Recursive to Authoritative DNS with Opportunistic Encryption
draft-ietf-dprive-opportunistic-adotq-00

Network Working Group                                         P. Hoffman
Internet-Draft                                                     ICANN
Intended status: Standards Track                             P. van Dijk
Expires: 18 August 2021                                         PowerDNS
                                                        14 February 2021

      Recursive to Authoritative DNS with Opportunistic Encryption
               draft-ietf-dprive-opportunistic-adotq-00

Abstract

   This document describes a use case and a method for a DNS recursive
   resolver to use opportunistic encryption (that is, encryption with
   optional authentication) when communicating with authoritative
   servers.  The motivating use case for this method is that more
   encryption on the Internet is better, and opportunistic encryption is
   better than no encryption at all.  The method described here is
   optional for both the recursive resolver and the authoritative
   server.  Nothing in this method prevents use cases and methods that
   require authenticated encryption.

Status of This Memo

   This Internet-Draft is submitted in full conformance with the
   provisions of BCP 78 and BCP 79.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF).  Note that other groups may also distribute
   working documents as Internet-Drafts.  The list of current Internet-
   Drafts is at https://datatracker.ietf.org/drafts/current/.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress."

   This Internet-Draft will expire on 18 August 2021.

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

Hoffman & van Dijk       Expires 18 August 2021                 [Page 1]
Internet-Draft             Opportunistic ADoTQ             February 2021

   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
     1.1.  Use Case  . . . . . . . . . . . . . . . . . . . . . . . .   3
     1.2.  Summary of Protocol . . . . . . . . . . . . . . . . . . .   3
     1.3.  Definitions . . . . . . . . . . . . . . . . . . . . . . .   4
   2.  Method for Opportunistic Encryption . . . . . . . . . . . . .   4
     2.1.  Resolvers . . . . . . . . . . . . . . . . . . . . . . . .   4
     2.2.  Authoritative Servers . . . . . . . . . . . . . . . . . .   5
   3.  Discovering Whether an Authoritative Server Uses
           Encryption  . . . . . . . . . . . . . . . . . . . . . . .   6
   4.  The Transport Cache . . . . . . . . . . . . . . . . . . . . .   6
   5.  Authentication  . . . . . . . . . . . . . . . . . . . . . . .   7
   6.  IANA Considerations . . . . . . . . . . . . . . . . . . . . .   8
   7.  Security Considerations . . . . . . . . . . . . . . . . . . .   8
   8.  Acknowledgements  . . . . . . . . . . . . . . . . . . . . . .   8
   9.  References  . . . . . . . . . . . . . . . . . . . . . . . . .   8
     9.1.  Normative References  . . . . . . . . . . . . . . . . . .   8
     9.2.  Informative References  . . . . . . . . . . . . . . . . .   9
   Authors' Addresses  . . . . . . . . . . . . . . . . . . . . . . .   9

1.  Introduction

   A recursive resolver using traditional DNS over port 53 may wish
   instead to use encrypted communication with authoritative servers in
   order to limit passive snooping of its DNS traffic.  The recursive
   resolver can use opportunistic encryption (defined in [RFC7435] to
   achieve this goal.

   This document describes a use case and a method for recursive
   resolvers to use opportunistic encryption.  The use case is described
   in Section 1.1.  The method uses DNS-over-TLS [RFC7858] (DoT) with
   authoritative servers in an efficient manner; it is called "ADoT", as
   described in [I-D.ietf-dnsop-rfc8499bis]. (( A later version of this
   document will probably also describe the use of DNS-over-QUIC
   [I-D.ietf-dprive-dnsoquic] (DoQ). ))

   (( The following is about optional authentication; maybe will be
   removed soon )) Because opportunistic encryption means encryption
   with optional authentication, a resolver using the mechanism
   described here could achieve authenticated encryption with some
   authoritative servers, depending on how authentication for ADoT is
   defined.  To date, there has been no definition of how a resolver can

Hoffman & van Dijk       Expires 18 August 2021                 [Page 2]
Internet-Draft             Opportunistic ADoTQ             February 2021

   take advantage of DNS features that require authentication of
   authoritative servers.  If those advantages are defined in the
   future, this document would need to define the types of
   authentication for ADoT that would be allowed.

