Internet Engineering Task Force (IETF)                       C. Holmberg
Request for Comments: 8863                                      Ericsson
Updates: 8445, 8838                                            J. Uberti
Category: Standards Track                                         Google
ISSN: 2070-1721                                             January 2021


Interactive Connectivity Establishment Patiently Awaiting Connectivity
                              (ICE PAC)

Abstract

  During the process of establishing peer-to-peer connectivity,
  Interactive Connectivity Establishment (ICE) agents can encounter
  situations where they have no candidate pairs to check, and, as a
  result, conclude that ICE processing has failed.  However, because
  additional candidate pairs can be discovered during ICE processing,
  declaring failure at this point may be premature.  This document
  discusses when these situations can occur.

  This document updates RFCs 8445 and 8838 by requiring that an ICE
  agent wait a minimum amount of time before declaring ICE failure,
  even if there are no candidate pairs left to check.

Status of This Memo

  This is an Internet Standards Track document.

  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).  Further information on
  Internet Standards is available in 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/rfc8863.

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.  Conventions
  3.  Relevant Scenarios
    3.1.  No Candidates from Peer
    3.2.  All Candidates Discarded
    3.3.  Immediate Candidate Pair Failure
  4.  Update to RFC 8445
  5.  Update to RFC 8838
  6.  Security Considerations
  7.  IANA Considerations
  8.  Normative References
  Acknowledgements
  Authors' Addresses

1.  Introduction

  [RFC8445] describes a protocol, Interactive Connectivity
  Establishment (ICE), for Network Address Translator (NAT) traversal
  for UDP-based communication.

  When using ICE, endpoints will typically exchange ICE candidates,
  form a list of candidate pairs, and then test each candidate pair to
  see if connectivity can be established.  If the test for a given pair
  fails, it is marked accordingly, and if all pairs have failed, the
  overall ICE process typically is considered to have failed.

  During the process of connectivity checks, additional candidates may
  be created as a result of successful inbound checks from the remote
  peer.  Such candidates are referred to as peer-reflexive candidates;
  once discovered, these candidates will be used to form new candidate
  pairs, which will be tested like any other.  However, there is an
  inherent problem here; if, before learning about any peer-reflexive
  candidates, an endpoint runs out of candidate pairs to check, either
  because it has none or it considers them all to have failed, it will
  prematurely declare failure and terminate ICE processing.  This
  problem can occur in many common situations.

  This specification updates [RFC8445] and [RFC8838] by simply
  requiring that an ICE agent wait a minimum amount of time before
  declaring ICE failure, even if there are no candidate pairs to check
  or all candidate pairs have failed.  This delay provides enough time
  for the discovery of peer-reflexive candidates, which may eventually
  lead to ICE processing completing successfully.

2.  Conventions

  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.

3.  Relevant Scenarios

  As noted above, the core problem this specification attempts to
  address is the situation where even after local gathering and remote
  candidate signaling have completed, the ICE agent immediately ends up
  with no valid pairs and no candidate pairs left to check, resulting
  in a premature ICE failure.  This failure is premature because not
  enough time has elapsed to allow for discovery of peer-reflexive
  candidates from inbound connectivity checks; if discovered, these
  candidates are very likely to result in a valid pair.

  In most ICE scenarios, the lengthy timeouts for connectivity check
  transactions, typically tens of seconds, will prevent this problem
  from occurring.  However, there are certain specific cases where this
  problem will frequently occur.

3.1.  No Candidates from Peer

  Per [RFC8838], an ICE agent can provide zero candidates of its own.
  If the agent somehow knows that the remote endpoint is directly
  reachable, gathering local candidates is unnecessary and will only
  cause delays; the peer agent can discover the appropriate local
  candidate via connectivity checks.

  However, following the procedures from [RFC8445] strictly will result
  in immediate ICE failure, since the checklist at the peer agent will
  be empty.

3.2.  All Candidates Discarded

  Even if the ICE agent provides candidates, they may be discarded by
  the peer agent if it does not know what to do with them.  For
  example, candidates may use an address family that the peer agent
  does not support (e.g., a host candidate with an IPv6 address in a
  NAT64 scenario) or that may not be usable for some other reason.

  In these scenarios, when the candidates are discarded, the checklist
  at the peer agent will once again be empty, leading to immediate ICE
  failure.

3.3.  Immediate Candidate Pair Failure

  Section 7.2.5.2 of [RFC8445] describes several situations in which a
  candidate pair will be considered to have failed, well before the
  connectivity check transaction timeout.

  As a result, even if the ICE agent provides usable candidates, the
  pairs created by the peer agent may fail immediately when checked,
  e.g., a check to a non-routable address that receives an immediate
  ICMP error.

  In this situation, the checklist at the peer agent may contain only
  failed pairs, resulting in immediate ICE failure.

4.  Update to RFC 8445

  In order to avoid the problem raised by this document, the ICE agent
  needs to wait enough time to allow peer-reflexive candidates to be
  discovered.  Accordingly, when a full ICE implementation begins its
  ICE processing, as described in [RFC8445], Section 6.1, it MUST set a
  timer, henceforth known as the "PAC timer" (Patiently Awaiting
  Connectivity), to ensure that ICE will run for a minimum amount of
  time before determining failure.

