RFC 9172

Internet Engineering Task Force (IETF)                   E. Birrane, III
Request for Comments: 9172                                   K. McKeever
Category: Standards Track                                        JHU/APL
ISSN: 2070-1721                                             January 2022

                    Bundle Protocol Security (BPSec)


   This document defines a security protocol providing data integrity
   and confidentiality services for the Bundle Protocol (BP).

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

Copyright Notice

   Copyright (c) 2022 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
     1.1.  Supported Security Services
     1.2.  Specification Scope
     1.3.  Related Documents
     1.4.  Terminology
   2.  Design Decisions
     2.1.  Block-Level Granularity
     2.2.  Multiple Security Sources
     2.3.  Mixed Security Policy
     2.4.  User-Defined Security Contexts
     2.5.  Deterministic Processing
   3.  Security Blocks
     3.1.  Block Definitions
     3.2.  Uniqueness
     3.3.  Target Multiplicity
     3.4.  Target Identification
     3.5.  Block Representation
     3.6.  Abstract Security Block
     3.7.  Block Integrity Block
     3.8.  Block Confidentiality Block
     3.9.  Block Interactions
     3.10. Parameter and Result Identification
     3.11. BPSec Block Examples
       3.11.1.  Example 1: Constructing a Bundle with Security
       3.11.2.  Example 2: Adding More Security at a New Node
   4.  Canonical Forms
   5.  Security Processing
     5.1.  Bundles Received from Other Nodes
       5.1.1.  Receiving BCBs
       5.1.2.  Receiving BIBs
     5.2.  Bundle Fragmentation and Reassembly
   6.  Key Management
   7.  Security Policy Considerations
     7.1.  Security Reason Codes
   8.  Security Considerations
     8.1.  Attacker Capabilities and Objectives
     8.2.  Attacker Behaviors and BPSec Mitigations
       8.2.1.  Eavesdropping Attacks
       8.2.2.  Modification Attacks
       8.2.3.  Topology Attacks
       8.2.4.  Message Injection
   9.  Security Context Considerations
     9.1.  Mandating Security Contexts
     9.2.  Identification and Configuration
     9.3.  Authorship
   10. Defining Other Security Blocks
   11. IANA Considerations
     11.1.  Bundle Block Types
     11.2.  Bundle Status Report Reason Codes
     11.3.  Security Context Identifiers
   12. References
     12.1.  Normative References
     12.2.  Informative References

   Authors' Addresses

1.  Introduction

   This document defines security features for the Bundle Protocol (BP)
   [RFC9171] and is intended for use in Delay-Tolerant Networking (DTN)
   to provide security services between a security source and a security
   acceptor.  When the security source is the bundle source and the
   security acceptor is the bundle destination, the security service
   provides end-to-end protection.

   The Bundle Protocol specification [RFC9171] defines DTN as referring
   to "a network architecture providing communications in and/or through
   highly stressed environments" where "BP may be viewed as sitting at
   the application layer of some number of constituent networks, forming
   a store-carry-forward overlay network".  The phrase "stressed
   environment" refers to multiple challenging conditions including
   intermittent connectivity, large and/or variable delays, asymmetric
   data rates, and high bit error rates.

   It should be presumed that the BP will be deployed in an untrusted
   network, which poses the usual security challenges related to
   confidentiality and integrity.  However, the stressed nature of the
   BP operating environment imposes unique conditions where usual
   transport security mechanisms may not be sufficient.  For example,
   the store-carry-forward nature of the network may require protecting
   data at rest, preventing unauthorized consumption of critical
   resources such as storage space, and operating without regular
   contact with a centralized security oracle (such as a certificate

   An end-to-end security service that operates in all of the
   environments where the BP operates is needed.

1.1.  Supported Security Services

   BPSec provides integrity and confidentiality services for BP bundles,
   as defined in this section.

   Integrity services ensure that changes to target data within a bundle
   can be discovered.  Data changes may be caused by processing errors,
   environmental conditions, or intentional manipulation.  In the
   context of BPSec, integrity services apply to plaintext in the

   Confidentiality services ensure that target data is unintelligible to
   nodes in DTN, except for authorized nodes possessing special
   information.  Generally, this means producing ciphertext from
   plaintext and generating authentication information for that
   ciphertext.  In this context, confidentiality applies to the contents
   of target data and does not extend to hiding the fact that
   confidentiality exists in the bundle.

   NOTE: Hop-by-hop authentication is NOT a supported security service
   in this specification, for two reasons:

   1.  The term "hop-by-hop" is ambiguous in a BP overlay, as nodes that
       are adjacent in the overlay may not be adjacent in physical
       connectivity.  This condition is difficult or impossible to
       detect; therefore, hop-by-hop authentication is difficult or
       impossible to enforce.

   2.  Hop-by-hop authentication cannot be deployed in a network if
       adjacent nodes in the network have incompatible security

1.2.  Specification Scope

   This document defines the security services provided by the BPSec.
   This includes the data specification for representing these services
   as BP extension blocks and the rules for adding, removing, and
   processing these blocks at various points during the bundle's
   traversal of a delay-tolerant network.

   BPSec addresses only the security of data traveling over the DTN, not
   the underlying DTN itself.  Furthermore, while the BPSec protocol can
   provide security-at-rest in a store-carry-forward network, it does
   not address threats that share computing resources with the DTN and/
   or BPSec software implementations.  These threats may be malicious
   software or compromised libraries that intend to intercept data or
   recover cryptographic material.  Here, it is the responsibility of
   the BPSec implementer to ensure that any cryptographic material,
   including shared secrets or private keys, is protected against access
   within both memory and storage devices.

   Completely trusted networks are extremely uncommon.  Among untrusted
   networks, different networking conditions and operational
   considerations require security mechanisms of varying strengths.
   Mandating a single security context, which is a set of assumptions,
   algorithms, configurations, and policies used to implement security
   services, may result in too much security for some networks and too
   little security in others.  Default security contexts are defined in
   [RFC9173] to provide basic security services for interoperability
   testing and for operational use on the terrestrial Internet.  It is
   expected that separate documents will define different security
   contexts for use in different networks.

   This specification addresses neither the fitness of externally
   defined cryptographic methods nor the security of their

   This specification does not address the implementation of security
   policies and does not provide a security policy for the BPSec.
   Similar to cipher suites, security policies are based on the nature
   and capabilities of individual networks and network operational
   concepts.  This specification does provide policy considerations that
   can be taken into account when building a security policy.

   With the exception of the Bundle Protocol, this specification does
   not address how to combine the BPSec security blocks with other
   protocols, other BP extension blocks, or other best practices to
   achieve security in any particular network implementation.

1.3.  Related Documents

   This document is best read and understood within the context of the
   following other DTN documents:

   *  "Delay-Tolerant Networking Architecture" [RFC4838] defines the
      architecture for DTN and identifies certain security assumptions
      made by existing Internet protocols that are not valid in DTN.

   *  "Bundle Protocol Version 7" [RFC9171] defines the format and
      processing of bundles, the extension block format used to
      represent BPSec security blocks, and the canonical block structure
      used by this specification.

