RFC 9145




Internet Engineering Task Force (IETF)                      M. Boucadair
Request for Comments: 9145                                        Orange
Category: Standards Track                                     T. Reddy.K
ISSN: 2070-1721                                                   Akamai
                                                                 D. Wing
                                                                  Citrix
                                                           December 2021


Integrity Protection for the Network Service Header (NSH) and Encryption
                      of Sensitive Context Headers

Abstract



   This specification presents an optional method to add integrity
   protection directly to the Network Service Header (NSH) used for
   Service Function Chaining (SFC).  Also, this specification allows for
   the encryption of sensitive metadata (MD) that is carried in the NSH.

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/rfc9145.

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 Revised BSD License text as described in Section 4.e of the
   Trust Legal Provisions and are provided without warranty as described
   in the Revised BSD License.

Table of Contents



   1.  Introduction
   2.  Terminology
   3.  Assumptions and Basic Requirements
   4.  Design Overview
     4.1.  Supported Security Services
       4.1.1.  Encrypt All or a Subset of Context Headers
       4.1.2.  Integrity Protection
     4.2.  One Secret Key, Two Security Services
     4.3.  Mandatory-to-Implement Authenticated Encryption and HMAC
           Algorithms
     4.4.  Key Management
     4.5.  New NSH Variable-Length Context Headers
     4.6.  Encapsulation of NSH within NSH
   5.  New NSH Variable-Length Context Headers
     5.1.  MAC#1 Context Header
     5.2.  MAC#2 Context Header
   6.  Timestamp Format
   7.  Processing Rules
     7.1.  Generic Behavior
     7.2.  MAC NSH Data Generation
     7.3.  Encrypted NSH Metadata Generation
     7.4.  Timestamp for Replay Attack Prevention
     7.5.  NSH Data Validation
     7.6.  Decryption of NSH Metadata
   8.  MTU Considerations
   9.  Security Considerations
     9.1.  MAC#1
     9.2.  MAC#2
     9.3.  Time Synchronization
   10. IANA Considerations
   11. References
     11.1.  Normative References
     11.2.  Informative References
   Acknowledgements

   Authors' Addresses



1.  Introduction



   Many advanced Service Functions (SFs) are enabled for the delivery of
   value-added services.  Typically, SFs are used to meet various
   service objectives such as IP address sharing, avoiding covert
   channels, detecting Denial-of-Service (DoS) attacks and protecting
   network infrastructures against them, network slicing, etc.  Because
   of the proliferation of such advanced SFs together with complex
   service deployment constraints that demand more agile service
   delivery procedures, operators need to rationalize their service
   delivery logic and control its complexity while optimizing service
   activation time cycles.  The overall problem space is described in
   [RFC7498].

   [RFC7665] presents a data plane architecture addressing the
   problematic aspects of existing service deployments, including
   topological dependence and configuration complexity.  It also
   describes an architecture for the specification, creation, and
   maintenance of Service Function Chains (SFCs) within a network, that
   is, how to define an ordered set of SFs and ordering constraints that
   must be applied to packets/flows selected as a result of traffic
   classification.  [RFC8300] specifies the SFC encapsulation: Network
   Service Header (NSH).

   The NSH data is unauthenticated and unencrypted, forcing a service
   topology that requires security and privacy to use a transport
   encapsulation that supports such features (Section 8.2 of [RFC8300]).

   Note that some transport encapsulations (e.g., IPsec) only provide
   hop-by-hop security between two SFC data plane elements (e.g., two
   Service Function Forwarders (SFFs), SFF to SF) and do not provide SF-
   to-SF security of NSH metadata.  For example, if IPsec is used, SFFs
   or SFs within a Service Function Path (SFP) that are not authorized
   to access the sensitive metadata (e.g., privacy-sensitive
   information) will have access to the metadata.  As a reminder, the
   metadata referred to is information that is inserted by Classifiers
   or intermediate SFs and shared with downstream SFs; such information
   is not visible to the communication endpoints (Section 4.9 of
   [RFC7665]).

   The lack of such capability was reported during the development of
   [RFC8300] and [RFC8459].  The reader may refer to Section 3.2.1 of
   [INTERNET-THREAT-MODEL] for a discussion on the need for more
   awareness about attacks from within closed domains.

   This specification fills that gap for SFC (that is, it defines the
   "NSH Variable Header-Based Integrity" option mentioned in
   Section 8.2.1 of [RFC8300]).  Concretely, this document adds
   integrity protection and optional encryption of sensitive metadata
   directly to the NSH (Section 4).  The integrity protection covers the
   packet payload and provides replay protection (Section 7.4).  Thus,
   the NSH does not have to rely upon an underlying transport
   encapsulation for security.

   This specification introduces new Variable-Length Context Headers to
   carry fields necessary for integrity-protected NSH headers and
   encrypted Context Headers (Section 5).  This specification is only
   applicable to NSH MD Type 0x02 (Section 2.5 of [RFC8300]).  MTU
   considerations are discussed in Section 8.  This specification is not
   applicable to NSH MD Type 0x01 (Section 2.4 of [RFC8300]) because
   that MD Type only allows a Fixed-Length Context Header whose size is
   16 bytes; that is not sufficient to accommodate both the metadata and
   message integrity of the NSH data.

   This specification limits access to NSH-supplied information along an
   SFP to entities that have a need to interpret it.

   The mechanism specified in this document does not preclude the use of
   transport security.  Such considerations are deployment specific.

   It is out of the scope of this document to specify an NSH-aware
   control plane solution.

2.  Terminology



   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.

   This document makes use of the terms defined in [RFC7665] and
   [RFC8300].  The term "transport encapsulation" used in this document
   refers to the outer encapsulation (e.g., Generic Routing
   Encapsulation (GRE), IPsec, and Generic Protocol Extension for
   Virtual eXtensible Local Area Network (VXLAN-GPE)) that is used to
   carry NSH-encapsulated packets as per Section 4 of [RFC8300].

   The document defines the following terms:

   SFC data plane element:  Refers to NSH-aware SF, SFF, the SFC Proxy,
      or the Classifier as defined in the SFC data plane architecture
      [RFC7665] and further refined in [RFC8300].

   SFC control element:  Is a logical entity that instructs one or more
      SFC data plane elements on how to process NSH packets within an
      SFC-enabled domain.

   Key Identifier:  Is used to identify keys to authorized entities.
      See, for example, "kid" usage in [RFC7635].

   NSH data:  The NSH is composed of a Base Header, a Service Path
      Header, and optional Context Headers.  NSH data refers to all the
      above headers and the packet or frame on which the NSH is imposed
      to realize an SFP.