1.1.  Use Case

   The use case in this document is recursive resolver operators who are
   happy to use TLS [RFC8446] encryption with authoritative servers if
   doing so doesn't significantly slow down getting answers, and
   authoritative server operators that are happy to use encryption with
   recursive resolvers if it doesn't cost much.

   Both parties understand that using encryption costs something, but
   are willing to absorb the costs for the benefit of more Internet
   traffic being encrypted.  The extra costs (compared to using
   traditional DNS on port 53) include:

   *  Extra round trips to establish TCP for every session (but not
      necessarily for every query)

   *  Extra round trips for TLS establishment

   *  Greater CPU use for TLS establishment

   *  Greater CPU use for encryption after TLS establishment

   *  Greater memory use for holding TLS state

1.2.  Summary of Protocol

   This protocol has four main parts.  This summary gives an overview of
   how the work together.

   *  A resolver that uses this protocol has a transport cache that it
      uses to know whether to attempt using ADoT with a particular
      authoritative server, as described in Section 4.

   *  A resolver fills its transport cache by discovering whether any
      authoritative server of interest supports encrypted DNS, as
      described in Section 3.

   *  If there is no entry for that server in the cache, or the cache
      says that the authoritative server doesn't support encrypted
      transport, the resolver uses classic DNS; only if the transport
      cache indicates ADoT support, the resolver attempts to connect to
      the authoritative server with ADoT, as described in Section 2.

Hoffman & van Dijk       Expires 18 August 2021                 [Page 3]
Internet-Draft             Opportunistic ADoTQ             February 2021

   *  (( The following is about optional authentication; maybe will be
      removed soon )) If the TLS session is authenticated and the
      resolver has use for this authentication, the resolver can mark
      responses it gets as authenticated, as described in Section 5.  If
      the TLS session is not authenticated, the resolver treats the
      answers it receives as if they were received over classic DNS.

1.3.  Definitions

   The terms "recursive resolver", "authoritative server", "ADoT", and
   "classic DNS" are defined in [I-D.ietf-dnsop-rfc8499bis].

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
   "OPTIONAL" in this document are to be interpreted as described in BCP
   14 [RFC2119] [RFC8174] when, and only when, they appear in all
   capitals, as shown here.

2.  Method for Opportunistic Encryption

   [RFC7435] defines opportunistic encryption.  In this document, the
   only difference between normal TLS session establishment and
   opportunistic encryption is that the the TLS client (the recursive
   resolver) optionally authenticates the server. (( The following is
   about optional authentication; maybe will be removed soon )) See
   Section 5 for a fuller description of the use of authentication.

2.1.  Resolvers

   A resolver following this protocol uses its transport cache
   (described in Section 4) to decide whether to use classic DNS or this
   protocol to contact authoritative servers.  If the transport cache
   indicates that the authoritative server is known to support encrypted
   DNS, the resolver attempts to connect to it with ADoT on port TBD1.

   The resolver is configured with a set of timeouts that it uses when
   it is setting up ADoT.  This document does has suggested values for
   those timeouts; they are marked here with (( timeout_ )).  Resolver
   software might use these suggested values for defaults, or might
   choose their own default values.

   (( The proposed default values here are based on research that I have
   done but not published.  The research is expected to be published
   before IETF 110. ))

   The resolver MUST fall back to using classic DNS with a server if any
   of the following happens when using ADoT:

Hoffman & van Dijk       Expires 18 August 2021                 [Page 4]
Internet-Draft             Opportunistic ADoTQ             February 2021

   *  The resolver receives a TCP RST response

   *  The resolver does not receive a reply to the TCP SYN message
      within timeout "timeout_syn"; the suggested default is 1.3 seconds

   *  The resolver does not receive a reply to its first TLS message
      within timeout "timeout_tls_start"; the suggested default (which
      includes the TCP startup time) is 2.4 seconds

   *  The TLS handshake gets a definitive failure

   *  The TLS session is set up, but the resolver does not receive a
      response to its first DNS query in the TLS session within timeout
      "timeout_dns_answ"; the suggested default is 5 seconds (which
      includes the TCP and TLS startup times)

   *  The TLS session fails for reasons other than for authentication,
      such as incorrect algorithm choices or TLS record failures

   In any of those cases, the resolver needs to update its transport
   cache to indicate that the server is not currently available over
   DoT.  The time-to-live value for that entry, "timeout_ttl", could be
   as long as the TTL on the NS RRset.