   *  "Concise Binary Object Representation (CBOR)" [RFC8949] defines a
      data format that allows for small code size, fairly small message
      size, and extensibility without version negotiation.  The block-
      type-specific data associated with BPSec security blocks is
      encoded in this data format.

   *  "Bundle Security Protocol Specification" [RFC6257] introduces the
      concept of using BP extension blocks for security services in DTN.
      BPSec is a continuation and refinement of this document.

1.4.  Terminology

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "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 section defines terminology that either is unique to the BPSec
   or is necessary for understanding the concepts defined in this

   Bundle Destination:  the Bundle Protocol Agent (BPA) that receives a
      bundle and delivers the payload of the bundle to an Application
      Agent.  Also, an endpoint comprising the node(s) at which the
      bundle is to be delivered.  The bundle destination acts as the
      security acceptor for every security target in every security
      block in every bundle it receives.

   Bundle Source:  the BPA that originates a bundle.  Also, any node ID
      of the node of which the BPA is a component.

   Cipher Suite:  a set of one or more algorithms providing integrity
      and/or confidentiality services.  Cipher suites may define user
      parameters (e.g., secret keys to use), but they do not provide
      values for those parameters.

   Forwarder:  any BPA that transmits a bundle in DTN.  Also, any node
      ID of the node of which the BPA that sent the bundle on its most
      recent hop is a component.

   Intermediate Receiver, Waypoint, or Next Hop:  any BPA that receives
      a bundle from a forwarder that is not the bundle destination.
      Also, any node ID of the node of which the BPA is a component.

   Path:  the ordered sequence of nodes through which a bundle passes on
      its way from source to destination.  The path is not necessarily
      known in advance by the bundle or any BPAs in DTN.

   Security Acceptor:  a BPA that processes and dispositions one or more
      security blocks in a bundle.  Security acceptors act as the
      endpoint of a security service represented in a security block.
      They remove the security blocks they act upon as part of
      processing and disposition.  Also, any node ID of the node of
      which the BPA is a component.

   Security Block:  a BPSec extension block in a bundle.

   Security Context:  the set of assumptions, algorithms,
      configurations, and policies used to implement security services.

   Security Operation:  the application of a given security service to a
      security target, notated as OP(security service, security target).
      For example, OP(bcb-confidentiality, payload).  Every security
      operation in a bundle MUST be unique, meaning that a given
      security service can only be applied to a security target once in
      a bundle.  A security operation is implemented by a security

   Security Service:  a process that gives some protection to a security
      target.  For example, this specification defines security services
      for plaintext integrity (bib-integrity) and authenticated
      plaintext confidentiality with additional authenticated data (bcb-

   Security Source:  a BPA that adds a security block to a bundle.
      Also, any node ID of the node of which the BPA is a component.

   Security Target:  the block within a bundle that receives a security
      service as part of a security operation.

   Security Verifier:  a BPA that verifies the data integrity of one or
      more security blocks in a bundle.  Unlike security acceptors,
      security verifiers do not act as the endpoint of a security
      service, and they do not remove verified security blocks.  Also,
      any node ID of the node of which the BPA is a component.

2.  Design Decisions

   The application of security services in DTN is a complex endeavor
   that must consider physical properties of the network (such as
   connectivity and propagation times), policies at each node,
   application security requirements, and current and future threat
   environments.  This section identifies those desirable properties
   that guide design decisions for this specification and that are
   necessary for understanding the format and behavior of the BPSec

2.1.  Block-Level Granularity

   Security services within this specification must allow different
   blocks within a bundle to have different security services applied to

   Blocks within a bundle represent different types of information.  The
   primary block contains identification and routing information.  The
   payload block carries application data.  Extension blocks carry a
   variety of data that may augment or annotate the payload or that
   otherwise provide information necessary for the proper processing of
   a bundle along a path.  Therefore, applying a single level and type
   of security across an entire bundle fails to recognize that blocks in
   a bundle represent different types of information with different
   security needs.

   For example, a payload block might be encrypted to protect its
   contents and an extension block containing summary information
   related to the payload might be integrity signed but unencrypted to
   provide waypoints access to payload-related data without providing
   access to the payload.

2.2.  Multiple Security Sources

   A bundle can have multiple security blocks, and these blocks can have
   different security sources.  BPSec implementations MUST NOT assume
   that all blocks in a bundle have the same security operations applied
   to them.

   The Bundle Protocol allows extension blocks to be added to a bundle
   at any time during its existence in DTN.  When a waypoint adds a new
   extension block to a bundle, that extension block MAY have security
   services applied to it by that waypoint.  Similarly, a waypoint MAY
   add a security service to an existing block, consistent with its
   security policy.

   When a waypoint adds a security service to the bundle, the waypoint
   is the security source for that service.  The security block(s) that
   represent that service in the bundle may need to record this security
   source, as the bundle destination might need this information for

   For example, a bundle source may choose to apply an integrity service
   to its plaintext payload.  Later a waypoint node, representing a
   gateway to another portion of the delay-tolerant network, may receive
   the bundle and choose to apply a confidentiality service.  In this
   case, the integrity security source is the bundle source and the
   confidentiality security source is the waypoint node.

   In cases where the security source and security acceptor are not the
   bundle source and bundle destination, respectively, it is possible
   that the bundle will reach the bundle destination prior to reaching a
   security acceptor.  In cases where this may be a practical problem,
   it is recommended that solutions such as bundle encapsulation be used
   to ensure that a bundle be delivered to a security acceptor prior to
   being delivered to the bundle destination.  Generally, if a bundle
   reaches a waypoint that has the appropriate configuration and policy
   to act as a security acceptor for a security service in the bundle,
   then the waypoint should act as that security acceptor.

2.3.  Mixed Security Policy

   The security policy enforced by nodes in the delay-tolerant network
   may differ.

   Some waypoints will have security policies that require the waypoint
   to evaluate security services even if the waypoint is neither the
   bundle destination nor the final intended acceptor of the service.
   For example, a waypoint could choose to verify an integrity service
   even though the waypoint is not the bundle destination and the
   integrity service will be needed by other nodes along the bundle's

   Some waypoints will determine, through policy, that they are the
   intended recipient of the security service and will terminate the
   security service in the bundle.  For example, a gateway node could
   determine that, even though it is not the destination of the bundle,
   it should verify and remove a particular integrity service or attempt
   to decrypt a confidentiality service, before forwarding the bundle
   along its path.

   Some waypoints could understand security blocks but refuse to process
   them unless they are the bundle destination.

2.4.  User-Defined Security Contexts

   A security context is the set of assumptions, algorithms,
   configurations, and policies used to implement security services.
   Different contexts may specify different algorithms, different
   polices, or different configuration values used in the implementation
   of their security services.  BPSec provides a mechanism to define
   security contexts.  Users may select from registered security
   contexts and customize those contexts through security context

   For example, some users might prefer a SHA2 hash function for
   integrity, whereas other users might prefer a SHA3 hash function.
   Providing either separate security contexts or a single,
   parameterized security context allows users flexibility in applying
   the desired cipher suite, policy, and configuration when populating a
   security block.

