Internet Engineering Task Force (IETF)                  T. Wicinski, Ed.
Request for Comments: 9076                                     July 2021
Obsoletes: 7626
Category: Informational
ISSN: 2070-1721


                      DNS Privacy Considerations

Abstract

  This document describes the privacy issues associated with the use of
  the DNS by Internet users.  It provides general observations about
  typical current privacy practices.  It is intended to be an analysis
  of the present situation and does not prescribe solutions.  This
  document obsoletes RFC 7626.

Status of This Memo

  This document is not an Internet Standards Track specification; it is
  published for informational purposes.

  This document is a product of the Internet Engineering Task Force
  (IETF).  It represents the consensus of the IETF community.  It has
  received public review and has been approved for publication by the
  Internet Engineering Steering Group (IESG).  Not all documents
  approved by the IESG are candidates for any level of Internet
  Standard; see Section 2 of RFC 7841.

  Information about the current status of this document, any errata,
  and how to provide feedback on it may be obtained at
  https://www.rfc-editor.org/info/rfc9076.

Copyright Notice

  Copyright (c) 2021 IETF Trust and the persons identified as the
  document authors.  All rights reserved.

  This document is subject to BCP 78 and the IETF Trust's Legal
  Provisions Relating to IETF Documents
  (https://trustee.ietf.org/license-info) in effect on the date of
  publication of this document.  Please review these documents
  carefully, as they describe your rights and restrictions with respect
  to this document.  Code Components extracted from this document must
  include Simplified BSD License text as described in Section 4.e of
  the Trust Legal Provisions and are provided without warranty as
  described in the Simplified BSD License.

Table of Contents

  1.  Introduction
  2.  Scope
  3.  Risks
  4.  Risks in the DNS Data
    4.1.  The Public Nature of DNS Data
    4.2.  Data in the DNS Request
      4.2.1.  Data in the DNS Payload
    4.3.  Cache Snooping
  5.  Risks on the Wire
    5.1.  Unencrypted Transports
    5.2.  Encrypted Transports
  6.  Risks in the Servers
    6.1.  In the Recursive Resolvers
      6.1.1.  Resolver Selection
      6.1.2.  Active Attacks on Resolver Configuration
      6.1.3.  Blocking of DNS Resolution Services
      6.1.4.  Encrypted Transports and Recursive Resolvers
    6.2.  In the Authoritative Name Servers
  7.  Other Risks
    7.1.  Re-identification and Other Inferences
    7.2.  More Information
  8.  Actual "Attacks"
  9.  Legalities
  10. Security Considerations
  11. IANA Considerations
  12. References
    12.1.  Normative References
    12.2.  Informative References
  Appendix A.  Updates since RFC 7626
  Acknowledgments
  Contributions
  Author's Address

1.  Introduction

  This document is an analysis of the DNS privacy issues, in the spirit
  of Section 8 of [RFC6973].

  The Domain Name System (DNS) is specified in [RFC1034], [RFC1035],
  and many later RFCs, which have never been consolidated.  It is one
  of the most important infrastructure components of the Internet and
  is often ignored or misunderstood by Internet users (and even by many
  professionals).  Almost every activity on the Internet starts with a
  DNS query (and often several).  Its use has many privacy
  implications, and this document is an attempt at a comprehensive and
  accurate list.

  Let us begin with a simplified reminder of how the DNS works (see
  also [RFC8499]).  A client, the stub resolver, issues a DNS query to
  a server called the recursive resolver (also called caching resolver,
  full resolver, or recursive name server).  Let's use the query "What
  are the AAAA records for www.example.com?" as an example.  AAAA is
  the QTYPE (Query Type), and www.example.com is the QNAME (Query
  Name).  (The description that follows assumes a cold cache, for
  instance, because the server just started.)  The recursive resolver
  will first query the root name servers.  In most cases, the root name
  servers will send a referral.  In this example, the referral will be
  to the .com name servers.  The resolver repeats the query to one of
  the .com name servers.  The .com name servers, in turn, will refer to
  the example.com name servers.  The example.com name servers will then
  return the answers.  The root name servers, the name servers of .com,
  and the name servers of example.com are called authoritative name
  servers.  It is important, when analyzing the privacy issues, to
  remember that the question asked to all these name servers is always
  the original question, not a derived question.  The question sent to
  the root name servers is "What are the AAAA records for
  www.example.com?", not "What are the name servers of .com?".  By
  repeating the full question, instead of just the relevant part of the
  question to the next in line, the DNS provides more information than
  necessary to the name server.  In this simplified description,
  recursive resolvers do not implement QNAME minimization as described
  in [RFC7816], which will only send the relevant part of the question
  to the upstream name server.

  DNS relies heavily on caching, so the algorithm described above is
  actually a bit more complicated, and not all questions are sent to
  the authoritative name servers.  If the stub resolver asks the
  recursive resolver a few seconds later, "What are the SRV records of
  _xmpp-server._tcp.example.com?", the recursive resolver will remember
  that it knows the name servers of example.com and will just query
  them, bypassing the root and .com.  Because there is typically no
  caching in the stub resolver, the recursive resolver, unlike the
  authoritative servers, sees all the DNS traffic.  (Applications, like
  web browsers, may have some form of caching that does not follow DNS
  rules, for instance, because it may ignore the TTL.  So, the
  recursive resolver does not see all the name resolution activity.)

  It should be noted that DNS recursive resolvers sometimes forward
  requests to other recursive resolvers, typically bigger machines,
  with a larger and more shared cache (and the query hierarchy can be
  even deeper, with more than two levels of recursive resolvers).  From
  the point of view of privacy, these forwarders are like resolvers
  except that they do not see all of the requests being made (due to
  caching in the first resolver).

  At the time of writing, almost all this DNS traffic is currently sent
  unencrypted.  However, there is increasing deployment of DNS over TLS
  (DoT) [RFC7858] and DNS over HTTPS (DoH) [RFC8484], particularly in
  mobile devices, browsers, and by providers of anycast recursive DNS
  resolution services.  There are a few cases where there is some
  alternative channel encryption, for instance, in an IPsec VPN tunnel,
  at least between the stub resolver and the resolver.  Some recent
  analysis on the service quality of encrypted DNS traffic can be found
  in [dns-over-encryption].

  Today, almost all DNS queries are sent over UDP [thomas-ditl-tcp].
  This has practical consequences when considering encryption of the
  traffic as a possible privacy technique.  Some encryption solutions
  are only designed for TCP, not UDP, although new solutions are still
  emerging [RFC9000] [DPRIVE-DNSOQUIC].

  Another important point to keep in mind when analyzing the privacy
  issues of DNS is the fact that DNS requests received by a server are
  triggered for different reasons.  Let's assume an eavesdropper wants
  to know which web page is viewed by a user.  For a typical web page,
  there are three sorts of DNS requests being issued:

  Primary request:
     This is the domain name in the URL that the user typed, selected
     from a bookmark, or chose by clicking on a hyperlink.  Presumably,
     this is what is of interest for the eavesdropper.