   A resolver SHOULD keep a TLS session to a particular server open if
   it expects to send additional queries to that server in a short
   period of time, "timeout_hold_open".  If the server closes the TLS
   session, the resolver can re-establish a TLS session of the version
   of TLS in use allows for session resumption.

2.2.  Authoritative Servers

   An authoritative server following this protocol SHOULD support an
   ADoT service at port TBD1 for each IP address on which it offers
   service for classic DNS on port 53.

   A server MAY close a TLS connection at any time.  For example, it can
   close the TLS session if it has not received a DNS query in a defined
   length of time, "timeout_dns_query".  It can also close the TLS
   session after it sends a DNS response; however, it might also want to
   keep the TLS session open waiting for another DNS query from the
   resolver.

Hoffman & van Dijk       Expires 18 August 2021                 [Page 5]
Internet-Draft             Opportunistic ADoTQ             February 2021

3.  Discovering Whether an Authoritative Server Uses Encryption

   A recursive resolver can discover whether an authoritative server
   supports ADoT by attempting to open a TLS session to port TBD1 of an
   IP address for the server.  If the server completes the TLS
   handshake, the resolver can be fairly confident that the server
   supports ADoT.

   (( Note that there are likely better ways to do discovery.  The
   DPRIVE WG requested that this version of this draft only specify
   port-probing.  Future drafts might describe other methods, and how to
   use multiple methods at the same time for discovery, depending on
   what the WG chooses for discovery. ))

   The following are indications of failure for the ability to use ADoT
   with the server:

   *  The resolver receives a TCP RST response

   *  The resolver does not receive a reply to the TCP SYN message
      within timeout "timeout_syn"

   *  The resolver does not receive a reply to its first TLS message
      within timeout "timeout_tls_start"

   *  The TLS handshake gets a definitive failure

4.  The Transport Cache

   A recursive resolver that attempted to use encrypted transport every
   time it connected to any authoritative server would inherently be
   slower than one that did not.  Similarly, a recursive resolver that
   made an external lookup of what secure transports every authoritative
   server supports each time it connected would also likely be slower
   than one that did not. (( Proposals such as
   [I-D.vandijk-dprive-ds-dot-signal-and-pin] could be used to not cause
   extra lookups. )) Recursive resolver operators desire to give answers
   to stub resolvers as quickly as possible, so neither of these two
   strategies would make sense.

   Instead, recursive resolvers following the method described in this
   document MUST keep a cache of relevant information about how DNS-
   over-TLS is supported by authoritative servers.  This is called a
   "transport cache" in this document.  The relevant information could
   include things such as support for encryption, expected round-trip
   times, authentication mechanisms, and so on.  The transport cache is
   likely to store both positive and negative information about a
   server's ability to support encrypted DNS.

Hoffman & van Dijk       Expires 18 August 2021                 [Page 6]
Internet-Draft             Opportunistic ADoTQ             February 2021

   The recursive resolver MUST look in its transport cache before
   sending DNS queries to an authoritative server.  If there is no entry
   for an authoritative server in its transport cache, the recursive
   resolver MUST use classic DNS over port 53.  It MAY then probe for
   encrypted transports, and cache that information for later
   connections.

   This document explicitly does not mandate the contents of the
   transport cache.  Different recursive resolver implementers are
   likely to have different cache structures based on many factors, such
   as research results, active measurements, secure protocols supported,
   and customer feedback, There will likely be different strategies for
   the time-to-live for parts of the transport cache, such as how often
   to refresh the data in the cache, how often to refresh negative data,
   whether to prioritize refreshing certain zones or types of zones, and
   so on.

   This document also explicitly doesn't mandate the strategy for
   filling transport caches.  Some strategies might include one or more
   of "test NS entries from the main cache", "try to send a refresh
   query over ADoT", "use external data", "trust a third-party service
   for filling the transport cache", and so on.

   There are no interoperability issues with different implementors
   making different choices for the contents and fill strategies of
   their transport caches, and having many different options available
   will likely cause the cache designs to get better over time.