2.5.  Deterministic Processing

   Whenever a node determines that it must process more than one
   security block in a received bundle (either because the policy at a
   waypoint states that it should process security blocks or because the
   node is the bundle destination), the order in which security blocks
   are processed must be deterministic.  All nodes must impose this same
   deterministic processing order for all security blocks.  This
   specification provides determinism in the application and evaluation
   of security services, even when doing so results in a loss of

3.  Security Blocks

3.1.  Block Definitions

   This specification defines two types of security block: the Block
   Integrity Block (BIB) and the Block Confidentiality Block (BCB).

   *  The BIB is used to ensure the integrity of its plaintext security
      target(s).  The integrity information in the BIB MAY be verified
      by any node along the bundle path from the BIB security source to
      the bundle destination.  Waypoints add or remove BIBs from bundles
      in accordance with their security policy.  BIBs are never used for
      integrity protection of the ciphertext provided by a BCB.  Because
      security policy at BPSec nodes may differ regarding integrity
      verification, BIBs do not guarantee hop-by-hop authentication, as
      discussed in Section 1.1.

   *  The BCB indicates that the security target or targets have been
      encrypted at the BCB security source in order to protect their
      content while in transit.  As a matter of security policy, the BCB
      is decrypted by security acceptor nodes in the network, up to and
      including the bundle destination.  BCBs additionally provide
      integrity-protection mechanisms for the ciphertext they generate.

3.2.  Uniqueness

   Security operations in a bundle MUST be unique; the same security
   service MUST NOT be applied to a security target more than once in a
   bundle.  Since a security operation is represented by a security
   block, this means that multiple security blocks of the same type
   cannot share the same security targets.  A new security block MUST
be added to a bundle if a preexisting security block of the same
   type is already defined for the security target of the new security

   This uniqueness requirement ensures that there is no ambiguity
   related to the order in which security blocks are processed or how
   security policy can be specified to require certain security services
   be present in a bundle.

   Using the notation OP(service, target), several examples illustrate
   this uniqueness requirement.

   Signing the payload twice:  The two operations OP(bib-integrity,
      payload) and OP(bib-integrity, payload) are redundant and MUST NOT
      both be present in the same bundle at the same time.

   Signing different blocks:  The two operations OP(bib-integrity,
      payload) and OP(bib-integrity, extension_block_1) are not
      redundant and both may be present in the same bundle at the same
      time.  Similarly, the two operations OP(bib-integrity,
      extension_block_1) and OP(bib-integrity, extension_block_2) are
      also not redundant and may both be present in the bundle at the
      same time.

   Different services on same block:  The two operations OP(bib-
      integrity, payload) and OP(bcb-confidentiality, payload) are not
      inherently redundant and may both be present in the bundle at the
      same time, pursuant to other processing rules in this

   Different services from different block types:  The notation
      OP(service, target) refers specifically to a security block, as
      the security block is the embodiment of a security service applied
      to a security target in a bundle.  Were some Other Security Block
      (OSB) to be defined providing an integrity service, then the
      operations OP(bib-integrity, target) and OP(osb-integrity, target)
      MAY both be present in the same bundle if so allowed by the
      definition of the OSB, as discussed in Section 10.


   *  A security block may be removed from a bundle as part of security
      processing at a waypoint node with a new security block being
      added to the bundle by that node.  In this case, conflicting
      security blocks never coexist in the bundle at the same time and
      the uniqueness requirement is not violated.

   *  A ciphertext integrity-protection mechanism (such as associated
      authenticated data) calculated by a cipher suite and transported
      in a BCB is considered part of the confidentiality service;
      therefore, it is unique from the plaintext integrity service
      provided by a BIB.

   *  The security blocks defined in this specification (BIB and BCB)
      are designed with the intention that the BPA adding these blocks
      is the authoritative source of the security service.  If a BPA
      adds a BIB on a security target, then the BIB is expected to be
      the authoritative source of integrity for that security target.
      If a BPA adds a BCB to a security target, then the BCB is expected
      to be the authoritative source of confidentiality for that
      security target.  More complex scenarios, such as having multiple
      nodes in a network sign the same security target, can be
      accommodated using the definition of custom security contexts (see
      Section 9) and/or the definition of OSBs (see Section 10).

3.3.  Target Multiplicity

   A single security block MAY represent multiple security operations as
   a way of reducing the overall number of security blocks present in a
   bundle.  In these circumstances, reducing the number of security
   blocks in the bundle reduces the amount of redundant information in
   the bundle.

   A set of security operations can be represented by a single security
   block when all of the following conditions are true.

   *  The security operations apply the same security service.  For
      example, they are all integrity operations or all confidentiality

   *  The security context parameters for the security operations are

   *  The security source for the security operations is the same,
      meaning the set of operations are being added by the same node.

   *  No security operations have the same security target, as that
      would violate the need for security operations to be unique.

   *  None of the security operations conflict with security operations
      already present in the bundle.

   When representing multiple security operations in a single security
   block, the information that is common across all operations is
   represented once in the security block; the information that is
   different (e.g., the security targets) is represented individually.

   If a node processes any security operation in a security block, it is
   RECOMMENDED that it process all security operations in the security
   block.  This allows security sources to assert that the set of
   security operations in a security block are expected to be processed
   by the same security acceptor.  However, the determination of whether
   a node actually is a security acceptor or not is a matter of the
   policy of the node itself.  In cases where a receiving node
   determines that it is the security acceptor of only a subset of the
   security operations in a security block, the node may choose to only
   process that subset of security operations.

3.4.  Target Identification

   A security target is a block in the bundle to which a security
   service applies.  This target must be uniquely and unambiguously
   identifiable when processing a security block.  The definition of the
   extension block header from [RFC9171] provides a "block number" field
   suitable for this purpose.  Therefore, a security target in a
   security block MUST be represented as the block number of the target

3.5.  Block Representation

   Each security block uses the Canonical Bundle Block Format as defined
   in [RFC9171].  That is, each security block is comprised of the
   following elements:

   *  block type code
   *  block number
   *  block processing control flags
   *  cyclic redundancy check (CRC) type
   *  block-type-specific data
   *  CRC field (if present)

   Security-specific information for a security block is captured in the
   block-type-specific data field.

3.6.  Abstract Security Block

   The structure of the security-specific portions of a security block
   is identical for both the BIB and BCB block types.  Therefore, this
   section defines an Abstract Security Block (ASB) data structure and
   discusses its definition, its processing, and other constraints for
   using this structure.  An ASB is never directly instantiated within a
   bundle, it is only a mechanism for discussing the common aspects of
   BIB and BCB security blocks.

   The fields of the ASB SHALL be as follows, listed in the order in
   which they must appear.  The encoding of these fields MUST be in
   accordance with the canonical forms provided in Section 4.

   Security Targets:
         This field identifies the block(s) targeted by the security
         operation(s) represented by this security block.  Each target
         block is represented by its unique block number.  This field
         SHALL be represented by a Concise Binary Object Representation
         (CBOR) array of data items.  Each target within this CBOR array
         SHALL be represented by a CBOR unsigned integer.  This array
         MUST have at least one entry and each entry MUST represent the
         block number of a block that exists in the bundle.  There MUST
be duplicate entries in this array.  The order of elements
         in this list has no semantic meaning outside of the context of
         this block.  Within the block, the ordering of targets must
         match the ordering of results associated with these targets.