  Secondary requests:
     These are the additional requests performed by the user agent
     (here, the web browser) without any direct involvement or
     knowledge of the user.  For the Web, they are triggered by
     embedded content, Cascading Style Sheets (CSS), JavaScript code,
     embedded images, etc.  In some cases, there can be dozens of
     domain names in different contexts on a single web page.

  Tertiary requests:
     These are the additional requests performed by the DNS service
     itself.  For instance, if the answer to a query is a referral to a
     set of name servers and the glue records are not returned, the
     resolver will have to send additional requests to turn the name
     servers' names into IP addresses.  Similarly, even if glue records
     are returned, a careful recursive server will send tertiary
     requests to verify the IP addresses of those records.

  It can also be noted that, in the case of a typical web browser, more
  DNS requests than strictly necessary are sent, for instance, to
  prefetch resources that the user may query later or when
  autocompleting the URL in the address bar.  Both are a significant
  privacy concern since they may leak information even about non-
  explicit actions.  For instance, just reading a local HTML page, even
  without selecting the hyperlinks, may trigger DNS requests.

  Privacy-related terminology is from [RFC6973].  This document
  obsoletes [RFC7626].

2.  Scope

  This document focuses mostly on the study of privacy risks for the
  end user (the one performing DNS requests).  The risks of pervasive
  surveillance [RFC7258] are considered as well as risks coming from a
  more focused surveillance.  In this document, the term "end user" is
  used as defined in [RFC8890].

  This document does not attempt a comparison of specific privacy
  protections provided by individual networks or organizations; it
  makes only general observations about typical current practices.

  Privacy risks for the holder of a zone (the risk that someone gets
  the data) are discussed in [RFC5155] and [RFC5936].

  Privacy risks for recursive operators (including access providers and
  operators in enterprise networks) such as leakage of private
  namespaces or blocklists are out of scope for this document.

  Non-privacy risks (e.g., security-related considerations such as
  cache poisoning) are also out of scope.

  The privacy risks associated with the use of other protocols that
  make use of DNS information are not considered here.

3.  Risks

  The following four sections outline the privacy considerations
  associated with different aspects of the DNS for the end user.  When
  reading these sections, it needs to be kept in mind that many of the
  considerations (for example, recursive resolver and transport
  protocol) can be specific to the network context that a device is
  using at a given point in time.  A user may have many devices, and
  each device might utilize many different networks (e.g., home, work,
  public, or cellular) over a period of time or even concurrently.  An
  exhaustive analysis of the privacy considerations for an individual
  user would need to take into account the set of devices used and the
  multiple dynamic contexts of each device.  This document does not
  attempt such a complex analysis; instead, it presents an overview of
  the various considerations that could form the basis of such an
  analysis.

4.  Risks in the DNS Data

4.1.  The Public Nature of DNS Data

  It has been stated that "the data in the DNS is public".  This
  sentence makes sense for an Internet-wide lookup system, and there
  are multiple facets to the data and metadata involved that deserve a
  more detailed look.  First, access control lists (ACLs) and private
  namespaces notwithstanding, the DNS operates under the assumption
  that public-facing authoritative name servers will respond to "usual"
  DNS queries for any zone they are authoritative for, without further
  authentication or authorization of the client (resolver).  Due to the
  lack of search capabilities, only a given QNAME will reveal the
  resource records associated with that name (or that name's
  nonexistence).  In other words: one needs to know what to ask for in
  order to receive a response.  There are many ways in which supposedly
  "private" resources currently leak.  A few examples are DNSSEC NSEC
  zone walking [RFC4470], passive DNS services [passive-dns], etc.  The
  zone transfer QTYPE [RFC5936] is often blocked or restricted to
  authenticated/authorized access to enforce this difference (and maybe
  for other reasons).

  Another difference between the DNS data and a particular DNS
  transaction (i.e., a DNS name lookup): DNS data and the results of a
  DNS query are public, within the boundaries described above, and may
  not have any confidentiality requirements.  However, the same is not
  true of a single transaction or a sequence of transactions; those
  transactions are not / should not be public.  A single transaction
  reveals both the originator of the query and the query contents; this
  potentially leaks sensitive information about a specific user.  A
  typical example from outside the DNS world is that the website of
  Alcoholics Anonymous is public but the fact that you visit it should
  not be.  Furthermore, the ability to link queries reveals information
  about individual use patterns.

4.2.  Data in the DNS Request

  The DNS request includes many fields, but two of them seem
  particularly relevant for the privacy issues: the QNAME and the
  source IP address.  "Source IP address" is used in a loose sense of
  "source IP address + maybe source port number", because the port
  number is also in the request and can be used to differentiate
  between several users sharing an IP address (behind a Carrier-Grade
  NAT (CGN), for instance [RFC6269]).

  The QNAME is the full name sent by the user.  It gives information
  about what the user does ("What are the MX records of example.net?"
  means they probably want to send email to someone at example.net,
  which may be a domain used by only a few persons and is therefore
  very revealing about communication relationships).  Some QNAMEs are
  more sensitive than others.  For instance, querying the A record of a
  well-known web statistics domain reveals very little (everybody
  visits websites that use this analytics service), but querying the A
  record of www.verybad.example where verybad.example is the domain of
  an organization that some people find offensive or objectionable may
  create more problems for the user.  Also, sometimes, the QNAME embeds
  the software one uses, which could be a privacy issue (for instance,
  _ldap._tcp.Default-First-Site-Name._sites.gc._msdcs.example.org.
  There are also some BitTorrent clients that query an SRV record for
  _bittorrent-tracker._tcp.domain.example.

  Another important thing about the privacy of the QNAME is future
  usages.  Today, the lack of privacy is an obstacle to putting
  potentially sensitive or personally identifiable data in the DNS.  At
  the moment, your DNS traffic might reveal that you are exchanging
  emails but not with whom.  If your Mail User Agent (MUA) starts
  looking up Pretty Good Privacy (PGP) keys in the DNS [RFC7929], then
  privacy becomes a lot more important.  And email is just an example;
  there would be other really interesting uses for a more privacy-
  friendly DNS.

  For the communication between the stub resolver and the recursive
  resolver, the source IP address is the address of the user's machine.
  Therefore, all the issues and warnings about collection of IP
  addresses apply here.  For the communication between the recursive
  resolver and the authoritative name servers, the source IP address
  has a different meaning; it does not have the same status as the
  source address in an HTTP connection.  It is typically the IP address
  of the recursive resolver that, in a way, "hides" the real user.
  However, hiding does not always work.  The edns-client-subnet (ECS)
  EDNS0 option [RFC7871] is sometimes used (see one privacy analysis in
  [denis-edns-client-subnet]).  Sometimes the end user has a personal
  recursive resolver on their machine.  In both cases, the IP address
  originating queries to the authoritative server is as sensitive as it
  is for HTTP [sidn-entrada].

  A note about IP addresses: there is currently no IETF document that
  describes in detail all the privacy issues around IP addressing in
  general, although [RFC7721] does discuss privacy considerations for
  IPv6 address generation mechanisms.  In the meantime, the discussion
  here is intended to include both IPv4 and IPv6 source addresses.  For
  a number of reasons, their assignment and utilization characteristics
  are different, which may have implications for details of information
  leakage associated with the collection of source addresses.  (For
  example, a specific IPv6 source address seen on the public Internet
  is less likely than an IPv4 address to originate behind an address-
  sharing scheme.)  However, for both IPv4 and IPv6 addresses, it is
  important to note that source addresses are propagated with queries
  via the ECS option and comprise metadata about the host, user, or
  application that originated them.