  Specifically, the ICE agent will start its timer once it believes ICE
  connectivity checks are starting.  This occurs when the agent has
  sent the values needed to perform connectivity checks (e.g., the
  Username Fragment and Password denoted in [RFC8445], Section 5.3) and
  has received some indication that the remote side is ready to start
  connectivity checks, typically via receipt of the values mentioned
  above.  Note that the agent will start the timer even if it has not
  sent or received any ICE candidates.

  The RECOMMENDED duration for the PAC timer is equal to the agent's
  connectivity check transaction timeout, including all
  retransmissions.  When using default values for retransmission
  timeout (RTO) and Rc, this amounts to 39.5 seconds, as explained in
  [RFC5389], Section 7.2.1.  This timeout value is chosen to roughly
  coincide with the maximum possible duration of ICE connectivity
  checks from the remote peer, which, if successful, could create peer-
  reflexive candidates.  Because the ICE agent doesn't know the exact
  number of candidate pairs and pacing interval in use by the remote
  side, this timeout value is simply a guess, albeit an educated one.
  Regardless, for this particular problem, the desired benefits will be
  realized as long as the agent waits some reasonable amount of time,
  and, as usual, the application is in the best position to determine
  what is reasonable for its scenario.

  While the timer is still running, the ICE agent MUST NOT update a
  checklist state from Running to Failed, even if there are no pairs
  left in the checklist to check.  As a result, the ICE agent will not
  remove any data streams or set the state of the ICE session to Failed
  as long as the timer is running.

  When the timer period eventually elapses, the ICE agent MUST resume
  typical ICE processing, including setting the state of any checklists
  to Failed if they have no pairs left to check and handling any
  consequences as indicated in [RFC8445], Section 8.1.2.  Naturally, if
  there are no such checklists, no action is necessary.

  One consequence of this behavior is that in cases where ICE should
  fail, e.g., where both sides provide candidates with unsupported
  address families, ICE will no longer fail immediately -- it will only
  fail when the PAC timer expires.  However, because most ICE scenarios
  require an extended period of time to determine failure, the fact
  that some specific scenarios no longer fail quickly should have
  minimal application impact, if any.

  Note also that the PAC timer is potentially relevant to the ICE
  nomination procedure described in [RFC8445], Section 8.1.1.  That
  specification does not define a minimum duration for ICE processing
  prior to nomination of a candidate pair, but in order to select the
  best candidate pair, ICE needs to run for enough time in order to
  allow peer-reflexive candidates to be discovered and checked, as
  noted above.  Accordingly, the controlling ICE agent SHOULD wait a
  sufficient amount of time before nominating candidate pairs, and it
  MAY use the PAC timer to do so.  As always, the controlling ICE agent
  retains full discretion and MAY decide, based on its own criteria, to
  nominate pairs prior to the PAC timer period elapsing.

5.  Update to RFC 8838

  Trickle ICE [RFC8838] considers a similar problem, namely whether an
  ICE agent should allow a checklist to enter the Failed state if more
  candidates might still be provided by the remote peer.  The solution,
  specified in [RFC8838], Section 8, is to wait until an end-of-
  candidates indication has been received before determining ICE
  failure.

  However, for the same reasons described above, the ICE agent may
  discover peer-reflexive candidates after it has received the end-of-
  candidates indication, and so the solution proposed by this document
  MUST still be used even when the ICE agent is using Trickle ICE.

  Note also that sending an end-of-candidates indication is only a
  SHOULD-strength requirement, which means that ICE agents will need to
  implement a backup mechanism to decide when all candidates have been
  received, typically a timer.  Accordingly, ICE agents MAY use the PAC
  timer to also serve as an end-of-candidates fallback.

6.  Security Considerations

  The security considerations for ICE are defined in [RFC8445].  This
  specification only recommends that ICE agents wait for a certain
  period of time before they declare ICE failure; it does not introduce
  new security considerations.

7.  IANA Considerations

  This document has no IANA actions.

8.  Normative References

  [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>.

  [RFC5389]  Rosenberg, J., Mahy, R., Matthews, P., and D. Wing,
             "Session Traversal Utilities for NAT (STUN)", RFC 5389,
             DOI 10.17487/RFC5389, October 2008,
             <https://www.rfc-editor.org/info/rfc5389>.

  [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>.

  [RFC8445]  Keranen, A., Holmberg, C., and J. Rosenberg, "Interactive
             Connectivity Establishment (ICE): A Protocol for Network
             Address Translator (NAT) Traversal", RFC 8445,
             DOI 10.17487/RFC8445, July 2018,
             <https://www.rfc-editor.org/info/rfc8445>.

  [RFC8838]  Ivov, E., Uberti, J., and P. Saint-Andre, "Trickle ICE:
             Incremental Provisioning of Candidates for the Interactive
             Connectivity Establishment (ICE) Protocol", RFC 8838,
             DOI 10.17487/RFC8838, January 2021,
             <https://www.rfc-editor.org/info/rfc8838>.

Acknowledgements

  Roman Shpount, Nils Ohlmeier, and Peter Thatcher provided lots of
  useful input and comments.

Authors' Addresses

  Christer Holmberg
  Ericsson
  Hirsalantie 11
  FI-02420 Jorvas
  Finland

  Email: [email protected]


  Justin Uberti
  Google
  747 6th St W
  Kirkland, WA 98033
  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.