   NSH imposer:  Refers to an SFC data plane element that is entitled to
      impose the NSH with the Context Headers defined in this document.

3.  Assumptions and Basic Requirements



   Section 2 of [RFC8300] specifies that the NSH data can be spread over
   three headers:

   Base Header:  Provides information about the service header and the
      payload protocol.

   Service Path Header:  Provides path identification and location
      within an SFP.

   Context Header(s):  Carries metadata (i.e., context data) along a
      service path.

   The NSH allows sharing context information (a.k.a. metadata) with
   downstream NSH-aware data plane elements on a per-SFC/SFP basis.  To
   that aim:

   *  The Classifier is instructed by an SFC control element about the
      set of context information to be supplied for a given service
      function chain.

   *  An NSH-aware SF is instructed by an SFC control element about any
      metadata it needs to attach to packets for a given service
      function chain.  This instruction may occur any time during the
      validity lifetime of an SFC/SFP.  For a given service function
      chain, the NSH-aware SF is also provided with an order for
      consuming a set of contexts supplied in a packet.

   *  An NSH-aware SF can also be instructed by an SFC control element
      about the behavior it should adopt after consuming context
      information that was supplied in the NSH.  For example, the
      context information can be maintained, updated, or stripped.

   *  An SFC Proxy may be instructed by an SFC control element about the
      behavior it should adopt to process the context information that
      was supplied in the NSH on behalf of an NSH-unaware SF (e.g., the
      context information can be maintained or stripped).  The SFC Proxy
      may also be instructed to add some new context information into
      the NSH on behalf of an NSH-unaware SF.

   In reference to Table 1:

   *  Classifiers, NSH-aware SFs, and SFC proxies are entitled to update
      the Context Header(s).

   *  Only NSH-aware SFs and SFC proxies are entitled to update the
      Service Path Header.

   *  SFFs are entitled to modify the Base Header (TTL value, for
      example).  Nevertheless, SFFs are not supposed to act on the
      Context Headers or look into the content of the Context Headers
      (Section 4.3 of [RFC7665]).

   Thus, the following requirements:

   *  Only Classifiers, NSH-aware SFs, and SFC proxies must be able to
      encrypt and decrypt a given Context Header.

   *  Both encrypted and unencrypted Context Headers may be included in
      the same NSH.

   *  The solution must provide integrity protection for the Service
      Path Header.

   *  The solution must provide optional integrity protection for the
      Base Header.  The implications of disabling such checks are
      discussed in Section 9.1.

    +=============+===========================+=======================+
    |   SFC Data  |     Insert, remove, or    |     Update the NSH    |
    |    Plane    |      replace the NSH      |                       |
    |   Element   +========+========+=========+===========+===========+
    |             | Insert | Remove | Replace | Decrement |   Update  |
    |             |        |        |         |  Service  |  Context  |
    |             |        |        |         |   Index   | Header(s) |
    +=============+========+========+=========+===========+===========+
    | Classifier  |   +    |        |    +    |           |     +     |
    +-------------+--------+--------+---------+-----------+-----------+
    | Service     |        |   +    |         |           |           |
    | Function    |        |        |         |           |           |
    | Forwarder   |        |        |         |           |           |
    | (SFF)       |        |        |         |           |           |
    +-------------+--------+--------+---------+-----------+-----------+
    | Service     |        |        |         |     +     |     +     |
    | Function    |        |        |         |           |           |
    | (SF)        |        |        |         |           |           |
    +-------------+--------+--------+---------+-----------+-----------+
    | Service     |   +    |   +    |         |     +     |     +     |
    | Function    |        |        |         |           |           |
    | Chaining    |        |        |         |           |           |
    | (SFC) Proxy |        |        |         |           |           |
    +-------------+--------+--------+---------+-----------+-----------+

                      Table 1: Summary of NSH Actions

4.  Design Overview



4.1.  Supported Security Services



   This specification provides the functions described in the following
   subsections.

4.1.1.  Encrypt All or a Subset of Context Headers



   The solution allows encrypting all or a subset of NSH Context Headers
   by Classifiers, NSH-aware SFs, and SFC proxies.

   As depicted in Table 2, SFFs are not involved in data encryption.

      +====================+=======================+================+
      | Data Plane Element | Base and Service Path | Context Header |
      |                    | Headers Encryption    | Encryption     |
      +====================+=======================+================+
      | Classifier         | No                    | Yes            |
      +--------------------+-----------------------+----------------+
      | SFF                | No                    | No             |
      +--------------------+-----------------------+----------------+
      | NSH-aware SF       | No                    | Yes            |
      +--------------------+-----------------------+----------------+
      | SFC Proxy          | No                    | Yes            |
      +--------------------+-----------------------+----------------+
      | NSH-unaware SF     | No                    | No             |
      +--------------------+-----------------------+----------------+

          Table 2: Encryption Function Supported by SFC Data Plane
                                  Elements

   Classifier(s), NSH-aware SFs, and SFC proxies are instructed with the
   set of Context Headers (privacy-sensitive metadata, typically) that
   must be encrypted.  Encryption keying material is only provided to
   these SFC data plane elements.

   The control plane may indicate the set of SFC data plane elements
   that are entitled to supply a given Context Header (e.g., in
   reference to their identifiers as assigned within the SFC-enabled
   domain).  It is out of the scope of this document to elaborate on how
   such instructions are provided to the appropriate SFC data plane
   elements nor to detail the structure used to store the instructions.

   The Service Path Header (Section 2 of [RFC8300]) is not encrypted
   because SFFs use the Service Index (SI) in conjunction with the
   Service Path Identifier (SPI) for determining the next SF in the
   path.

4.1.2.  Integrity Protection



   The solution provides integrity protection for the NSH data.  Two
   levels of assurance (LoAs) are supported.

   The first level of assurance is where all NSH data except the Base
   Header are integrity protected (Figure 1).  In this case, the NSH
   imposer may be a Classifier, an NSH-aware SF, or an SFC Proxy.  SFFs
   are not provided with authentication material.  Further details are
   discussed in Section 5.1.

         +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
         |                Transport Encapsulation                |
         +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+...
         |                Base Header                            |  |
      +->+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+  N
      |  |                Service Path Header                    |  S
      |  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+  H
      |  |                Context Header(s)                      |  |
      |  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+...
      |  |                Original Packet                        |
      +->+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |
      +------Scope of integrity-protected data

                     Figure 1: First Level of Assurance

   The second level of assurance is where all NSH data, including the
   Base Header, are integrity protected (Figure 2).  In this case, the
   NSH imposer may be a Classifier, an NSH-aware SF, an SFF, or an SFC
   Proxy.  Further details are provided in Section 5.2.