4.2.1.  Data in the DNS Payload

  At the time of writing, there are no standardized client identifiers
  contained in the DNS payload itself (ECS, as described in [RFC7871],
  is widely used; however, [RFC7871] is only an Informational RFC).

  DNS Cookies [RFC7873] are a lightweight DNS transaction security
  mechanism that provides limited protection against a variety of
  increasingly common denial-of-service and amplification/forgery or
  cache poisoning attacks by off-path attackers.  It is noted, however,
  that they are designed to just verify IP addresses (and should change
  once a client's IP address changes), but they are not designed to
  actively track users (like HTTP cookies).

  There are anecdotal accounts of Media Access Control (MAC) addresses
  (https://lists.dns-oarc.net/pipermail/dns-
  operations/2016-January/014143.html) and even user names being
  inserted in nonstandard EDNS(0) options [RFC6891] for stub-to-
  resolver communications to support proprietary functionality
  implemented at the resolver (e.g., parental filtering).

4.3.  Cache Snooping

  The content of recursive resolvers' caches can reveal data about the
  clients using it (the privacy risks depend on the number of clients).
  This information can sometimes be examined by sending DNS queries
  with RD=0 to inspect cache content, particularly looking at the DNS
  TTLs [grangeia.snooping].  Since this also is a reconnaissance
  technique for subsequent cache poisoning attacks, some
  countermeasures have already been developed and deployed
  [cache-snooping-defence].

5.  Risks on the Wire

5.1.  Unencrypted Transports

  For unencrypted transports, DNS traffic can be seen by an
  eavesdropper like any other traffic.  (DNSSEC, specified in
  [RFC4033], explicitly excludes confidentiality from its goals.)  So,
  if an initiator starts an HTTPS communication with a recipient, the
  HTTP traffic will be encrypted, but the DNS exchange prior to it will
  not be.  When other protocols become more and more privacy aware and
  secured against surveillance (e.g., [RFC8446], [RFC9000]), the use of
  unencrypted transports for DNS may become "the weakest link" in
  privacy.  It is noted that, at the time of writing, there is ongoing
  work attempting to encrypt the Server Name Identification (SNI) in
  the TLS handshake [RFC8744], which is one of the last remaining non-
  DNS cleartext identifiers of a connection target.

  An important specificity of the DNS traffic is that it may take a
  different path than the communication between the initiator and the
  recipient.  For instance, an eavesdropper may be unable to tap the
  wire between the initiator and the recipient but may have access to
  the wire going to the recursive resolver or to the authoritative name
  servers.

  The best place to tap, from an eavesdropper's point of view, is
  clearly between the stub resolvers and the recursive resolvers,
  because traffic is not limited by DNS caching.

  The attack surface between the stub resolver and the rest of the
  world can vary widely depending upon how the end user's device is
  configured.  By order of increasing attack surface:

  *  The recursive resolver can be on the end user's device.  In
     (currently) a small number of cases, individuals may choose to
     operate their own DNS resolver on their local machine.  In this
     case, the attack surface for the connection between the stub
     resolver and the caching resolver is limited to that single
     machine.  The recursive resolver will expose data to authoritative
     resolvers as discussed in Section 6.2.

  *  The recursive resolver may be at the local network edge.  For
     many/most enterprise networks and for some residential networks,
     the caching resolver may exist on a server at the edge of the
     local network.  In this case, the attack surface is the local
     network.  Note that in large enterprise networks, the DNS resolver
     may not be located at the edge of the local network but rather at
     the edge of the overall enterprise network.  In this case, the
     enterprise network could be thought of as similar to the Internet
     Access Provider (IAP) network referenced below.

  *  The recursive resolver can be in the IAP network.  For most
     residential networks and potentially other networks, the typical
     case is for the user's device to be configured (typically
     automatically through DHCP or relay agent options) with the
     addresses of the DNS proxy in the Customer Premises Equipment
     (CPE), which in turn points to the DNS recursive resolvers at the
     IAP.  The attack surface for on-the-wire attacks is therefore from
     the end user system across the local network and across the IAP
     network to the IAP's recursive resolvers.

  *  The recursive resolver can be a public DNS service (or a privately
     run DNS resolver hosted on the public Internet).  Some machines
     may be configured to use public DNS resolvers such as those
     operated by Google Public DNS or OpenDNS.  The user may have
     configured their machine to use these DNS recursive resolvers
     themselves -- or their IAP may have chosen to use the public DNS
     resolvers rather than operating their own resolvers.  In this
     case, the attack surface is the entire public Internet between the
     user's connection and the public DNS service.  It can be noted
     that if the user selects a single resolver with a small client
     population (even when using an encrypted transport), it can
     actually serve to aid tracking of that user as they move across
     network environments.

  It is also noted that, typically, a device connected _only_ to a
  modern cellular network is

  *  directly configured with only the recursive resolvers of the IAP
     and

  *  afforded some level of protection against some types of
     eavesdropping for all traffic (including DNS traffic) due to the
     cellular network link-layer encryption.

  The attack surface for this specific scenario is not considered here.

5.2.  Encrypted Transports

  The use of encrypted transports directly mitigates passive
  surveillance of the DNS payload; however, some privacy attacks are
  still possible.  This section enumerates the residual privacy risks
  to an end user when an attacker can passively monitor encrypted DNS
  traffic flows on the wire.

  These are cases where user identification, fingerprinting, or
  correlations may be possible due to the use of certain transport
  layers or cleartext/observable features.  These issues are not
  specific to DNS, but DNS traffic is susceptible to these attacks when
  using specific transports.

  Some general examples exist; for example, certain studies highlight
  that the OS fingerprint values (http://netres.ec/?b=11B99BD) of IPv4
  TTL, IPv6 Hop Limit, or TCP Window size can be used to fingerprint
  client OSes or that various techniques can be used to de-NAT DNS
  queries [dns-de-nat].

  Note that even when using encrypted transports, the use of cleartext
  transport options to decrease latency can provide correlation of a
  user's connections, e.g., using TCP Fast Open [RFC7413].

  Implementations that support encrypted transports also commonly reuse
  connections for multiple DNS queries to optimize performance (e.g.,
  via DNS pipelining or HTTPS multiplexing).  Default configuration
  options for encrypted transports could, in principle, fingerprint a
  specific client application.  For example:

  *  TLS version or cipher suite selection

  *  session resumption

  *  the maximum number of messages to send and

  *  a maximum connection time before closing a connections and
     reopening.

  If libraries or applications offer user configuration of such options
  (e.g., [getdns]), then they could, in principle, help to identify a
  specific user.  Users may want to use only the defaults to avoid this
  issue.

  While there are known attacks on older versions of TLS, the most
  recent recommendations [RFC7525] and the development of TLS 1.3
  [RFC8446] largely mitigate those.