   Security Context Id:
         This field identifies the security context used to implement
         the security service represented by this block and applied to
         each security target.  This field SHALL be represented by a
         CBOR unsigned integer.  The values for this Id should come from
         the registry defined in Section 11.3.

   Security Context Flags:
         This field identifies which optional fields are present in the
         security block.  This field SHALL be represented as a CBOR
         unsigned integer whose contents shall be interpreted as a bit
         field.  Each bit in this bit field indicates the presence (bit
         set to 1) or absence (bit set to 0) of optional data in the
         security block.  The association of bits to security block data
         is defined as follows.

         Bit 0     (the least-significant bit, 0x01): "Security context
                   parameters present" flag.

         Bit >0    Reserved

         Implementations MUST set reserved bits to 0 when writing this
         field and MUST ignore the values of reserved bits when reading
         this field.  For unreserved bits, a value of 1 indicates that
         the associated security block field MUST be included in the
         security block.  A value of 0 indicates that the associated
         security block field MUST NOT be in the security block.

   Security Source:
         This field identifies the BPA that inserted the security block
         in the bundle.  Also, any node ID of the node of which the BPA
         is a component.  This field SHALL be represented by a CBOR
         array in accordance with the rules in [RFC9171] for
         representing endpoint IDs (EIDs).

   Security Context Parameters (Optional):
         This field captures one or more security context parameters
         that should be used when processing the security service
         described by this security block.  This field SHALL be
         represented by a CBOR array.  Each entry in this array is a
         single security context parameter.  A single parameter SHALL
         also be represented as a CBOR array comprising a 2-tuple of the
         Id and value of the parameter, as follows.

         Parameter Id:  This field identifies which parameter is being
            specified.  This field SHALL be represented as a CBOR
            unsigned integer.  Parameter Ids are selected as described
            in Section 3.10.

         Parameter Value:  This field captures the value associated with
            this parameter.  This field SHALL be represented by the
            applicable CBOR representation of the parameter, in
            accordance with Section 3.10.

         The logical layout of the parameters array is illustrated in
         Figure 1.

           +----------------+----------------+     +----------------+
           |  Parameter 1   |  Parameter 2   | ... |  Parameter N   |
           +------+---------+------+---------+     +------+---------+
           |  Id  |  Value  |  Id  |  Value  |     |  Id  |  Value  |
           +------+---------+------+---------+     +------+---------+

                      Figure 1: Security Context Parameters

   Security Results:
         This field captures the results of applying a security service
         to the security targets of the security block.  This field
         SHALL be represented as a CBOR array of target results.  Each
         entry in this array represents the set of security results for
         a specific security target.  The target results MUST be ordered
         identically to the Security Targets field of the security
         block.  This means that the first set of target results in this
         array corresponds to the first entry in the Security Targets
         field of the security block, and so on.  There MUST be one
         entry in this array for each entry in the Security Targets
         field of the security block.

         The set of security results for a target is also represented as
         a CBOR array of individual results.  An individual result is
         represented as a CBOR array comprising a 2-tuple of a result Id
         and a result value, defined as follows.

         Result Id:  This field identifies which security result is
            being specified.  Some security results capture the primary
            output of a cipher suite.  Other security results contain
            additional annotative information from cipher suite
            processing.  This field SHALL be represented as a CBOR
            unsigned integer.  Security result Ids will be as specified
            in Section 3.10.

         Result Value:  This field captures the value associated with
            the result.  This field SHALL be represented by the
            applicable CBOR representation of the result value, in
            accordance with Section 3.10.

         The logical layout of the security results array is illustrated
         in Figure 2.  In this figure, there are N security targets for
         this security block.  The first security target contains M
         results and the Nth security target contains K results.

         +--------------------------+     +---------------------------+
         |          Target 1        |     |         Target N          |
         +----------+----+----------+     +---------------------------+
         | Result 1 |    | Result M | ... | Result 1 |    |  Result K |
         +----+-----+ .. +----+-----+     +---+------+ .. +----+------+
         | Id |Value|    | Id |Value|     | Id |Value|    | Id | Value|
         +----+-----+    +----+-----+     +----+-----+    +----+------+

                            Figure 2: Security Results

3.7.  Block Integrity Block

   A BIB is a BP extension block with the following characteristics.

   *  The block type code value is as specified in Section 11.1.

   *  The block-type-specific data field follows the structure of the

   *  A security target listed in the Security Targets field MUST NOT
      reference a security block defined in this specification (e.g., a
      BIB or a BCB).

   *  The security context MUST utilize an authentication mechanism or
      an error detection mechanism.


   *  Designers SHOULD carefully consider the effect of setting flags
      that either discard the block or delete the bundle in the event
      that this block cannot be processed.

   *  Since OP(bib-integrity, target) is allowed only once in a bundle
      per target, it is RECOMMENDED that users wishing to support
      multiple integrity-protection mechanisms for the same target
      define a multi-result security context.  Such a context could
      generate multiple security results for the same security target
      using different integrity-protection mechanisms or different
      configurations for the same integrity-protection mechanism.

   *  A BIB is used to verify the plaintext integrity of its security
      target.  However, a single BIB MAY include security results for
      blocks other than its security target when doing so establishes a
      needed relationship between the BIB security target and other
      blocks in the bundle (such as the primary block).

   *  Security information MAY be checked at any hop on the way to the
      bundle destination that has access to the required keying
      information, in accordance with Section 3.9.

3.8.  Block Confidentiality Block

   A BCB is a BP extension block with the following characteristics.

   *  The block type code value is as specified in Section 11.1.

   *  The block processing control flags value can be set to whatever
      values are required by local policy with the following exceptions:

      -  BCBs MUST have the "Block must be replicated in every fragment"
         flag set if one of the targets is the payload block.  Having
         that BCB in each fragment indicates to a receiving node that
         the payload portion of each fragment represents ciphertext.

      -  BCBs MUST NOT have the "Block must be removed from bundle if it
         can't be processed" flag set.  Removing a BCB from a bundle
         without decrypting its security targets removes information
         from the bundle necessary for their later decryption.

   *  The block-type-specific data fields follow the structure of the

   *  A security target listed in the Security Targets field can
      reference the payload block, a non-security extension block, or a
      BIB.  A BCB MUST NOT include another BCB as a security target.  A
      BCB MUST NOT target the primary block.  A BCB MUST NOT target a
      BIB unless it shares a security target with that BIB.

   *  Any security context used by a BCB MUST utilize a confidentiality
      cipher that provides authenticated encryption with associated data

   *  Additional information created by a cipher suite (such as an
      authentication tag) can be placed either in a security result
      field or in the generated ciphertext.  The determination of where
      to place this information is a function of the cipher suite and
      security context used.

   The BCB modifies the contents of its security target(s).  When a BCB
   is applied, the security target body data are encrypted "in-place".
   Following encryption, the security target block-type-specific data
   field contains ciphertext, not plaintext.


   *  It is RECOMMENDED that designers carefully consider the effect of
      setting flags that delete the bundle in the event that this block
      cannot be processed.