         +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
         |                Transport Encapsulation                |
      +->+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+...
      |  |                Base Header                            |  |
      |  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+  N
      |  |                Service Path Header                    |  S
      |  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+  H
      |  |                Context Header(s)                      |  |
      |  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+...
      |  |                Original Packet                        |
      +->+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |
      +----Scope of integrity-protected data

                    Figure 2: Second Level of Assurance

   The integrity-protection scope is explicitly signaled to NSH-aware
   SFs, SFFs, and SFC proxies in the NSH by means of a dedicated MD Type
   (Section 5).

   In both levels of assurance, the Context Headers and the packet on
   which the NSH is imposed are subject to integrity protection.

   Table 3 classifies the data plane elements as being involved or not
   involved in providing integrity protection for the NSH.

         +====================+==================================+
         | Data Plane Element | Integrity Protection             |
         +====================+==================================+
         | Classifier         | Yes                              |
         +--------------------+----------------------------------+
         | SFF                | No (first LoA); Yes (second LoA) |
         +--------------------+----------------------------------+
         | NSH-aware SF       | Yes                              |
         +--------------------+----------------------------------+
         | SFC Proxy          | Yes                              |
         +--------------------+----------------------------------+
         | NSH-unaware SF     | No                               |
         +--------------------+----------------------------------+

            Table 3: Integrity Protection Supported by SFC Data
                               Plane Elements

4.2.  One Secret Key, Two Security Services



   The Authenticated Encryption with Associated Data (AEAD) interface
   defined in Section 5 of [RFC5116] is used to encrypt the Context
   Headers that carry sensitive metadata and to provide integrity
   protection for the encrypted Context Headers.

   The secondary keys "MAC_KEY" and "ENC_KEY" are generated from the
   input secret key (K) as follows; each of these two keys is an octet
   string:

   MAC_KEY:  Consists of the initial MAC_KEY_LEN octets of K, in order.
      The MAC_KEY is used for calculating the message integrity of the
      NSH data.  In other words, the integrity protection provided by
      the cryptographic mechanism is extended to also provide protection
      for the unencrypted Context Headers and the packet on which the
      NSH is imposed.

   ENC_KEY:  Consists of the final ENC_KEY_LEN octets of K, in order.
      The ENC_KEY is used as the symmetric encryption key for encrypting
      the Context Headers.

   The Hashed Message Authentication Code (HMAC) algorithm discussed in
   [RFC4868] is used to protect the integrity of the NSH data.  The
   algorithm for implementing and validating HMACs is provided in
   [RFC2104].

   The advantage of using both AEAD and HMAC algorithms (instead of just
   AEAD) is that NSH-aware SFs and SFC proxies only need to recompute
   the message integrity of the NSH data after decrementing the SI and
   do not have to recompute the ciphertext.  The other advantage is that
   SFFs do not have access to the ENC_KEY and cannot act on the
   encrypted Context Headers, and (in the case of the second level of
   assurance) SFFs do have access to the MAC_KEY.  Similarly, an NSH-
   aware SF or SFC Proxy not allowed to decrypt the Context Headers will
   not have access to the ENC_KEY.

   The authenticated encryption algorithm or HMAC algorithm to be used
   by SFC data plane elements is typically controlled using the SFC
   control plane.  Mandatory-to-implement authenticated encryption and
   HMAC algorithms are listed in Section 4.3.

   The authenticated encryption process takes four inputs, each of which
   is an octet string: a secret key (ENC_KEY, referred to as "K" in
   [RFC5116]), a plaintext (P), associated data (A) (which contains the
   data to be authenticated but not encrypted), and a nonce (N).  A
   ciphertext (C) is generated as an output as discussed in Section 2.1
   of [RFC5116].

   In order to decrypt and verify, the cipher takes ENC_KEY, N, A, and C
   as input.  The output is either the plaintext or an error indicating
   that the decryption failed as described in Section 2.2 of [RFC5116].

4.3.  Mandatory-to-Implement Authenticated Encryption and HMAC
      Algorithms



   Classifiers, NSH-aware SFs, and SFC proxies MUST implement the
   AES_128_GCM [GCM][RFC5116] algorithm and SHOULD implement the
   AES_192_GCM and AES_256_GCM algorithms.

   Classifiers, NSH-aware SFs, and SFC proxies MUST implement the HMAC-
   SHA-256-128 algorithm and SHOULD implement the HMAC-SHA-384-192 and
   HMAC-SHA-512-256 algorithms.

   SFFs MAY implement the aforementioned cipher suites and HMAC
   algorithms.

      Note: The use of the AES_128_CBC_HMAC_SHA_256 algorithm would have
      avoided the need for a second 128-bit authentication tag, but due
      to the nature of how Cipher Block Chaining (CBC) mode operates,
      AES_128_CBC_HMAC_SHA_256 allows for the malleability of
      plaintexts.  This malleability allows for attackers that know the
      Message Authentication Code (MAC) key but not the encryption key
      to make changes in the ciphertext and, if parts of the plaintext
      are known, create arbitrary blocks of plaintext.  This
      specification mandates the use of separate AEAD and MAC
      protections to prevent this type of attack.

4.4.  Key Management



   The procedure for the allocation/provisioning of secret keys (K) and
   the authenticated encryption algorithm or MAC_KEY and HMAC algorithm
   is outside the scope of this specification.  As such, this
   specification does not mandate the support of any specific mechanism.

   The document does not assume nor preclude the following:

   *  The same keying material is used for all the service functions
      used within an SFC-enabled domain.

   *  Distinct keying material is used per SFP by all involved SFC data
      path elements.

   *  Per-tenant keys are used.

   In order to accommodate deployments relying upon keying material per
   SFC/SFP and also the need to update keys after encrypting NSH data
   for a certain amount of time, this document uses key identifiers to
   unambiguously identify the appropriate keying material and associated
   algorithms for MAC and encryption.  This use of in-band identifiers
   addresses the problem of the synchronization of keying material.

   Additional information on manual vs. automated key management and
   when one should be used over the other can be found in [RFC4107].

   The risk involved in using a group-shared symmetric key increases
   with the size of the group to which it is shared.  Additional
   security issues are discussed in Section 9.