  Traffic analysis of unpadded encrypted traffic is also possible
  [pitfalls-of-dns-encryption] because the sizes and timing of
  encrypted DNS requests and responses can be correlated to unencrypted
  DNS requests upstream of a recursive resolver.

6.  Risks in the Servers

  Using the terminology of [RFC6973], the DNS servers (recursive
  resolvers and authoritative servers) are enablers: "they facilitate
  communication between an initiator and a recipient without being
  directly in the communications path".  As a result, they are often
  forgotten in risk analysis.  But, to quote [RFC6973] again, "Although
  [...] enablers may not generally be considered as attackers, they may
  all pose privacy threats (depending on the context) because they are
  able to observe, collect, process, and transfer privacy-relevant
  data".  In [RFC6973] parlance, enablers become observers when they
  start collecting data.

  Many programs exist to collect and analyze DNS data at the servers --
  from the "query log" of some programs like BIND to tcpdump and more
  sophisticated programs like PacketQ [packetq] and DNSmezzo
  [dnsmezzo].  The organization managing the DNS server can use this
  data itself, or it can be part of a surveillance program like PRISM
  [prism] and pass data to an outside observer.

  Sometimes this data is kept for a long time and/or distributed to
  third parties for research purposes [ditl] [day-at-root], security
  analysis, or surveillance tasks.  These uses are sometimes under some
  sort of contract, with various limitations, for instance, on
  redistribution, given the sensitive nature of the data.  Also, there
  are observation points in the network that gather DNS data and then
  make it accessible to third parties for research or security purposes
  ("passive DNS" [passive-dns]).

6.1.  In the Recursive Resolvers

  Recursive resolvers see all the traffic since there is typically no
  caching before them.  To summarize: your recursive resolver knows a
  lot about you.  The resolver of a large IAP, or a large public
  resolver, can collect data from many users.

6.1.1.  Resolver Selection

  Given all the above considerations, the choice of recursive resolver
  has direct privacy considerations for end users.  Historically, end
  user devices have used the DHCP-provided local network recursive
  resolver.  The choice by a user to join a particular network (e.g.,
  by physically plugging in a cable or selecting a network in an OS
  dialogue) typically updates a number of system resources -- these can
  include IP addresses, the availability of IPv4/IPv6, DHCP server, and
  DNS resolver.  These individual changes, including the change in DNS
  resolver, are not normally communicated directly to the user by the
  OS when the network is joined.  The choice of network has
  historically determined the default system DNS resolver selection;
  the two are directly coupled in this model.

  The vast majority of users do not change their default system DNS
  settings and so implicitly accept the network settings for the DNS.
  The network resolvers have therefore historically been the sole
  destination for all of the DNS queries from a device.  These
  resolvers may have varied privacy policies depending on the network.
  Privacy policies for these servers may or may not be available, and
  users need to be aware that privacy guarantees will vary with the
  network.

  All major OSes expose the system DNS settings and allow users to
  manually override them if desired.

  More recently, some networks and users have actively chosen to use a
  large public resolver, e.g., Google Public DNS
  (https://developers.google.com/speed/public-dns), Cloudflare
  (https://developers.cloudflare.com/1.1.1.1/setting-up-1.1.1.1/), or
  Quad9 (https://www.quad9.net).  There can be many reasons: cost
  considerations for network operators, better reliability, or anti-
  censorship considerations are just a few.  Such services typically do
  provide a privacy policy, and the user can get an idea of the data
  collected by such operators by reading one, e.g., Google Public DNS -
  Your Privacy (https://developers.google.com/speed/public-dns/
  privacy).

  In general, as with many other protocols, issues around
  centralization also arise with DNS.  The picture is fluid with
  several competing factors contributing, where these factors can also
  vary by geographic region.  These include:

  *  ISP outsourcing, including to third-party and public resolvers

  *  regional market domination by one or only a few ISPs

  *  applications directing DNS traffic by default to a limited subset
     of resolvers (see Section 6.1.1.2)

  An increased proportion of the global DNS resolution traffic being
  served by only a few entities means that the privacy considerations
  for users are highly dependent on the privacy policies and practices
  of those entities.  Many of the issues around centralization are
  discussed in [centralisation-and-data-sovereignty].

6.1.1.1.  Dynamic Discovery of DoH and Strict DoT

  While support for opportunistic DoT can be determined by probing a
  resolver on port 853, there is currently no standardized discovery
  mechanism for DoH and Strict DoT servers.

  This means that clients that might want to dynamically discover such
  encrypted services, and where users are willing to trust such
  services, are not able to do so.  At the time of writing, efforts to
  provide standardized signaling mechanisms to discover the services
  offered by local resolvers are in progress [DNSOP-RESOLVER].  Note
  that an increasing number of ISPs are deploying encrypted DNS; for
  example, see the Encrypted DNS Deployment Initiative [EDDI].

6.1.1.2.  Application-Specific Resolver Selection

  An increasing number of applications are offering application-
  specific encrypted DNS resolution settings, rather than defaulting to
  using only the system resolver.  A variety of heuristics and
  resolvers are available in different applications, including hard-
  coded lists of recognized DoH/DoT servers.

  Generally, users are not aware of application-specific DNS settings
  and may not have control over those settings.  To address these
  limitations, users will only be aware of and have the ability to
  control such settings if applications provide the following
  functions:

  *  communicate the change clearly to users when the default
     application resolver changes away from the system resolver

  *  provide configuration options to change the default application
     resolver, including a choice to always use the system resolver

  *  provide mechanisms for users to locally inspect, selectively
     forward, and filter queries (either via the application itself or
     use of the system resolver)

  Application-specific changes to default destinations for users' DNS
  queries might increase or decrease user privacy; it is highly
  dependent on the network context and the application-specific
  default.  This is an area of active debate, and the IETF is working
  on a number of issues related to application-specific DNS settings.

6.1.2.  Active Attacks on Resolver Configuration

  The previous section discussed DNS privacy, assuming that all the
  traffic was directed to the intended servers (i.e., those that would
  be used in the absence of an active attack) and that the potential
  attacker was purely passive.  But, in reality, there can be active
  attackers in the network.

  The Internet Threat model, as described in [RFC3552], assumes that
  the attacker controls the network.  Such an attacker can completely
  control any insecure DNS resolution, both passively monitoring the
  queries and responses and substituting their own responses.  Even if
  encrypted DNS such as DoH or DoT is used, unless the client has been
  configured in a secure way with the server identity, an active
  attacker can impersonate the server.  This implies that opportunistic
  modes of DoH/DoT as well as modes where the client learns of the DoH/
  DoT server via in-network mechanisms such as DHCP are vulnerable to
  attack.  In addition, if the client is compromised, the attacker can
  replace the DNS configuration with one of its own choosing.

6.1.3.  Blocking of DNS Resolution Services

  User privacy can also be at risk if there is blocking of access to
  remote recursive servers that offer encrypted transports, e.g., when
  the local resolver does not offer encryption and/or has very poor
  privacy policies.  For example, active blocking of port 853 for DoT
  or blocking of specific IP addresses could restrict the resolvers
  available to the user.  The extent of the risk to user privacy is
  highly dependent on the specific network and user context; a user on
  a network that is known to perform surveillance would be compromised
  if they could not access such services, whereas a user on a trusted
  network might have no privacy motivation to do so.