   *  The BCB block processing control flags can be set independently
      from the processing control flags of the security target(s).  The
      setting of such flags should be an implementation/policy decision
      for the encrypting node.

3.9.  Block Interactions

   The security block types defined in this specification are designed
   to be as independent as possible.  However, there are some cases
   where security blocks may share a security target; this sharing
   creates processing dependencies.

   If a BCB and a BIB share a security target, an undesirable condition
   occurs: a waypoint would be unable to validate the BIB because the
   shared security target has been encrypted by the BCB.  To address
   this situation, the following processing rules MUST be followed:

   *  When adding a BCB to a bundle, if some (or all) of the security
      targets of the BCB match all of the security targets of an
      existing BIB, then the existing BIB MUST also be encrypted.  This
      can be accomplished either by adding a new BCB that targets the
      existing BIB or by adding the BIB to the list of security targets
      for the BCB.  Deciding which way to represent this situation is a
      matter of security policy.

   *  When adding a BCB to a bundle, if some (or all) of the security
      targets of the BCB match some (but not all) of the security
      targets of a BIB, then that BIB MUST be altered in the following
      way.  Any security results in the BIB associated with the BCB
      security targets MUST be removed from the BIB and placed in a new
      BIB.  This newly created BIB MUST then be encrypted.  The
      encryption of the new BIB can be accomplished either by adding a
      new BCB that targets the new BIB or by adding the new BIB to the
      list of security targets for the BCB.  Deciding which way to
      represent this situation is a matter of security policy.

   *  A BIB MUST NOT be added for a security target that is already the
      security target of a BCB as this would cause ambiguity in block
      processing order.

   *  A BIB integrity value MUST NOT be checked if the BIB is the
      security target of an existing BCB.  In this case, the BIB data is

   *  A BIB integrity value MUST NOT be checked if the security target
      associated with that value is also the security target of a BCB.
      In such a case, the security target data contains ciphertext as it
      has been encrypted.

   *  As mentioned in Section 3.7, a BIB MUST NOT have a BCB as its
      security target.

   These restrictions on block interactions impose a necessary ordering
   when applying security operations within a bundle.  Specifically, for
   a given security target, BIBs MUST be added before BCBs.  This
   ordering MUST be preserved in cases where the current BPA is adding
   all of the security blocks for the bundle or where the BPA is a
   waypoint adding new security blocks to a bundle that already contains
   security blocks.

   In cases where a security source wishes to calculate both a plaintext
   integrity-protection mechanism and encrypt a security target, a BCB
   with a security context that generates an integrity-protection
   mechanism as one or more additional security results MUST be used
   instead of adding both a BIB and then a BCB for the security target
   at the security source.

3.10.  Parameter and Result Identification

   Each security context MUST define its own context parameters and
   results.  Each defined parameter and result is represented as the
   tuple of an identifier and a value.  Identifiers are always
   represented as a CBOR unsigned integer.  The CBOR encoding of values
   is as defined by the security context specification.

   Identifiers MUST be unique for a given security context but do not
   need to be unique amongst all security contexts.

   An example of a security context can be found in [RFC9173].

3.11.  BPSec Block Examples

   This section provides two examples of BPSec blocks applied to
   bundles.  In the first example, a single node adds several security
   operations to a bundle.  In the second example, a waypoint node
   received the bundle created in the first example and adds additional
   security operations.  In both examples, the first column represents
   blocks within a bundle and the second column represents the block
   number for the block, using the terminology B1...Bn for the purpose
   of illustration.

3.11.1.  Example 1: Constructing a Bundle with Security

   In this example, a bundle has four non-security-related blocks: the
   primary block (B1), two extension blocks (B4, B5), and a payload
   block (B6).  The bundle source wishes to provide an integrity
   signature of the plaintext associated with the primary block, the
   second extension block, and the payload.  The bundle source also
   wishes to provide confidentiality for the first extension block.  The
   resultant bundle is illustrated in Figure 3 and the security actions
   are described below.

                           Block in Bundle                ID
             |              Primary Block               | B1 |
             |                    BIB                   | B2 |
             |   OP(bib-integrity, targets = B1, B5, B6)|    |
             |                    BCB                   | B3 |
             |    OP(bcb-confidentiality, target = B4)  |    |
             |       Extension Block (encrypted)        | B4 |
             |              Extension Block             | B5 |
             |               Payload Block              | B6 |

                   Figure 3: Security at Bundle Creation

   The following security actions were applied to this bundle at its
   time of creation.

   *  An integrity signature applied to the canonical form of the
      primary block (B1), the canonical form of the block-type-specific
      data field of the second extension block (B5), and the canonical
      form of the payload block (B6).  This is accomplished by a single
      BIB (B2) with multiple targets.  A single BIB is used in this case
      because all three targets share a security source, security
      context, and security context parameters.  Had this not been the
      case, multiple BIBs could have been added instead.

   *  Confidentiality for the first extension block (B4).  This is
      accomplished by a BCB (B3).  Once applied, the block-type-specific
      data field of extension block B4 is encrypted.  The BCB MUST hold
      an authentication tag for the ciphertext either in the ciphertext
      that now populates the first extension block or as a security
      result in the BCB itself, depending on which security context is
      used to form the BCB.  A plaintext integrity signature may also
      exist as a security result in the BCB if one is provided by the
      selected confidentiality security context.

3.11.2.  Example 2: Adding More Security at a New Node

   Consider that the bundle as it is illustrated in Figure 3 is now
   received by a waypoint node that wishes to encrypt the second
   extension block and the bundle payload.  The waypoint security policy
   is to allow existing BIBs for these blocks to persist, as they may be
   required as part of the security policy at the bundle destination.

   The resultant bundle is illustrated in Figure 4 and the security
   actions are described below.  Note that block IDs provided here are
   ordered solely for the purpose of this example and are not meant to
   impose an ordering for block creation.  The ordering of blocks added
   to a bundle MUST always be in compliance with [RFC9171].

                           Block in Bundle                ID
             |              Primary Block               | B1 |
             |                    BIB                   | B2 |
             |      OP(bib-integrity, target = B1)      |    |
             |                    BIB (encrypted)       | B7 |
             |      OP(bib-integrity, targets = B5, B6) |    |
             |                    BCB                   | B8 |
             |OP(bcb-confidentiality,targets = B5,B6,B7)|    |
             |                    BCB                   | B3 |
             |    OP(bcb-confidentiality, target = B4)  |    |
             |       Extension Block (encrypted)        | B4 |
             |       Extension Block (encrypted)        | B5 |
             |         Payload Block (encrypted)        | B6 |

                  Figure 4: Security at Bundle Forwarding

   The following security actions were applied to this bundle prior to
   its forwarding from the waypoint node.

   *  Since the waypoint node wishes to encrypt the block-type-specific
      data field of blocks B5 and B6, it MUST also encrypt the block-
      type-specific data field of the BIBs providing plaintext integrity
      over those blocks.  However, BIB B2 could not be encrypted in its
      entirety because it also held a signature for the primary block
      (B1).  Therefore, a new BIB (B7) is created and security results
      associated with B5 and B6 are moved out of BIB B2 and into BIB B7.