4.5.  New NSH Variable-Length Context Headers



   New NSH Variable-Length Context Headers are defined in Section 5 for
   NSH data integrity protection and, optionally, encryption of Context
   Headers carrying sensitive metadata.  Concretely, an NSH imposer
   includes (1) the key identifier to identify the keying material, (2)
   the timestamp to protect against replay attacks (Section 7.4), and
   (3) MAC for the target NSH data (depending on the integrity-
   protection scope) calculated using MAC_KEY and, optionally, Context
   Headers encrypted using ENC_KEY.

   An SFC data plane element that needs to check the integrity of the
   NSH data uses the MAC_KEY and HMAC algorithm for the key identifier
   being carried in the NSH.

   An NSH-aware SF or SFC Proxy that needs to decrypt some Context
   Headers uses ENC_KEY and the decryption algorithm for the key
   identifier being carried in the NSH.

   Section 7 specifies the detailed procedure.

4.6.  Encapsulation of NSH within NSH



   As discussed in Section 3 of [RFC8459], an SFC-enabled domain (called
   an upper-level domain) may be decomposed into many sub-domains
   (called lower-level domains).  In order to avoid maintaining state to
   restore upper-level NSH information at the boundaries of lower-level
   domains, two NSH levels are used: an Upper-NSH that is imposed at the
   boundaries of the upper-level domain and a Lower-NSH that is pushed
   by the Classifier of a lower-level domain in front of the original
   NSH (Figure 3).  As such, the Upper-NSH information is carried along
   the lower-level chain without modification.  The packet is forwarded
   in the top-level domain according to the Upper-NSH, while it is
   forwarded according to the Lower-NSH in a lower-level domain.

                     +---------------------------------+
                     |     Transport Encapsulation     |
                  +->+---------------------------------+
                  |  |        Lower-NSH Header         |
                  |  +---------------------------------+
                  |  |        Upper-NSH Header         |
                  |  +---------------------------------+
                  |  |          Original Packet        |
                  +->+---------------------------------+
                  |
                  |
                  +----Scope of NSH security protection
                       provided by a lower-level domain

                 Figure 3: Encapsulation of NSH within NSH

   SFC data plane elements of a lower-level domain include the Upper-NSH
   when computing the MAC.

   Keying material used at the upper-level domain SHOULD NOT be the same
   as the one used by a lower-level domain.

5.  New NSH Variable-Length Context Headers



   This section specifies the format of new Variable-Length Context
   Headers that are used for NSH integrity protection and, optionally,
   Context Header encryption.

   In particular, this section defines two "MAC and Encrypted Metadata"
   Context Headers, each having specific deployment constraints.  Unlike
   Section 5.1, the level of assurance provided in Section 5.2 requires
   sharing MAC_KEY with SFFs.  Both Context Headers have the same format
   as shown in Figure 4.

       0                   1                   2                   3
       0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |          Metadata Class       |      Type     |U|    Length   |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      | Key Id Len  |         Key Identifier (Variable)               ~
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      ~                      Timestamp (8 bytes)                      ~
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      | Nonce Length  |           Nonce  (Variable)                   ~
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |     Message Authentication Code and optional Encrypted        |
      ~                  Context Headers (Variable)                   ~
      |                                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

            Figure 4: MAC and Encrypted Metadata Context Header

   The "MAC and Encrypted Metadata" Context Headers are padded out to a
   multiple of 4 bytes as per Section 2.2 of [RFC8300].  The "MAC and
   Encrypted Metadata" Context Header, if included, MUST always be the
   last Context Header.

5.1.  MAC#1 Context Header



   The MAC#1 Context Header is a Variable-Length Context Header that
   carries MAC for the Service Path Header, Context Headers, and the
   inner packet on which NSH is imposed, calculated using MAC_KEY and,
   optionally, Context Headers encrypted using ENC_KEY.  The scope of
   the integrity protection provided by this Context Header is depicted
   in Figure 5.

   This MAC scheme does not require sharing MAC_KEY with SFFs.  It does
   not require recomputing the MAC by each SFF because of TTL
   processing.  Section 9.1 discusses the possible threat associated
   with this level of assurance.

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |Ver|O|U|    TTL    |   Length  |U|U|U|U|MD Type| Next Protocol |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<--+
   |          Service Path Identifier              | Service Index |   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   ~       Variable-Length Unencrypted Context Headers  (opt.)     ~   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   |          Metadata Class       |      Type     |U|    Length   |   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   | Key Id Len  |         Key Identifier (Variable)               ~   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   ~                      Timestamp (8 bytes)                      ~   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   | Nonce Length  |           Nonce  (Variable)                   ~   |
+->+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
|  ~                Encrypted Context Headers (opt.)               ~   |
+->+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
|  ~                 Message Authentication Code                   ~   |
|  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
|  |                                                               |   |
|  ~               Inner Packet on which NSH is imposed            ~   |
|  |                                                               |   |
|  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<--|
|                                                                      |
|                                       Integrity-Protection Scope ----+
+----Encrypted Data

                       Figure 5: Scope of MAC#1

   In reference to Figure 4, the description of the fields is as
   follows:

   Metadata Class:  MUST be set to 0x0 (Section 2.2 of [RFC8300]).

   Type:       0x02 (see Section 10).

   U:          Unassigned bit (Section 2.5.1 of [RFC8300]).



   Length:     Indicates the length of the variable-length metadata in
               bytes.  Padding considerations are discussed in
               Section 2.5.1 of [RFC8300].

   Key Id Len:  Variable.  Carries the length of the key identifier in
               octets.

   Key Identifier:  Carries a variable-length Key Identifier object used
               to identify and deliver keys to SFC data plane elements.
               This identifier is helpful for accommodating deployments
               relying upon keying material per SFC/SFP.  The key
               identifier helps to resolve the problem of
               synchronization of keying material.  A single key
               identifier is used to look up both the ENC_KEY and the
               MAC_KEY associated with a key, and the corresponding
               encryption and MAC algorithms used with those keys.

   Timestamp:  Refer to Section 6 for more details about the structure
               of this field.

   Nonce Length:  Carries the length of the Nonce.  If the Context
               Headers are only integrity protected, "Nonce Length" is
               set to zero (that is, no "Nonce" is included).

   Nonce:      Carries the Nonce for the authenticated encryption
               operation (Section 3.1 of [RFC5116]).

   Encrypted Context Headers:  Carries the optional encrypted Context
               Headers.

   Message Authentication Code:  Covers the entire NSH data, excluding
               the Base Header.

5.2.  MAC#2 Context Header



   The MAC#2 Context Header is a Variable-Length Context Header that
   carries the MAC for the entire NSH data calculated using MAC_KEY and,
   optionally, Context Headers encrypted using ENC_KEY.  The scope of
   the integrity protection provided by this Context Header is depicted
   in Figure 6.