  As a matter of policy, some recursive resolvers use their position in
  the query path to selectively block access to certain DNS records.
  This is a form of rendezvous-based blocking as described in
  Section 4.3 of [RFC7754].  Such blocklists often include servers
  known to be used for malware, bots, or other security risks.  In
  order to prevent circumvention of their blocking policies, some
  networks also block access to resolvers with incompatible policies.

  It is also noted that attacks on remote resolver services, e.g.,
  DDoS, could force users to switch to other services that do not offer
  encrypted transports for DNS.

6.1.4.  Encrypted Transports and Recursive Resolvers

6.1.4.1.  DoT and DoH

  Use of encrypted transports does not reduce the data available in the
  recursive resolver and ironically can actually expose more
  information about users to operators.  As described in Section 5.2,
  use of session-based encrypted transports (TCP/TLS) can expose
  correlation data about users.

6.1.4.2.  DoH-Specific Considerations

  DoH inherits the full privacy properties of the HTTPS stack and as a
  consequence introduces new privacy considerations when compared with
  DNS over UDP, TCP, or TLS [RFC7858].  Section 8.2 of [RFC8484]
  describes the privacy considerations in the server of the DoH
  protocol.

  A brief summary of some of the issues includes the following:

  *  HTTPS presents new considerations for correlation, such as
     explicit HTTP cookies and implicit fingerprinting of the unique
     set and ordering of HTTP request header fields.

  *  The User-Agent and Accept-Language request header fields often
     convey specific information about the client version or locale.

  *  Utilizing the full set of HTTP features enables DoH to be more
     than an HTTP tunnel, but it is at the cost of opening up
     implementations to the full set of privacy considerations of HTTP.

  *  Implementations are advised to expose the minimal set of data
     needed to achieve the desired feature set.

  [RFC8484] specifically makes selection of HTTPS functionality vs.
  privacy an implementation choice.  At the extremes, there may be
  implementations that attempt to achieve parity with DoT from a
  privacy perspective at the cost of using no identifiable HTTP
  headers, and there might be others that provide feature-rich data
  flows where the low-level origin of the DNS query is easily
  identifiable.  Some implementations have, in fact, chosen to restrict
  the use of the User-Agent header so that resolver operators cannot
  identify the specific application that is originating the DNS
  queries.

  Privacy-focused users should be aware of the potential for additional
  client identifiers in DoH compared to DoT and may want to only use
  DoH client implementations that provide clear guidance on what
  identifiers they add.

6.2.  In the Authoritative Name Servers

  Unlike what happens for recursive resolvers, the observation
  capabilities of authoritative name servers are limited by caching;
  they see only the requests for which the answer was not in the cache.
  For aggregated statistics ("What is the percentage of LOC queries?"),
  this is sufficient, but it prevents an observer from seeing
  everything.  Similarly, the increasing deployment of QNAME
  minimization [ripe-qname-measurements] reduces the data visible at
  the authoritative name server.  Still, the authoritative name servers
  see a part of the traffic, and this subset may be sufficient to
  violate some privacy expectations.

  Also, the user often has some legal/contractual link with the
  recursive resolver (they have chosen the IAP, or they have chosen to
  use a given public resolver) while having no control and perhaps no
  awareness of the role of the authoritative name servers and their
  observation abilities.

  As noted before, using a local resolver or a resolver close to the
  machine decreases the attack surface for an on-the-wire eavesdropper.
  But it may decrease privacy against an observer located on an
  authoritative name server.  This authoritative name server will see
  the IP address of the end client instead of the address of a big
  recursive resolver shared by many users.

  This "protection", when using a large resolver with many clients, is
  no longer present if ECS [RFC7871] is used because, in this case, the
  authoritative name server sees the original IP address (or prefix,
  depending on the setup).

  As of today, all the instances of one root name server, L-root,
  receive together around 50,000 queries per second.  While most of it
  is "junk" (errors on the Top-Level Domain (TLD) name), it gives an
  idea of the amount of big data that pours into name servers.  (And
  even "junk" can leak information; for instance, if there is a typing
  error in the TLD, the user will send data to a TLD that is not the
  usual one.)

  Many domains, including TLDs, are partially hosted by third-party
  servers, sometimes in a different country.  The contracts between the
  domain manager and these servers may or may not take privacy into
  account.  Whatever the contract, the third-party hoster may or may
  not be honest; in any case, it will have to follow its local laws.
  For example, requests to a given ccTLD may go to servers managed by
  organizations outside of the ccTLD's country.  Users may not
  anticipate that when doing a security analysis.

  Also, it seems (see the survey described in [aeris-dns]) that there
  is a strong concentration of authoritative name servers among
  "popular" domains (such as the Alexa Top N list).  For instance,
  among the Alexa Top 100K (https://www.alexa.com/topsites), one DNS
  provider hosts 10% of the domains today.  The ten most important DNS
  providers together host one-third of all domains.  With the control
  (or the ability to sniff the traffic) of a few name servers, you can
  gather a lot of information.

7.  Other Risks

7.1.  Re-identification and Other Inferences

  An observer has access not only to the data they directly collect but
  also to the results of various inferences about this data.  The term
  "observer" here is used very generally; for example, the observer
  might passively observe cleartext DNS traffic or be in the network
  that is actively attacking the user by redirecting DNS resolution, or
  it might be a local or remote resolver operator.

  For instance, a user can be re-identified via DNS queries.  If the
  adversary knows a user's identity and can watch their DNS queries for
  a period, then that same adversary may be able to re-identify the
  user solely based on their pattern of DNS queries later on regardless
  of the location from which the user makes those queries.  For
  example, one study [herrmann-reidentification] found that such re-
  identification is possible so that "73.1% of all day-to-day links
  were correctly established, i.e. user u was either re-identified
  unambiguously (1) or the classifier correctly reported that u was not
  present on day t + 1 any more (2)".  While that study related to web
  browsing behavior, equally characteristic patterns may be produced
  even in machine-to-machine communications or without a user taking
  specific actions, e.g., at reboot time if a characteristic set of
  services are accessed by the device.

  For instance, one could imagine that an intelligence agency
  identifies people going to a site by putting in a very long DNS name
  and looking for queries of a specific length.  Such traffic analysis
  could weaken some privacy solutions.

  The IAB Privacy and Security Program also has a document [RFC7624]
  that considers such inference-based attacks in a more general
  framework.

7.2.  More Information

  Useful background information can also be found in [tor-leak]
  (regarding the risk of privacy leaks through DNS) and in a few
  academic papers: [yanbin-tsudik], [castillo-garcia],
  [fangming-hori-sakurai], and [federrath-fuchs-herrmann-piosecny].

8.  Actual "Attacks"

  A very quick examination of DNS traffic may lead to the false
  conclusion that extracting the needle from the haystack is difficult.
  "Interesting" primary DNS requests are mixed with useless (for the
  eavesdropper) secondary and tertiary requests (see the terminology in
  Section 1).  But, in this time of "big data" processing, powerful
  techniques now exist to get from the raw data to what the
  eavesdropper is actually interested in.