   *  Now that there is no longer confusion about which plaintext
      integrity signatures must be encrypted, a BCB is added to the
      bundle with the security targets being the second extension block
      (B5) and the payload (B6) as well as the newly created BIB holding
      their plaintext integrity signatures (B7).  A single new BCB is
      used in this case because all three targets share a security
      source, security context, and security context parameters.  Had
      this not been the case, multiple BCBs could have been added

4.  Canonical Forms

   Security services require consistency and determinism in how
   information is presented to cipher suites at security sources,
   verifiers, and acceptors.  For example, integrity services require
   that the same target information (e.g., the same bits in the same
   order) is provided to the cipher suite when generating an original
   signature and when validating a signature.  Canonicalization
   algorithms transcode the contents of a security target into a
   canonical form.

   Canonical forms are used to generate input to a security context for
   security processing at a BP node.  If the values of a security target
   are unchanged, then the canonical form of that target will be the
   same even if the encoding of those values for wire transmission is

   BPSec operates on data fields within bundle blocks (e.g., the block-
   type-specific data field).  In their canonical form, these fields
   MUST include their own CBOR encoding and MUST NOT include any other
   encapsulating CBOR encoding.  For example, the canonical form of the
   block-type-specific data field is a CBOR byte string existing within
   the CBOR array containing the fields of the extension block.  The
   entire CBOR byte string is considered the canonical block-type-
   specific data field.  The CBOR array framing is not considered part
   of the field.

   The canonical form of the primary block is as specified in [RFC9171]
   with the following constraint.

   *  CBOR values from the primary block MUST be canonicalized using the
      rules for Deterministically Encoded CBOR, as specified in

   All non-primary blocks share the same block structure and are
   canonicalized as specified in [RFC9171] with the following

   *  CBOR values from the non-primary block MUST be canonicalized using
      the rules for Deterministically Encoded CBOR, as specified in

   *  Only the block-type-specific data field may be provided to a
      cipher suite for encryption as part of a confidentiality security
      service.  Other fields within a non-primary block MUST NOT be
      encrypted or decrypted and MUST NOT be included in the canonical
      form used by the cipher suite for encryption and decryption.  An
      integrity-protection mechanism MAY be applied to these other
      fields as supported by the security context.  For example, these
      fields might be treated as associated authenticated data.

   *  Reserved and unassigned flags in the block processing control
      flags field MUST be set to 0 in a canonical form as it is not
      known if those flags will change in transit.

   Security contexts MAY define their own canonicalization algorithms
   and require the use of those algorithms over the ones provided in
   this specification.  In the event of conflicting canonicalization
   algorithms, algorithms defined in a security context take precedence
   over this specification when constructing canonical forms for that
   security context.

5.  Security Processing

   This section describes the security aspects of bundle processing.

5.1.  Bundles Received from Other Nodes

   Security blocks must be processed in a specific order when received
   by a BP node.  The processing order is as follows.

   *  When BIBs and BCBs share a security target, BCBs MUST be evaluated
      first and BIBs second.

5.1.1.  Receiving BCBs

   If a received bundle contains a BCB, the receiving node MUST
   determine whether it is the security acceptor for any of the security
   operations in the BCB.  If so, the node MUST process those operations
   and remove any operation-specific information from the BCB prior to
   delivering data to an application at the node or forwarding the
   bundle.  If processing a security operation fails, the target SHALL
   be processed according to the security policy.  A bundle status
   report indicating the failure MAY be generated.  When all security
   operations for a BCB have been removed from the BCB, the BCB MUST be
   removed from the bundle.

   If the receiving node is the destination of the bundle, the node MUST
   decrypt any BCBs remaining in the bundle.  If the receiving node is
   not the destination of the bundle, the node MUST process the BCB if
   directed to do so as a matter of security policy.

   If the security policy of a node specifies that a node should have
   applied confidentiality to a specific security target and no such BCB
   is present in the bundle, then the node MUST process this security
   target in accordance with the security policy.  It is RECOMMENDED
   that the node remove the security target from the bundle because the
   confidentiality (and possibly the integrity) of the security target
   cannot be guaranteed.  If the removed security target is the payload
   block, the bundle MUST be discarded.

   If an encrypted payload block cannot be decrypted (i.e., the
   ciphertext cannot be authenticated), then the bundle MUST be
   discarded and processed no further.  If an encrypted security target
   other than the payload block cannot be decrypted, then the associated
   security target and all security blocks associated with that target
   MUST be discarded and processed no further.  In both cases, requested
   status reports (see [RFC9171]) MAY be generated to reflect bundle or
   block deletion.

   When a BCB is decrypted, the recovered plaintext for each security
   target MUST replace the ciphertext in each of the security targets'
   block-type-specific data fields.  If the plaintext is of a different
   size than the ciphertext, the framing of the CBOR byte string of this
   field must be updated to ensure this field remains a valid CBOR byte
   string.  The length of the recovered plaintext is known by the
   decrypting security context.

   If a BCB contains multiple security operations, each operation
   processed by the node MUST be treated as if the security operation
   has been represented by a single BCB with a single security operation
   for the purposes of report generation and policy processing.

5.1.2.  Receiving BIBs

   If a received bundle contains a BIB, the receiving node MUST
   determine whether it is the security acceptor for any of the security
   operations in the BIB.  If so, the node MUST process those operations
   and remove any operation-specific information from the BIB prior to
   delivering data to an application at the node or forwarding the
   bundle.  If processing a security operation fails, the target SHALL
   be processed according to the security policy.  A bundle status
   report indicating the failure MAY be generated.  When all security
   operations for a BIB have been removed from the BIB, the BIB MUST be
   removed from the bundle.

   A BIB MUST NOT be processed if the security target of the BIB is also
   the security target of a BCB in the bundle.  Given the order of
   operations mandated by this specification, when both a BIB and a BCB
   share a security target, it means that the security target must have
   been encrypted after it was integrity signed; therefore, the BIB
   cannot be verified until the security target has been decrypted by
   processing the BCB.

   If the security policy of a node specifies that a node should have
   applied integrity to a specific security target and no such BIB is
   present in the bundle, then the node MUST process this security
   target in accordance with the security policy.  It is RECOMMENDED
   that the node remove the security target from the bundle if the
   security target is not the payload or primary block.  If the security
   target is the payload or primary block, the bundle MAY be discarded.
   This action can occur at any node that has the ability to verify an
   integrity signature, not just the bundle destination.

   If a receiving node is not the security acceptor of a security
   operation in a BIB, it MAY attempt to verify the security operation
   anyway to prevent forwarding corrupt data.  If the verification
   fails, the node SHALL process the security target in accordance with
   local security policy.  If a payload integrity check fails at a
   waypoint, it is RECOMMENDED that it be processed in the same way as a
   failure of a payload integrity check at the bundle destination.  If
   the check passes, the node MUST NOT remove the security operation
   from the BIB prior to forwarding.

   If a BIB contains multiple security operations, each operation
   processed by the node MUST be treated as if the security operation
   has been represented by a single BIB with a single security operation
   for the purposes of report generation and policy processing.