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<--+
   |Ver|O|U|    TTL    |   Length  |U|U|U|U|MD Type| Next Protocol |   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   |          Service Path Identifier              | Service Index |   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   ~       Variable-Length Unencrypted Context Headers  (opt.)     ~   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   |          Metadata Class       |      Type     |U|    Length   |   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   | Key Id Len  |         Key Identifier (Variable)               ~   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   ~                      Timestamp (8 bytes)                      ~   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
   | Nonce Length  |           Nonce  (Variable)                   ~   |
+->+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
|  ~                Encrypted Context Headers (opt.)               ~   |
+->+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
|  ~                 Message Authentication Code                   ~   |
|  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+   |
|  |                                                               |   |
|  ~               Inner Packet on which NSH is imposed            ~   |
|  |                                                               |   |
|  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<--|
|                                                                      |
|                                       Integrity-Protection Scope ----+
+----Encrypted Data

                       Figure 6: Scope of MAC#2

   In reference to Figure 4, the description of the fields is as
   follows:

   Metadata Class:  MUST be set to 0x0 (Section 2.2 of [RFC8300]).

   Type:       0x03 (see Section 10).

   U:          Unassigned bit (Section 2.5.1 of [RFC8300]).

   Length:     Indicates the length of the variable-length metadata in
               bytes.  Padding considerations are discussed in
               Section 2.5.1 of [RFC8300].

   Key Id Len:  See Section 5.1.

   Key Identifier:  See Section 5.1.

   Timestamp:  See Section 6.

   Nonce Length:  See Section 5.1.

   Nonce:      See Section 5.1.

   Encrypted Context Headers:  Carries the optional encrypted Context
               Headers.

   Message Authentication Code:  Covers the entire NSH data.

6.  Timestamp Format



   This section follows the template provided in Section 3 of [RFC8877].

   The format of the Timestamp field introduced in Section 5 is depicted
   in Figure 7.

       0                   1                   2                   3
       0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                            Seconds                            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                           Fraction                            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                      Figure 7: Timestamp Field Format

   Timestamp field format:
      Seconds:  Specifies the integer portion of the number of seconds
         since the epoch.

      + Size:  32 bits

      + Units:  Seconds

      Fraction:  Specifies the fractional portion of the number of
         seconds since the epoch.

      + Size:  32 bits

      + Units:  The unit is 2^(-32) seconds, which is roughly equal to
         233 picoseconds.

   Epoch:
      The epoch is 1970-01-01T00:00 in UTC time.  Note that this epoch
      value is different from the one used in Section 6 of [RFC5905]
      (which will wrap around in 2036).

   Leap seconds:
      This timestamp format is affected by leap seconds.  The timestamp
      represents the number of seconds elapsed since the epoch minus the
      number of leap seconds.

   Resolution:
      The resolution is 2^(-32) seconds.

   Wraparound:
      This time format wraps around every 2^32 seconds, which is roughly
      136 years.  The next wraparound will occur in the year 2106.

   Synchronization aspects:
      It is assumed that SFC data plane elements are synchronized to UTC
      using a synchronization mechanism that is outside the scope of
      this document.  In typical deployments, SFC data plane elements
      use NTP [RFC5905] for synchronization.  Thus, the timestamp may be
      derived from the NTP-synchronized clock, allowing the timestamp to
      be measured with respect to the clock of an NTP server.  Since
      this time format is specified in terms of UTC, it is affected by
      leap seconds (in a manner analogous to the NTP time format, which
      is similar).  Therefore, the value of a timestamp during or
      slightly after a leap second may be temporarily inaccurate.

7.  Processing Rules



   The following subsections describe the processing rules for
   integrity-protected NSH and, optionally, encrypted Context Headers.

7.1.  Generic Behavior



   This document adheres to the recommendations in Section 8.1 of
   [RFC8300] for handling the Context Headers at both ingress and egress
   SFC boundary nodes (i.e., to strip the entire NSH, including Context
   Headers).

   Failures of a Classifier to inject the Context Headers defined in
   this document SHOULD be logged locally while a notification alarm MAY
   be sent to an SFC control element.  Failures of an NSH-aware node to
   validate the integrity of the NSH data MUST cause that packet to be
   discarded while a notification alarm MAY be sent to an SFC control
   element.  The details of sending notification alarms (i.e., the
   parameters that affect the transmission of the notification alarms
   depending on the information in the Context Header such as frequency,
   thresholds, and content in the alarm) SHOULD be configurable.

   NSH-aware SFs and SFC proxies MAY be instructed to strip some
   encrypted Context Headers from the packet or to pass the data to the
   next SF in the service function chain after processing the content of
   the Context Headers.  If no instruction is provided, the default
   behavior for intermediary NSH-aware nodes is to maintain such Context
   Headers so that the information can be passed to the next NSH-aware
   hops.  NSH-aware SFs and SFC proxies MUST reapply the integrity
   protection if any modification is made to the Context Headers (e.g.,
   strip a Context Header, update the content of an existing Context
   Header, insert a new Context Header).

   An NSH-aware SF or SFC Proxy that is not allowed to decrypt any
   Context Headers MUST NOT be given access to the ENC_KEY.

   Otherwise, an NSH-aware SF or SFC Proxy that receives encrypted
   Context Headers, for which it is not allowed to consume a specific
   Context Header it decrypts (but consumes others), MUST keep that
   Context Header unaltered when forwarding the packet upstream.

   Only one instance of a "MAC and Encrypted Metadata" Context Header
   (Section 5) is allowed in an NSH level.  If multiple instances of a
   "MAC and Encrypted Metadata" Context Header are included in an NSH
   level, the SFC data plane element MUST process the first instance and
   ignore subsequent instances and MAY log or increase a counter for
   this event as per Section 2.5.1 of [RFC8300].  If NSH within NSH is
   used (Section 4.6), distinct LoAs may be used for each NSH level.

   MTU and fragmentation considerations are discussed in Section 8.

7.2.  MAC NSH Data Generation



   After performing any Context Header encryption, the HMAC algorithm
   discussed in [RFC4868] is used to integrity protect the target NSH
   data.  An NSH imposer inserts a "MAC and Encrypted Metadata" Context
   Header for integrity protection (Section 5).

   The NSH imposer sets the MAC field to zero and then computes the
   message integrity for the target NSH data (depending on the
   integrity-protection scope discussed in Section 5) using the MAC_KEY
   and HMAC algorithm.  It inserts the computed digest in the MAC field
   of the "MAC and Encrypted Metadata" Context Header.  The length of
   the MAC is decided by the HMAC algorithm adopted for the particular
   key identifier.