  Many research papers about malware detection use DNS traffic to
  detect "abnormal" behavior that can be traced back to the activity of
  malware on infected machines.  Yes, this research was done for the
  greater good, but technically it is a privacy attack and it
  demonstrates the power of the observation of DNS traffic.  See
  [dns-footprint], [dagon-malware], and [darkreading-dns].

  Passive DNS services [passive-dns] allow reconstruction of the data
  of sometimes an entire zone.  Well-known passive DNS services keep
  only the DNS responses and not the source IP address of the client,
  precisely for privacy reasons.  Other passive DNS services may not be
  so careful.  And there are still potential problems with revealing
  QNAMEs.

  The revelations from the Edward Snowden documents, which were leaked
  from the National Security Agency (NSA), provide evidence of the use
  of the DNS in mass surveillance operations [morecowbell].  For
  example, the MORECOWBELL surveillance program uses a dedicated covert
  monitoring infrastructure to actively query DNS servers and perform
  HTTP requests to obtain meta-information about services and to check
  their availability.  Also, the QUANTUMTHEORY
  (https://theintercept.com/document/2014/03/12/nsa-gchqs-
  quantumtheory-hacking-tactics/) project, which includes detecting
  lookups for certain addresses and injecting bogus replies, is another
  good example showing that the lack of privacy protections in the DNS
  is actively exploited.

9.  Legalities

  To our knowledge, there are no specific privacy laws for DNS data in
  any country.  Interpreting general privacy laws, like the European
  Union's [data-protection-directive] or GDPR (https://gdpr.eu/tag/
  gdpr/), in the context of DNS traffic data is not an easy task, and
  there is no known court precedent.  See an interesting analysis in
  [sidn-entrada].

10.  Security Considerations

  This document is entirely about security -- more precisely, privacy.
  It just lays out the problem; it does not try to set requirements
  (with the choices and compromises they imply), much less define
  solutions.  Possible solutions to the issues described here are
  discussed in other documents (currently too many to all be
  mentioned); see, for instance, "Recommendations for DNS Privacy
  Operators" [RFC8932].

11.  IANA Considerations

  This document has no IANA actions.

12.  References

12.1.  Normative References

  [RFC1034]  Mockapetris, P., "Domain names - concepts and facilities",
             STD 13, RFC 1034, DOI 10.17487/RFC1034, November 1987,
             <https://www.rfc-editor.org/info/rfc1034>.

  [RFC1035]  Mockapetris, P., "Domain names - implementation and
             specification", STD 13, RFC 1035, DOI 10.17487/RFC1035,
             November 1987, <https://www.rfc-editor.org/info/rfc1035>.

  [RFC6973]  Cooper, A., Tschofenig, H., Aboba, B., Peterson, J.,
             Morris, J., Hansen, M., and R. Smith, "Privacy
             Considerations for Internet Protocols", RFC 6973,
             DOI 10.17487/RFC6973, July 2013,
             <https://www.rfc-editor.org/info/rfc6973>.

  [RFC7258]  Farrell, S. and H. Tschofenig, "Pervasive Monitoring Is an
             Attack", BCP 188, RFC 7258, DOI 10.17487/RFC7258, May
             2014, <https://www.rfc-editor.org/info/rfc7258>.

12.2.  Informative References

  [aeris-dns]
             Vinot, N., "Vie privée: et le DNS alors? [Privacy: what
             about DNS?]", February 2015,
             <https://blog.imirhil.fr/vie-privee-et-le-dns-alors.html>.

  [cache-snooping-defence]
             ISC, "DNS Cache snooping - should I be concerned?",
             October 2018, <https://kb.isc.org/docs/aa-00482>.

  [castillo-garcia]
             Castillo-Perez, S. and J. Garcia-Alfaro, "Anonymous
             Resolution of DNS Queries", Lecture Notes in Computer
             Science, Vol. 5332, DOI 10.1007/978-3-540-88873-4_5, 2008,
             <https://dl.acm.org/doi/10.1007/978-3-540-88873-4_5>.

  [centralisation-and-data-sovereignty]
             De Filippi, P. and S. McCarthy, "Cloud Computing:
             Centralization and Data Sovereignty", European Journal of
             Law and Technology, Vol. 3, No. 2, October 2012,
             <https://papers.ssrn.com/sol3/
             papers.cfm?abstract_id=2167372>.

  [dagon-malware]
             Dagon, D., "Corrupted DNS Resolution Paths: The Rise of a
             Malicious Resolution Authority", ISC/OARC Workshop, 2007,
             <https://www.dns-oarc.net/files/workshop-2007/Dagon-
             Resolution-corruption.pdf>.

  [darkreading-dns]
             Lemos, R., "Got Malware? Three Signs Revealed In DNS
             Traffic", May 2013,
             <https://www.darkreading.com/analytics/security-
             monitoring/got-malware-three-signs-revealed-in-dns-
             traffic/d/d-id/1139680>.

  [data-protection-directive]
             European Parliament, "Directive 95/46/EC of the European
             Parliament and of the Council of 24 October 1995 on the
             protection of individuals with regard to the processing of
             personal data and on the free movement of such data",
             Official Journal L 281, pp. 31-50, November 1995,
             <https://eur-lex.europa.eu/LexUriServ/
             LexUriServ.do?uri=CELEX:31995L0046:EN:HTML>.

  [day-at-root]
             Castro, S., Wessels, D., Fomenkov, M., and K. Claffy, "A
             Day at the Root of the Internet", ACM SIGCOMM Computer
             Communication Review, Vol. 38, No. 5,
             DOI 10.1145/1452335.1452341, October 2008,
             <https://www.sigcomm.org/sites/default/files/ccr/
             papers/2008/October/1452335-1452341.pdf>.

  [denis-edns-client-subnet]
             Denis, F., "Security and privacy issues of edns-client-
             subnet", August 2013,
             <https://00f.net/2013/08/07/edns-client-subnet/>.

  [ditl]     CAIDA, "A Day in the Life of the Internet (DITL)",
             <https://www.caida.org/projects/ditl/>.

  [dns-de-nat]
             Orevi, L., Herzberg, A., Zlatokrilov, H., and D. Sigron,
             "DNS-DNS: DNS-based De-NAT Scheme", January 2017,
             <https://www.researchgate.net/publication/320322146_DNS-
             DNS_DNS-based_De-NAT_Scheme>.

  [dns-footprint]
             Stoner, E., "DNS Footprint of Malware", OARC Workshop,
             October 2010, <https://www.dns-oarc.net/files/workshop-
             201010/OARC-ers-20101012.pdf>.

  [dns-over-encryption]
             Lu, C., Liu, B., Li, Z., Hao, S., Duan, H., Zhang, M.,
             Leng, C., Liu, Y., Zhang, Z., and J. Wu, "An End-to-End,
             Large-Scale Measurement of DNS-over-Encryption: How Far
             Have We Come?", IMC '19: Proceedings of the Internet
             Measurement Conference, pp. 22-35,
             DOI 10.1145/3355369.3355580, October 2019,
             <https://dl.acm.org/citation.cfm?id=3355369.3355580>.

  [dnsmezzo] Bortzmeyer, S., "DNSmezzo", <http://www.dnsmezzo.net/>.