5.2.  Bundle Fragmentation and Reassembly

   If it is necessary for a node to fragment a bundle payload, and
   security services have been applied to that bundle, the fragmentation
   rules described in [RFC9171] MUST be followed.  As defined there and
   summarized here for completeness, only the payload block can be
   fragmented; security blocks, like all extension blocks, can never be

   Due to the complexity of payload-block fragmentation, including the
   possibility of fragmenting payload-block fragments, integrity and
   confidentiality operations are not to be applied to a bundle
   representing a fragment.  Specifically, a BCB or BIB MUST NOT be
   added to a bundle if the "Bundle is a fragment" flag is set in the
   bundle processing control flags field.

   Security processing in the presence of payload-block fragmentation
   may be handled by other mechanisms outside of the BPSec protocol or
   by applying BPSec blocks in coordination with an encapsulation
   mechanism.  A node should apply any confidentiality protection prior
   to performing any fragmentation.

6.  Key Management

   There exists a myriad of ways to establish, communicate, and
   otherwise manage key information in DTN.  Certain DTN deployments
   might follow established protocols for key management, whereas other
   DTN deployments might require new and novel approaches.  BPSec
   assumes that key management is handled as a separate part of network
   management; this specification neither defines nor requires a
   specific strategy for key management.

7.  Security Policy Considerations

   When implementing BPSec, several policy decisions must be considered.
   This section describes key policies that affect the generation,
   forwarding, and receipt of bundles that are secured using this
   specification.  No single set of policy decisions is envisioned to
   work for all secure DTN deployments.

   *  If a bundle is received that contains combinations of security
      operations that are disallowed by this specification, the BPA must
      determine how to handle the bundle: the bundle may be discarded,
      the block affected by the security operation may be discarded, or
      one security operation may be favored over another.

   *  BPAs in the network must understand what security operations they
      should apply to bundles.  This decision may be based on the source
      of the bundle, the destination of the bundle, or some other
      information related to the bundle.

   *  If a waypoint has been configured to add a security operation to a
      bundle, and the received bundle already has the security operation
      applied, then the receiver must understand what to do.  The
      receiver may discard the bundle, discard the security target and
      associated BPSec blocks, replace the security operation, or take
      some other action.

   *  It is RECOMMENDED that security operations be applied to every
      block in a bundle and that the default behavior of a BPA be to use
      the security services defined in this specification.  Designers
      should only deviate from the use of security operations when the
      deviation can be justified -- such as when doing so causes
      downstream errors when processing blocks whose contents must be
      inspected or changed at one or more hops along the path.

   *  BCB security contexts can alter the size of extension blocks and
      the payload block.  Security policy SHOULD consider how changes to
      the size of a block could negatively effect bundle processing
      (e.g., calculating storage needs and scheduling transmission

   *  Adding a BIB to a security target that has already been encrypted
      by a BCB is not allowed.  If this condition is likely to be
      encountered, there are (at least) three possible policies that
      could handle this situation.

      1.  At the time of encryption, a security context can be selected
          that computes a plaintext integrity-protection mechanism that
          is included as a security context result field.

      2.  The encrypted block may be replicated as a new block with a
          new block number and may be given integrity protection.

      3.  An encapsulation scheme may be applied to encapsulate the
          security target (or the entire bundle) such that the
          encapsulating structure is, itself, no longer the security
          target of a BCB and may therefore be the security target of a

   *  Security policy SHOULD address whether cipher suites whose
      ciphertext is larger than the initial plaintext are permitted and,
      if so, for what types of blocks.  Changing the size of a block may
      cause processing difficulties for networks that calculate block
      offsets into bundles or predict transmission times or storage
      availability as a function of bundle size.  In other cases,
      changing the size of a payload as part of encryption has no
      significant impact.

7.1.  Security Reason Codes

   BPAs must process blocks and bundles in accordance with both BP
   policy and BPSec policy.  The decision to receive, forward, deliver,
   or delete a bundle may be communicated to the report-to address of
   the bundle in the form of a status report, as a method of tracking
   the progress of the bundle through the network.  The status report
   for a bundle may be augmented with a "reason code" explaining why the
   particular action was taken on the bundle.

   This section describes a set of reason codes associated with the
   security processing of a bundle.  The communication of security-
   related status reports might reduce the security of a network if
   these reports are intercepted by unintended recipients.  BPSec policy
   SHOULD specify the conditions in which sending security reason codes
   are appropriate.  Examples of appropriate conditions for the use of
   security reason codes could include the following.

   *  When the report-to address is verified as unchanged from the
      bundle source.  This can occur by placing an appropriate BIB on
      the bundle primary block.

   *  When the block containing a status report with a security reason
      code is encrypted by a BCB.

   *  When a status report containing a security reason code is only
      sent for security issues relating to bundles and/or blocks
      associated with non-operational user data or test data.

   *  When a status report containing a security reason code is only
      sent for security issues associated with non-operational security
      contexts, or security contexts using non-operational
      configurations, such as test keys.

   Security reason codes are assigned in accordance with Section 11.2
   and are as described below.

   Missing security operation:
         This reason code indicates that a bundle was missing one or
         more required security operations.  This reason code is
         typically used by a security verifier or security acceptor.

   Unknown security operation:
         This reason code indicates that one or more security operations
         present in a bundle cannot be understood by the security
         verifier or security acceptor for the operation.  For example,
         this reason code may be used if a security block references an
         unknown security context identifier or security context
         parameter.  This reason code should not be used for security
         operations for which the node is not a security verifier or
         security acceptor; there is no requirement that all nodes in a
         network understand all security contexts, security context
         parameters, and security services for every bundle in a

   Unexpected security operation:
         This reason code indicates that a receiving node is neither a
         security verifier nor a security acceptor for at least one
         security operation in a bundle.  This reason code should not be
         seen as an error condition: not every node is a security
         verifier or security acceptor for every security operation in
         every bundle.  In certain networks, this reason code may be
         useful in identifying misconfigurations of security policy.

   Failed security operation:
         This reason code indicates that one or more security operations
         in a bundle failed to process as expected for reasons other
         than misconfiguration.  This may occur when a security-source
         is unable to add a security block to a bundle.  This may occur
         if the target of a security operation fails to verify using the
         defined security context at a security verifier.  This may also
         occur if a security operation fails to be processed without
         error at a security acceptor.

   Conflicting security operation:
         This reason code indicates that two or more security operations
         in a bundle are not conformant with the BPSec specification and
         that security processing was unable to proceed because of a
         BPSec protocol violation.

8.  Security Considerations

   Given the nature of DTN applications, it is expected that bundles may
   traverse a variety of environments and devices that each pose unique
   security risks and requirements on the implementation of security
   within BPSec.  For this reason, it is important to introduce key
   threat models and describe the roles and responsibilities of the
   BPSec protocol in protecting the confidentiality and integrity of the
   data against those threats.  This section provides additional
   discussion on security threats that BPSec will face and describes how
   BPSec security mechanisms operate to mitigate these threats.

   The threat model described here is assumed to have a set of
   capabilities identical to those described by the Internet Threat
   Model in [RFC3552], but the BPSec threat model is scoped to
   illustrate threats specific to BPSec operating within DTN
   environments; therefore, it focuses on on-path attackers (OPAs).  In
   doing so, it is assumed that the delay-tolerant network (or
   significant portions of the delay-tolerant network) are completely
   under the control of an attacker.