5.  Authentication

   (( The following is about optional authentication; maybe will be
   removed soon ))

   In the opportunistic encryption described here, there is no
   requirement, and no advantage, for the recursive resolver to
   authenticate the authoritative server because any certificate
   authentication failure does not prevent the TLS session from being
   set up.  If it is easier programmatically for the recursive resolver
   to authenticate the authoritative server and then ignore the negative
   result for certificate authentication, than to just not authenticate,
   the recursive resolver MAY do that.

Hoffman & van Dijk       Expires 18 August 2021                 [Page 7]
Internet-Draft             Opportunistic ADoTQ             February 2021

   This document does not describe what to do with successful
   authentication of a ADoT TLS session.  Some suggestions have been
   floated in the DPRIVE WG, but none have been written into drafts.  If
   there later are reasons to note authentication of the server,
   resolvers following this protocol MAY use that authenticated data. ((
   Change this paragraph if the WG later defines DNS-related reasons to
   authenticate. ))

   Later protocols for encrypted resolver-to-authoritative communication
   might require normal TLS authentication.  Because of this,
   authoritative servers SHOULD use TLS certificates that can be used in
   authenticated TLS communication, such as those issued by trusted
   third parties or self-issued certificates that can be authenticated
   with DANE [RFC6698] records.  However, if an authoritative server
   does not care about the use cases for such future protocols, it MAY
   use self-issued certificates that cannot be authenticated.

6.  IANA Considerations

   (( Add a registration for port TBD1 for TCP }}

7.  Security Considerations

   The method described in this document explicitly allows a resolver to
   perform DNS communications over traditional unencrypted,
   unauthenticated DNS on port 53.

   The method described in this document explicitly allows a resolver to
   choose to allow unauthenticated TLS.  In this case, the resulting
   communication will be susceptible to obvious and well-understood
   attacks from an attacker in the path of the communications.

8.  Acknowledgements

   Puneet Sood contributed many ideas to early drafts of this document.

9.  References

9.1.  Normative References

   [I-D.ietf-dnsop-rfc8499bis]
              Hoffman, P. and K. Fujiwara, "DNS Terminology", Work in
              Progress, Internet-Draft, draft-ietf-dnsop-rfc8499bis-01,
              20 November 2020, <http://www.ietf.org/internet-drafts/
              draft-ietf-dnsop-rfc8499bis-01.txt>.

Hoffman & van Dijk       Expires 18 August 2021                 [Page 8]
Internet-Draft             Opportunistic ADoTQ             February 2021

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

   [RFC7435]  Dukhovni, V., "Opportunistic Security: Some Protection
              Most of the Time", RFC 7435, DOI 10.17487/RFC7435,
              December 2014, <https://www.rfc-editor.org/info/rfc7435>.

   [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>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

   [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>.

9.2.  Informative References

   [I-D.ietf-dprive-dnsoquic]
              Huitema, C., Mankin, A., and S. Dickinson, "Specification
              of DNS over Dedicated QUIC Connections", Work in Progress,
              Internet-Draft, draft-ietf-dprive-dnsoquic-01, 20 October
              2020, <http://www.ietf.org/internet-drafts/draft-ietf-
              dprive-dnsoquic-01.txt>.

   [I-D.vandijk-dprive-ds-dot-signal-and-pin]
              Dijk, P., Geuze, R., and E. Bretelle, "Signalling
              Authoritative DoT support in DS records, with key
              pinning", Work in Progress, Internet-Draft, draft-vandijk-
              dprive-ds-dot-signal-and-pin-01, 13 July 2020,
              <http://www.ietf.org/internet-drafts/draft-vandijk-dprive-
              ds-dot-signal-and-pin-01.txt>.

   [RFC6698]  Hoffman, P. and J. Schlyter, "The DNS-Based Authentication
              of Named Entities (DANE) Transport Layer Security (TLS)
              Protocol: TLSA", RFC 6698, DOI 10.17487/RFC6698, August
              2012, <https://www.rfc-editor.org/info/rfc6698>.

Authors' Addresses

Hoffman & van Dijk       Expires 18 August 2021                 [Page 9]
Internet-Draft             Opportunistic ADoTQ             February 2021

   Paul Hoffman
   ICANN

   Email: paul.hoffman@icann.org

   Peter van Dijk
   PowerDNS

   Email: peter.van.dijk@powerdns.com

Hoffman & van Dijk       Expires 18 August 2021                [Page 10]