  [DNSOP-RESOLVER]
             Sood, P., Arends, R., and P. Hoffman, "DNS Resolver
             Information Self-publication", Work in Progress, Internet-
             Draft, draft-ietf-dnsop-resolver-information-01, 11
             February 2020, <https://datatracker.ietf.org/doc/html/
             draft-ietf-dnsop-resolver-information-01>.

  [DPRIVE-DNSOQUIC]
             Huitema, C., Dickinson, S., and A. Mankin, "Specification
             of DNS over Dedicated QUIC Connections", Work in Progress,
             Internet-Draft, draft-ietf-dprive-dnsoquic-03, 12 July
             2021, <https://datatracker.ietf.org/doc/html/draft-ietf-
             dprive-dnsoquic-03>.

  [EDDI]     EDDI, "Encrypted DNS Deployment Initiative",
             <https://www.encrypted-dns.org>.

  [fangming-hori-sakurai]
             Zhao, F., Hori, Y., and K. Sakurai, "Analysis of Privacy
             Disclosure in DNS Query", MUE '07: Proceedings of the 2007
             International Conference on Multimedia and Ubiquitous
             Engineering, pp. 952-957, DOI 10.1109/MUE.2007.84,
             ISBN 0-7695-2777-9, April 2007,
             <https://dl.acm.org/citation.cfm?id=1262690.1262986>.

  [federrath-fuchs-herrmann-piosecny]
             Federrath, H., Fuchs, K.-P., Herrmann, D., and C.
             Piosecny, "Privacy-Preserving DNS: Analysis of Broadcast,
             Range Queries and Mix-based Protection Methods", ESORICS
             2011, pp. 665-683, DOI 10.1007/978-3-642-23822-2_36,
             ISBN 978-3-642-23822-2, 2011, <https://svs.informatik.uni-
             hamburg.de/publications/2011/2011-09-14_FFHP_PrivacyPreser
             vingDNS_ESORICS2011.pdf>.

  [getdns]   "getdns", <https://getdnsapi.net>.

  [grangeia.snooping]
             Grangeia, L., "Cache Snooping or Snooping the Cache for
             Fun and Profit", 2005,
             <https://www.semanticscholar.org/paper/Cache-Snooping-or-
             Snooping-the-Cache-for-Fun-and-
             1-Grangeia/9b22f606e10b3609eafbdcbfc9090b63be8778c3>.

  [herrmann-reidentification]
             Herrmann, D., Gerber, C., Banse, C., and H. Federrath,
             "Analyzing Characteristic Host Access Patterns for Re-
             Identification of Web User Sessions", Lecture Notes in
             Computer Science, Vol. 7127,
             DOI 10.1007/978-3-642-27937-9_10, 2012, <https://epub.uni-
             regensburg.de/21103/1/Paper_PUL_nordsec_published.pdf>.

  [morecowbell]
             Grothoff, C., Wachs, M., Ermert, M., and J. Appelbaum,
             "NSA's MORECOWBELL: Knell for DNS", January 2015, <https:/
             /pdfs.semanticscholar.org/2610/2b99bdd6a258a98740af8217ba8
             da8a1e4fa.pdf>.

  [packetq]  DNS-OARC, "A tool that provides a basic SQL-frontend to
             PCAP-files", Release 1.4.3, commit 29a8288, October 2020,
             <https://github.com/DNS-OARC/PacketQ>.

  [passive-dns]
             Weimer, F., "Passive DNS Replication", 17th Annual FIRST
             Conference, April 2005,
             <https://www.first.org/conference/2005/papers/florian-
             weimer-slides-1.pdf>.

  [pitfalls-of-dns-encryption]
             Shulman, H., "Pretty Bad Privacy: Pitfalls of DNS
             Encryption", WPES '14: Proceedings of the 13th Workshop on
             Privacy in the Electronic Society, pp. 191-200,
             DOI 10.1145/2665943.2665959, November 2014,
             <https://dl.acm.org/citation.cfm?id=2665959>.

  [prism]    Wikipedia, "PRISM (surveillance program)", July 2015,
             <https://en.wikipedia.org/w/index.php?title=PRISM_(surveil
             lance_program)&oldid=673789455>.

  [RFC3552]  Rescorla, E. and B. Korver, "Guidelines for Writing RFC
             Text on Security Considerations", BCP 72, RFC 3552,
             DOI 10.17487/RFC3552, July 2003,
             <https://www.rfc-editor.org/info/rfc3552>.

  [RFC4033]  Arends, R., Austein, R., Larson, M., Massey, D., and S.
             Rose, "DNS Security Introduction and Requirements",
             RFC 4033, DOI 10.17487/RFC4033, March 2005,
             <https://www.rfc-editor.org/info/rfc4033>.

  [RFC4470]  Weiler, S. and J. Ihren, "Minimally Covering NSEC Records
             and DNSSEC On-line Signing", RFC 4470,
             DOI 10.17487/RFC4470, April 2006,
             <https://www.rfc-editor.org/info/rfc4470>.

  [RFC5155]  Laurie, B., Sisson, G., Arends, R., and D. Blacka, "DNS
             Security (DNSSEC) Hashed Authenticated Denial of
             Existence", RFC 5155, DOI 10.17487/RFC5155, March 2008,
             <https://www.rfc-editor.org/info/rfc5155>.

  [RFC5936]  Lewis, E. and A. Hoenes, Ed., "DNS Zone Transfer Protocol
             (AXFR)", RFC 5936, DOI 10.17487/RFC5936, June 2010,
             <https://www.rfc-editor.org/info/rfc5936>.

  [RFC6269]  Ford, M., Ed., Boucadair, M., Durand, A., Levis, P., and
             P. Roberts, "Issues with IP Address Sharing", RFC 6269,
             DOI 10.17487/RFC6269, June 2011,
             <https://www.rfc-editor.org/info/rfc6269>.

  [RFC6891]  Damas, J., Graff, M., and P. Vixie, "Extension Mechanisms
             for DNS (EDNS(0))", STD 75, RFC 6891,
             DOI 10.17487/RFC6891, April 2013,
             <https://www.rfc-editor.org/info/rfc6891>.

  [RFC7413]  Cheng, Y., Chu, J., Radhakrishnan, S., and A. Jain, "TCP
             Fast Open", RFC 7413, DOI 10.17487/RFC7413, December 2014,
             <https://www.rfc-editor.org/info/rfc7413>.

  [RFC7525]  Sheffer, Y., Holz, R., and P. Saint-Andre,
             "Recommendations for Secure Use of Transport Layer
             Security (TLS) and Datagram Transport Layer Security
             (DTLS)", BCP 195, RFC 7525, DOI 10.17487/RFC7525, May
             2015, <https://www.rfc-editor.org/info/rfc7525>.

  [RFC7624]  Barnes, R., Schneier, B., Jennings, C., Hardie, T.,
             Trammell, B., Huitema, C., and D. Borkmann,
             "Confidentiality in the Face of Pervasive Surveillance: A
             Threat Model and Problem Statement", RFC 7624,
             DOI 10.17487/RFC7624, August 2015,
             <https://www.rfc-editor.org/info/rfc7624>.