   The Message Authentication Code (T) computation process for the
   target NSH data with HMAC-SHA-256-128() can be illustrated as
   follows:

         T = HMAC-SHA-256-128(MAC_KEY, target NSH data)

   An entity in the SFP that updates the NSH MUST follow the above
   behavior to maintain message integrity of the NSH for subsequent
   validations.

7.3.  Encrypted NSH Metadata Generation



   An NSH imposer can encrypt Context Headers carrying sensitive
   metadata, i.e., encrypted and unencrypted metadata may be carried
   simultaneously in the same NSH packet (Sections 5 and 6).

   In order to prevent pervasive monitoring [RFC7258], it is RECOMMENDED
   to encrypt all Context Headers.  All Context Headers carrying
   privacy-sensitive metadata MUST be encrypted; by doing so, privacy-
   sensitive metadata is not revealed to attackers.  Privacy-specific
   threats are discussed in Section 5.2 of [RFC6973].

   Using the secret key (ENC_KEY) and authenticated encryption
   algorithm, the NSH imposer encrypts the Context Headers (as set, for
   example, in Section 3) and inserts the resulting payload in the "MAC
   and Encrypted Metadata" Context Header (Section 5).  The additional
   authenticated data input to the AEAD function is a zero-length byte
   string.  The entire Context Header carrying sensitive metadata is
   encrypted (that is, including the MD Class, Type, Length, and
   associated metadata of each Context Header).

   More details about the exact encryption procedure are provided in
   Section 2.1 of [RFC5116].  In this case, the associated data (A)
   input is zero length for AES Galois/Counter Mode (AES-GCM).

   An authorized entity in the SFP that updates the content of an
   encrypted Context Header or needs to add a new encrypted Context
   Header MUST also follow the aforementioned behavior.

7.4.  Timestamp for Replay Attack Prevention



   The Timestamp imposed by an initial Classifier is left untouched
   along an SFP.  However, it can be updated when reclassification
   occurs (Section 4.8 of [RFC7665]).  The same considerations for
   setting the Timestamp are followed in both initial classification and
   reclassification (Section 6).

   The received NSH is accepted by an NSH-aware node if the Timestamp
   (TS) in the NSH is recent enough to the reception time of the NSH
   (TSrt).  The following formula is used for this check:

         -Delta < (TSrt - TS) < +Delta

   The Delta interval is a configurable parameter.  The default value
   for the allowed Delta is 2 seconds.  Special care should be taken
   when setting very low Delta values as this may lead to dropping
   legitimate traffic.  If the timestamp is not within the boundaries,
   then the SFC data plane element receiving such packets MUST discard
   the NSH message.

   Replay attacks within the Delta window may be detected by an NSH-
   aware node by recording a unique value derived from the packet, such
   as a unique value from the MAC field value.  Such an NSH-aware node
   will detect and reject duplicates.  If for legitimate service reasons
   some flows have to be duplicated but still share a portion of an SFP
   with the original flow, legitimate duplicate packets will be tagged
   by NSH-aware nodes involved in that segment as replay packets unless
   sufficient entropy is added to the duplicate packet.  How such an
   entropy is added is implementation specific.

      Note: Within the timestamp Delta window, defining a sequence
      number to protect against replay attacks may be considered.  In
      such a mode, NSH-aware nodes must discard packets with duplicate
      sequence numbers within the timestamp Delta window.  However, in
      deployments with several instances of the same SF (e.g., cluster
      or load-balanced SFs), a mechanism to coordinate among those
      instances to discard duplicate sequence numbers is required.
      Because the coordination mechanism to comply with this requirement
      is service specific, this document does not include this
      protection.

   All SFC data plane elements must be synchronized among themselves.
   These elements may be synchronized to a global reference time.

7.5.  NSH Data Validation



   When an SFC data plane element that conforms to this specification
   needs to check the validity of the NSH data, it MUST ensure that a
   "MAC and Encrypted Metadata" Context Header is included in a received
   NSH packet.  The imposer MUST silently discard the packet and MUST
   log an error at least once per the SPI if at least one of the
   following is observed:

   *  the "MAC and Encrypted Metadata" Context Header is missing,

   *  the enclosed key identifier is unknown or invalid (e.g., the
      corresponding key expired), or

   *  the timestamp is invalid (Section 7.4).

   If the timestamp check is successfully passed, the SFC data plane
   element proceeds with NSH data integrity validation.  After storing
   the value of the MAC field in the "MAC and Encrypted Metadata"
   Context Header, the SFC data plane element fills the MAC field with
   zeros.  Then, the SFC data plane element generates the message
   integrity for the target NSH data (depending on the integrity-
   protection scope discussed in Section 5) using the MAC_KEY and HMAC
   algorithm.  If the value of the newly generated digest is identical
   to the stored one, the SFC data plane element is certain that the NSH
   data has not been tampered with and validation is therefore
   successful.  Otherwise, the NSH packet MUST be discarded.  The
   comparison of the computed HMAC value to the stored value MUST be
   done in a constant-time manner to thwart timing attacks.

7.6.  Decryption of NSH Metadata



   If entitled to consume a supplied encrypted Context Header, an NSH-
   aware SF or SFC Proxy decrypts metadata using (K) and a decryption
   algorithm for the key identifier in the NSH.

   The authenticated encryption algorithm has only a single output,
   either a plaintext or a special symbol (FAIL) that indicates that the
   inputs are not authentic (Section 2.2 of [RFC5116]).

8.  MTU Considerations



   The SFC architecture prescribes that additional information be added
   to packets to:

   *  Identify SFPs: this is typically the NSH Base Header and Service
      Path Header.

   *  Carry metadata such as that defined in Section 5.

   *  Steer the traffic along the SFPs: This is realized by means of
      transport encapsulation.

   This added information increases the size of the packet to be carried
   along an SFP.

   Aligned with Section 5 of [RFC8300], it is RECOMMENDED that network
   operators increase the underlying MTU so that NSH traffic is
   forwarded within an SFC-enabled domain without fragmentation.  The
   available underlying MTU should be taken into account by network
   operators when providing SFs with the required Context Headers to be
   injected per SFP and the size of the data to be carried in these
   Context Headers.

   If the underlying MTU cannot be increased to accommodate the NSH
   overhead, network operators may rely upon a transport encapsulation
   protocol with the required fragmentation handling.  The impact of
   activating such features on SFFs should be carefully assessed by
   network operators (Section 5.6 of [RFC7665]).

