have anonymity! anonymity! DNS Privacy Tutorial @ IETF 97 Nov 2016, Seoul 14
DNS: It’s not just for names • MX records (email domain) • SRV records (services) • OPENPGPKEY (email addresses) • …this is only going to increase…. July 2017, Paris, France DNS Privacy @ AFNIC JCSA 15
DNS: It’s not just for names • MX records (email domain) • SRV records (services) • OPENPGPKEY (email addresses) • …this is only going to increase…. July 2017, Paris, France DNS Privacy @ AFNIC JCSA 15
DNS Disclosure Example 3 • (AUTH) Who monitors or has access here ISP/ government/NSA/Passive DNS? • (AUTH) Does my ISP sell my (anonymous) data? • (UNAUTH) How safe is this data? Root Rec Auth for .org • When at home… • When in a coffee shop… July 2017, Paris, France DNS Privacy @ AFNIC JCSA 16
DNS Disclosure Example 3 • (AUTH) Who monitors or has access here ISP/ Who monitors or has government/NSA/Passive DNS? access here? • (AUTH) Does my ISP sell my (anonymous) data? • (UNAUTH) How safe is this data? Root Rec Auth for .org • When at home… • When in a coffee shop… Who monitors or has access here? July 2017, Paris, France DNS Privacy @ AFNIC JCSA 16
DNS - leakage • Basic problem is leakage of meta data • Allows fingerprinting and re-identification of individuals • Even without user meta data traffic analysis is possible based just on timings and cache snooping • Operators see (and log) your DNS queries DNS Privacy Tutorial @ IETF 97 17 Nov 2016, Seoul
DNS - leakage • Basic problem is leakage of meta data • Allows fingerprinting and re-identification of individuals • Even without user meta data traffic analysis is possible based just on timings and cache snooping • Operators see (and log) your DNS queries DNS Privacy Tutorial @ IETF 97 17 Nov 2016, Seoul
DNS Risk Matrix In-Flight At Rest Risk Stub => Rec Rec => Auth At At Recursive Authoritative Passive Monitoring Active Monitoring Other Disclosure Risks e.g. Data breaches July 2017, Paris, France DNS Privacy @ AFNIC JCSA 18
DPRIVE WG et al. July 2017, Paris, France DNS Privacy @ AFNIC JCSA 19
DPRIVE WG • DPRIVE WG create in 2014 Charter: Primary Focus is Stub to recursive Why not tackle whole problem? • • Don’t boil the ocean, stepwise solution • Stub to Rec reveals most information • Rec to Auth is a particularly hard problem July 2017, Paris, France DNS Privacy @ AFNIC JCSA 20
DNS Privacy problem Relationship: Root 1 to ‘a few’ some of whom are know (ISP) Relationship: 1 to many most of whom are not known Rec => Authentication is hard Auth for .org July 2017, Paris, France DNS Privacy @ AFNIC JCSA 21
Problem statement: RFC 7626 DNS Privacy Considerations: Expert coverage of risks throughout DNS ecosystem • Rebuts “alleged public nature of DNS data” • The data may be public, but a DNS ‘ transaction ’ is not/should not be. “A typical example from outside the DNS world is: the web site of Alcoholics Anonymous is public; the fact that you visit it should not be.” July 2017, Paris, France DNS Privacy @ AFNIC JCSA 22
Stub/Rec Encryption Options Pros Cons • Port 53 • Downgrade attack on negotiation • Known technique • Port 53 - middleboxes blocking? STARTTLS • Incrementation deployment • Latency from negotiation • New DNS port TLS • New port assignment (no interference with port 53) • Scalability? (new port) • Existing implementations • Truncation of DNS messages • UDP based DTLS (just like UDP) • Not as widely used/ ➡ Fallback to TLS or clear text (new port) deployed ❌ Can’t be standalone solution July 2017, Paris, France DNS Privacy @ AFNIC JCSA 23
Stub/Rec Encryption Options Pros Cons • Port 53 • Downgrade attack on negotiation • Known technique • Port 53 - middleboxes blocking? STARTTLS • Incrementation deployment • Latency from negotiation • New DNS port TLS • New port assignment (no interference with port 53) • Scalability? (new port) • Existing implementations • Truncation of DNS messages • UDP based DTLS (just like UDP) • Not as widely used/ ➡ Fallback to TLS or clear text (new port) deployed ❌ Can’t be standalone solution July 2017, Paris, France DNS Privacy @ AFNIC JCSA 23
Recommend
More recommend