  [RFC7626]  Bortzmeyer, S., "DNS Privacy Considerations", RFC 7626,
             DOI 10.17487/RFC7626, August 2015,
             <https://www.rfc-editor.org/info/rfc7626>.

  [RFC7721]  Cooper, A., Gont, F., and D. Thaler, "Security and Privacy
             Considerations for IPv6 Address Generation Mechanisms",
             RFC 7721, DOI 10.17487/RFC7721, March 2016,
             <https://www.rfc-editor.org/info/rfc7721>.

  [RFC7754]  Barnes, R., Cooper, A., Kolkman, O., Thaler, D., and E.
             Nordmark, "Technical Considerations for Internet Service
             Blocking and Filtering", RFC 7754, DOI 10.17487/RFC7754,
             March 2016, <https://www.rfc-editor.org/info/rfc7754>.

  [RFC7816]  Bortzmeyer, S., "DNS Query Name Minimisation to Improve
             Privacy", RFC 7816, DOI 10.17487/RFC7816, March 2016,
             <https://www.rfc-editor.org/info/rfc7816>.

  [RFC7858]  Hu, Z., Zhu, L., Heidemann, J., Mankin, A., Wessels, D.,
             and P. Hoffman, "Specification for DNS over Transport
             Layer Security (TLS)", RFC 7858, DOI 10.17487/RFC7858, May
             2016, <https://www.rfc-editor.org/info/rfc7858>.

  [RFC7871]  Contavalli, C., van der Gaast, W., Lawrence, D., and W.
             Kumari, "Client Subnet in DNS Queries", RFC 7871,
             DOI 10.17487/RFC7871, May 2016,
             <https://www.rfc-editor.org/info/rfc7871>.

  [RFC7873]  Eastlake 3rd, D. and M. Andrews, "Domain Name System (DNS)
             Cookies", RFC 7873, DOI 10.17487/RFC7873, May 2016,
             <https://www.rfc-editor.org/info/rfc7873>.

  [RFC7929]  Wouters, P., "DNS-Based Authentication of Named Entities
             (DANE) Bindings for OpenPGP", RFC 7929,
             DOI 10.17487/RFC7929, August 2016,
             <https://www.rfc-editor.org/info/rfc7929>.

  [RFC8446]  Rescorla, E., "The Transport Layer Security (TLS) Protocol
             Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018,
             <https://www.rfc-editor.org/info/rfc8446>.

  [RFC8484]  Hoffman, P. and P. McManus, "DNS Queries over HTTPS
             (DoH)", RFC 8484, DOI 10.17487/RFC8484, October 2018,
             <https://www.rfc-editor.org/info/rfc8484>.

  [RFC8499]  Hoffman, P., Sullivan, A., and K. Fujiwara, "DNS
             Terminology", BCP 219, RFC 8499, DOI 10.17487/RFC8499,
             January 2019, <https://www.rfc-editor.org/info/rfc8499>.

  [RFC8744]  Huitema, C., "Issues and Requirements for Server Name
             Identification (SNI) Encryption in TLS", RFC 8744,
             DOI 10.17487/RFC8744, July 2020,
             <https://www.rfc-editor.org/info/rfc8744>.

  [RFC8890]  Nottingham, M., "The Internet is for End Users", RFC 8890,
             DOI 10.17487/RFC8890, August 2020,
             <https://www.rfc-editor.org/info/rfc8890>.

  [RFC8932]  Dickinson, S., Overeinder, B., van Rijswijk-Deij, R., and
             A. Mankin, "Recommendations for DNS Privacy Service
             Operators", BCP 232, RFC 8932, DOI 10.17487/RFC8932,
             October 2020, <https://www.rfc-editor.org/info/rfc8932>.

  [RFC9000]  Iyengar, J., Ed. and M. Thomson, Ed., "QUIC: A UDP-Based
             Multiplexed and Secure Transport", RFC 9000,
             DOI 10.17487/RFC9000, May 2021,
             <https://www.rfc-editor.org/info/rfc9000>.

  [ripe-qname-measurements]
             de Vries, W., "Making the DNS More Private with QNAME
             Minimisation", April 2019,
             <https://labs.ripe.net/Members/wouter_de_vries/make-dns-a-
             bit-more-private-with-qname-minimisation>.

  [sidn-entrada]
             Hesselman, C., Jansen, J., Wullink, M., Vink, K., and M.
             Simon, "A privacy framework for 'DNS big data'
             applications", November 2014,
             <https://www.sidnlabs.nl/downloads/
             yBW6hBoaSZe4m6GJc_0b7w/2211058ab6330c7f3788141ea19d3db7/
             SIDN_Labs_Privacyraamwerk_Position_Paper_V1.4_ENG.pdf>.

  [thomas-ditl-tcp]
             Thomas, M. and D. Wessels, "An Analysis of TCP Traffic in
             Root Server DITL Data", DNS-OARC 2014 Fall Workshop,
             October 2014, <https://indico.dns-
             oarc.net/event/20/session/2/contribution/15/material/
             slides/1.pdf>.

  [tor-leak] Tor, "Tor FAQs: I keep seeing these warnings about SOCKS
             and DNS information leaks. Should I worry?",
             <https://www.torproject.org/docs/
             faq.html.en#WarningsAboutSOCKSandDNSInformationLeaks>.

  [yanbin-tsudik]
             Yanbin, L. and G. Tsudik, "Towards Plugging Privacy Leaks
             in Domain Name System", June 2010,
             <https://arxiv.org/abs/0910.2472>.

Appendix A.  Updates since RFC 7626

  Many references were updated.  Discussions of encrypted transports,
  including DoT and DoH, and sections on DNS payload, authentication of
  servers, and blocking of services were added.  With the publishing of
  [RFC7816] on QNAME minimization, text, references, and initial
  attempts to measure deployment were added to reflect this.  The text
  and references on the Snowden revelations were updated.

  The "Risks Overview" section was changed to "Scope" to help clarify
  the risks being considered.  Text on cellular network DNS, blocking,
  and security was added.  Considerations for recursive resolvers were
  collected and placed together.  A discussion on resolver selection
  was added.

Acknowledgments

  Thanks to Nathalie Boulvard and to the CENTR members for the original
  work that led to this document.  Thanks to Ondrej Sury for the
  interesting discussions.  Thanks to Mohsen Souissi and John Heidemann
  for proofreading and to Paul Hoffman, Matthijs Mekking, Marcos Sanz,
  Francis Dupont, Allison Mankin, and Warren Kumari for proofreading,
  providing technical remarks, and making many readability
  improvements.  Thanks to Dan York, Suzanne Woolf, Tony Finch, Stephen
  Farrell, Peter Koch, Simon Josefsson, and Frank Denis for good
  written contributions.  Thanks to Vittorio Bertola and Mohamed
  Boucadair for a detailed review of the -bis.  And thanks to the IESG
  members for the last remarks.

Contributions

  Sara Dickinson and Stephane Bortzmeyer were the original authors of
  the document, and their contribution to the initial draft of this
  document is greatly appreciated.

Author's Address

  Tim Wicinski (editor)
  Elkins, WV 26241
  United States of America

  Email: [email protected]

You are viewing proxied material from gopher.fnord.one. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.