   When dealing with MTU issues, network operators should consider the
   limitations of various tunnel mechanisms such as those discussed in
   [INTERNET-IP-TUNNELS].

9.  Security Considerations



   Data plane SFC-related security considerations, including privacy,
   are discussed in Section 6 of [RFC7665] and Section 8 of [RFC8300].
   In particular, Section 8.2.2 of [RFC8300] states that attached
   metadata (i.e., Context Headers) should be limited to that which is
   necessary for correct operation of the SFP.  Also, that section
   indicates that [RFC8165] discusses metadata considerations that
   operators can take into account when using NSH.

   The guidelines for cryptographic key management are discussed in
   [RFC4107].  The group key management protocol-related security
   considerations discussed in Section 8 of [RFC4046] need to be taken
   into consideration.

   The interaction between the SFC data plane elements and a key
   management system MUST NOT be transmitted unencrypted since this
   would completely destroy the security benefits of the integrity-
   protection solution defined in this document.

   The secret key (K) must have an expiration time assigned as the
   latest point in time before which the key may be used for integrity
   protection of NSH data and encryption of Context Headers.  Prior to
   the expiration of the secret key, all participating NSH-aware nodes
   SHOULD have the control plane distribute a new key identifier and
   associated keying material so that when the secret key is expired,
   those nodes are prepared with the new secret key.  This allows the
   NSH imposer to switch to the new key identifier as soon as necessary.
   It is RECOMMENDED that the next key identifier and associated keying
   material be distributed by the control plane well prior to the secret
   key expiration time.  Additional guidance for users of AEAD functions
   about rekeying can be found in [AEAD-LIMITS].

   The security and integrity of the key-distribution mechanism is vital
   to the security of the SFC system as a whole.

   NSH data is exposed to several threats:

   *  An on-path attacker modifying the NSH data.

   *  An attacker spoofing the NSH data.

   *  An attacker capturing and replaying the NSH data.

   *  Data carried in Context Headers revealing privacy-sensitive
      information to attackers.

   *  An attacker replacing the packet on which the NSH is imposed with
      a modified or bogus packet.

   In an SFC-enabled domain where the above attacks are possible, (1)
   NSH data MUST be integrity protected and replay protected, and (2)
   privacy-sensitive NSH metadata MUST be encrypted for confidentiality
   preservation purposes.  The Base and Service Path Headers are not
   encrypted.

   MACs with two levels of assurance are defined in Section 5.
   Considerations specific to each level of assurance are discussed in
   Sections 9.1 and 9.2.

   The attacks discussed in [ARCH-SFC-THREATS] are handled based on the
   solution specified in this document, with the exception of attacks
   dropping packets.  Such attacks can be detected by relying upon
   statistical analysis; such analysis is out of the scope of this
   document.  Also, if SFFs are not involved in the integrity checks, a
   misbehaving SFF that decrements SI while this should be done by an SF
   (SF bypass attack) will be detected by an upstream SF because the
   integrity check will fail.

   Some events are logged locally with notification alerts sent by NSH-
   aware nodes to a Control Element.  These events SHOULD be rate
   limited.

   The solution specified in this document does not provide data origin
   authentication.

   In order to detect compromised nodes, it is assumed that appropriate
   mechanisms to monitor and audit an SFC-enabled domain to detect
   misbehavior and to deter misuse are in place.  Compromised nodes can
   thus be withdrawn from active service function chains using
   appropriate control plane mechanisms.

9.1.  MAC#1



   An active attacker can potentially modify the Base Header (e.g.,
   decrement the TTL so the next SFF in the SFP discards the NSH
   packet).  An active attacker can typically also drop NSH packets.  As
   such, this attack is not considered an attack against the security
   mechanism specified in the document.

   It is expected that specific devices in the SFC-enabled domain will
   be configured such that no device other than the Classifiers (when
   reclassification is enabled), NSH-aware SFs, and SFC proxies will be
   able to update the integrity-protected NSH data (Section 7.1), and no
   device other than the NSH-aware SFs and SFC proxies will be able to
   decrypt and update the Context Headers carrying sensitive metadata
   (Section 7.1).  In other words, it is expected that the NSH-aware SFs
   and SFC proxies in the SFC-enabled domain are considered fully
   trusted to act on the NSH data.  Only these elements can have access
   to sensitive NSH metadata and the keying material used to integrity
   protect NSH data and encrypt Context Headers.

9.2.  MAC#2



   SFFs can detect whether an illegitimate node has altered the content
   of the Base Header.  Such messages must be discarded with appropriate
   logs and alarms generated (see Section 7.1).

   Similar to Section 9.1, no device other than the NSH-aware SFs and
   SFC proxies in the SFC-enabled domain should be able to decrypt and
   update the Context Headers carrying sensitive metadata.

9.3.  Time Synchronization



   [RFC8633] describes best current practices to be considered in
   deployments where SFC data plane elements use NTP for time-
   synchronization purposes.

   Also, a mechanism to provide cryptographic security for NTP is
   specified in [RFC8915].

10.  IANA Considerations



   IANA has added the following types to the "NSH IETF-Assigned Optional
   Variable-Length Metadata Types" registry (0x0000 IETF Base NSH MD
   Class) at <https://www.iana.org/assignments/nsh>.

           +=======+===============================+===========+
           | Value | Description                   | Reference |
           +=======+===============================+===========+
           | 0x02  | MAC and Encrypted Metadata #1 | RFC 9145  |
           +-------+-------------------------------+-----------+
           | 0x03  | MAC and Encrypted Metadata #2 | RFC 9145  |
           +-------+-------------------------------+-----------+

                Table 4: Additions to the NSH IETF-Assigned
              Optional Variable-Length Metadata Types Registry

11.  References



11.1.  Normative References



   [GCM]      Dworkin, M., "Recommendation for Block Cipher Modes of
              Operation: Galois/Counter Mode (GCM) and GMAC",
              NIST Special Publication 800-38D,
              DOI 10.6028/NIST.SP.800-38D, November 2007,
              <https://doi.org/10.6028/NIST.SP.800-38D>.

   [RFC2104]  Krawczyk, H., Bellare, M., and R. Canetti, "HMAC: Keyed-
              Hashing for Message Authentication", RFC 2104,
              DOI 10.17487/RFC2104, February 1997,
              <https://www.rfc-editor.org/info/rfc2104>.

   [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>.

   [RFC4107]  Bellovin, S. and R. Housley, "Guidelines for Cryptographic
              Key Management", BCP 107, RFC 4107, DOI 10.17487/RFC4107,
              June 2005, <https://www.rfc-editor.org/info/rfc4107>.

   [RFC4868]  Kelly, S. and S. Frankel, "Using HMAC-SHA-256, HMAC-SHA-
              384, and HMAC-SHA-512 with IPsec", RFC 4868,
              DOI 10.17487/RFC4868, May 2007,
              <https://www.rfc-editor.org/info/rfc4868>.

   [RFC5116]  McGrew, D., "An Interface and Algorithms for Authenticated
              Encryption", RFC 5116, DOI 10.17487/RFC5116, January 2008,
              <https://www.rfc-editor.org/info/rfc5116>.

   [RFC7665]  Halpern, J., Ed. and C. Pignataro, Ed., "Service Function
              Chaining (SFC) Architecture", RFC 7665,
              DOI 10.17487/RFC7665, October 2015,
              <https://www.rfc-editor.org/info/rfc7665>.

   [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>.

   [RFC8300]  Quinn, P., Ed., Elzur, U., Ed., and C. Pignataro, Ed.,
              "Network Service Header (NSH)", RFC 8300,
              DOI 10.17487/RFC8300, January 2018,
              <https://www.rfc-editor.org/info/rfc8300>.

11.2.  Informative References



   [AEAD-LIMITS]
              Günther, F., Thomson, M., and C. A. Wood, "Usage Limits on
              AEAD Algorithms", Work in Progress, Internet-Draft, draft-
              irtf-cfrg-aead-limits-03, 12 July 2021,
              <https://datatracker.ietf.org/doc/html/draft-irtf-cfrg-
              aead-limits-03>.

   [ARCH-SFC-THREATS]
              THANG, N. C. and M. Park, "A Security Architecture Against
              Service Function Chaining Threats", Work in Progress,
              Internet-Draft, draft-nguyen-sfc-security-architecture-00,
              24 November 2019, <https://datatracker.ietf.org/doc/html/
              draft-nguyen-sfc-security-architecture-00>.

   [INTERNET-IP-TUNNELS]
              Touch, J. and M. Townsley, "IP Tunnels in the Internet
              Architecture", Work in Progress, Internet-Draft, draft-
              ietf-intarea-tunnels-10, 12 September 2019,
              <https://datatracker.ietf.org/doc/html/draft-ietf-intarea-
              tunnels-10>.

   [INTERNET-THREAT-MODEL]
              Arkko, J. and S. Farrell, "Challenges and Changes in the
              Internet Threat Model", Work in Progress, Internet-Draft,
              draft-arkko-farrell-arch-model-t-04, 13 July 2020,
              <https://datatracker.ietf.org/doc/html/draft-arkko-
              farrell-arch-model-t-04>.

   [RFC4046]  Baugher, M., Canetti, R., Dondeti, L., and F. Lindholm,
              "Multicast Security (MSEC) Group Key Management
              Architecture", RFC 4046, DOI 10.17487/RFC4046, April 2005,
              <https://www.rfc-editor.org/info/rfc4046>.

   [RFC5905]  Mills, D., Martin, J., Ed., Burbank, J., and W. Kasch,
              "Network Time Protocol Version 4: Protocol and Algorithms
              Specification", RFC 5905, DOI 10.17487/RFC5905, June 2010,
              <https://www.rfc-editor.org/info/rfc5905>.

   [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>.

   [RFC7498]  Quinn, P., Ed. and T. Nadeau, Ed., "Problem Statement for
              Service Function Chaining", RFC 7498,
              DOI 10.17487/RFC7498, April 2015,
              <https://www.rfc-editor.org/info/rfc7498>.

   [RFC7635]  Reddy, T., Patil, P., Ravindranath, R., and J. Uberti,
              "Session Traversal Utilities for NAT (STUN) Extension for
              Third-Party Authorization", RFC 7635,
              DOI 10.17487/RFC7635, August 2015,
              <https://www.rfc-editor.org/info/rfc7635>.

   [RFC8165]  Hardie, T., "Design Considerations for Metadata
              Insertion", RFC 8165, DOI 10.17487/RFC8165, May 2017,
              <https://www.rfc-editor.org/info/rfc8165>.

   [RFC8459]  Dolson, D., Homma, S., Lopez, D., and M. Boucadair,
              "Hierarchical Service Function Chaining (hSFC)", RFC 8459,
              DOI 10.17487/RFC8459, September 2018,
              <https://www.rfc-editor.org/info/rfc8459>.

   [RFC8633]  Reilly, D., Stenn, H., and D. Sibold, "Network Time
              Protocol Best Current Practices", BCP 223, RFC 8633,
              DOI 10.17487/RFC8633, July 2019,
              <https://www.rfc-editor.org/info/rfc8633>.

   [RFC8877]  Mizrahi, T., Fabini, J., and A. Morton, "Guidelines for
              Defining Packet Timestamps", RFC 8877,
              DOI 10.17487/RFC8877, September 2020,
              <https://www.rfc-editor.org/info/rfc8877>.

   [RFC8915]  Franke, D., Sibold, D., Teichel, K., Dansarie, M., and R.
              Sundblad, "Network Time Security for the Network Time
              Protocol", RFC 8915, DOI 10.17487/RFC8915, September 2020,
              <https://www.rfc-editor.org/info/rfc8915>.

Acknowledgements



   This document was created as a follow-up to the discussion in IETF
   104: <https://datatracker.ietf.org/meeting/104/materials/slides-104-
   sfc-sfc-chair-slides-01> (slide 7).

   Thanks to Joel Halpern, Christian Jacquenet, Dirk von Hugo, Tal
   Mizrahi, Daniel Migault, Diego Lopez, Greg Mirsky, and Dhruv Dhody
   for the comments.

   Many thanks to Steve Hanna for the valuable secdir review and Juergen
   Schoenwaelder for the opsdir review.

   Thanks to Greg Mirsky for the Shepherd review.

   Thanks to Alvaro Retana, Lars Eggert, Martin Duke, Erik Kline,
   Zaheduzzaman Sarker, Éric Vyncke, Roman Danyliw, Murray Kucherawy,
   John Scudder, and Benjamin Kaduk for the IESG review.

Authors' Addresses



   Mohamed Boucadair
   Orange
   35000 Rennes
   France

   Email: mohamed.boucadair@orange.com


   Tirumaleswar Reddy.K
   Akamai
   Embassy Golf Link Business Park
   Bangalore 560071
   Karnataka
   India

   Email: kondtir@gmail.com


   Dan Wing
   Citrix Systems, Inc.
   United States of America

   Email: dwing-ietf@fuggles.com