8.1.  Attacker Capabilities and Objectives

   BPSec was designed to protect against OPA threats that may have
   access to a bundle during transit from its source, Alice, to its
   destination, Bob.  An OPA node, Olive, is a noncooperative node
   operating on the delay-tolerant network between Alice and Bob that
   has the ability to receive bundles, examine bundles, modify bundles,
   forward bundles, and generate bundles at will in order to compromise
   the confidentiality or integrity of data within the delay-tolerant
   network.  There are three classes of OPA nodes that are
   differentiated based on their access to cryptographic material:

   Unprivileged Node:  Olive has not been provisioned within the secure
      environment and only has access to cryptographic material that has
      been publicly shared.

   Legitimate Node:  Olive is within the secure environment; therefore,
      Olive has access to cryptographic material that has been
      provisioned to Olive (i.e., K_M) as well as material that has been
      publicly shared.

   Privileged Node:  Olive is a privileged node within the secure
      environment; therefore, Olive has access to cryptographic material
      that has been provisioned to Olive, Alice, and/or Bob (i.e., K_M,
      K_A, and/or K_B) as well as material that has been publicly

   If Olive is operating as a privileged node, this is tantamount to
   compromise; BPSec does not provide mechanisms to detect or remove
   Olive from the delay-tolerant network or BPSec secure environment.
   It is up to the BPSec implementer or the underlying cryptographic
   mechanisms to provide appropriate capabilities if they are needed.
   It should also be noted that if the implementation of BPSec uses a
   single set of shared cryptographic material for all nodes, a
   legitimate node is equivalent to a privileged node because K_M == K_A
   == K_B.  For this reason, sharing cryptographic material in this way
   is not recommended.

   A special case of the legitimate node is when Olive is either Alice
   or Bob (i.e., K_M == K_A or K_M == K_B).  In this case, Olive is able
   to impersonate traffic as either Alice or Bob, respectively, which
   means that traffic to and from that node can be decrypted and
   encrypted, respectively.  Additionally, messages may be signed as
   originating from one of the endpoints.

8.2.  Attacker Behaviors and BPSec Mitigations

8.2.1.  Eavesdropping Attacks

   Once Olive has received a bundle, she is able to examine the contents
   of that bundle and attempt to recover any protected data or
   cryptographic keying material from the blocks contained within.  The
   protection mechanism that BPSec provides against this action is the
   BCB, which encrypts the contents of its security target, providing
   confidentiality of the data.  Of course, it should be assumed that
   Olive is able to attempt offline recovery of encrypted data, so the
   cryptographic mechanisms selected to protect the data should provide
   a suitable level of protection.

   When evaluating the risk of eavesdropping attacks, it is important to
   consider the lifetime of bundles on DTN.  Depending on the network,
   bundles may persist for days or even years.  Long-lived bundles imply
   that the data exists in the network for a longer period of time and,
   thus, there may be more opportunities to capture those bundles.
   Additionally, the implication is that long-lived bundles store
   information within that remains relevant and sensitive for long
   enough that, once captured, there is sufficient time to crack
   encryption associated with the bundle.  If a bundle does persist on
   the network for years and the cipher suite used for a BCB provides
   inadequate protection, Olive may be able to recover the protected
   data either before that bundle reaches its intended destination or
   before the information in the bundle is no longer considered

   NOTE: Olive is not limited by the bundle lifetime and may retain a
   given bundle indefinitely.

   NOTE: Irrespective of whether BPSec is used, traffic analysis will be

8.2.2.  Modification Attacks

   As a node participating in the delay-tolerant network between Alice
   and Bob, Olive will also be able to modify the received bundle,
   including non-BPSec data such as the primary block, payload blocks,
   or block processing control flags as defined in [RFC9171].  Olive
   will be able to undertake activities including modification of data
   within the blocks, replacement of blocks, addition of blocks, or
   removal of blocks.  Within BPSec, both the BIB and BCB provide
   integrity-protection mechanisms to detect or prevent data
   manipulation attempts by Olive.

   The BIB provides that protection to another block that is its
   security target.  The cryptographic mechanisms used to generate the
   BIB should be strong against collision attacks, and Olive should not
   have access to the cryptographic material used by the originating
   node to generate the BIB (e.g., K_A).  If both of these conditions
   are true, Olive will be unable to modify the security target or the
   BIB, and thus she cannot lead Bob to validate the security target as
   originating from Alice.

   Since BPSec security operations are implemented by placing blocks in
   a bundle, there is no in-band mechanism for detecting or correcting
   certain cases where Olive removes blocks from a bundle.  If Olive
   removes a BCB, but keeps the security target, the security target
   remains encrypted and there is a possibility that there may no longer
   be sufficient information to decrypt the block at its destination.
   If Olive removes both a BCB (or BIB) and its security target, there
   is no evidence left in the bundle of the security operation.
   Similarly, if Olive removes the BIB, but not the security target,
   there is no evidence left in the bundle of the security operation.
   In each of these cases, the implementation of BPSec must be combined
   with policy configuration at endpoints in the network that describe
   the expected and required security operations that must be applied on
   transmission and that are expected to be present on receipt.  This or
   other similar out-of-band information is required to correct for
   removal of security information in the bundle.

   A limitation of the BIB may exist within the implementation of BIB
   validation at the destination node.  If Olive is a legitimate node
   within the delay-tolerant network, the BIB generated by Alice with
   K_A can be replaced with a new BIB generated with K_M and forwarded
   to Bob.  If Bob is only validating that the BIB was generated by a
   legitimate user, Bob will acknowledge the message as originating from
   Olive instead of Alice.  Validating a BIB indicates only that the BIB
   was generated by a holder of the relevant key; it does not provide
   any guarantee that the bundle or block was created by the same
   entity.  In order to provide verifiable integrity checks, the BCB
   should require an encryption scheme that is Indistinguishable under
   adaptive Chosen Ciphertext Attack (IND-CCA2) secure.  Such an
   encryption scheme will guard against signature substitution attempts
   by Olive.  In this case, Alice creates a BIB with the protected data
   block as the security target and then creates a BCB with both the BIB
   and protected data block as its security targets.

8.2.3.  Topology Attacks

   If Olive is in an OPA position within the delay-tolerant network, she
   is able to influence how any bundles that come to her may pass
   through the network.  Upon receiving and processing a bundle that
   must be routed elsewhere in the network, Olive has three options as
   to how to proceed: not forward the bundle, forward the bundle as
   intended, or forward the bundle to one or more specific nodes within
   the network.

   Attacks that involve rerouting the bundles throughout the network are
   essentially a special case of the modification attacks described in
   this section, one where the attacker is modifying fields within the
   primary block of the bundle.  Given that BPSec cannot encrypt the
   contents of the primary block, alternate methods must be used to
   prevent this situation.  These methods may include requiring BIBs for
   primary blocks, using encapsulation, or otherwise strategically
   manipulating primary block data.  The details of any such mitigation
   technique are specific to the implementation of the deploying network
   and are outside of the scope of this document.

   Furthermore, routing rules and policies may be useful in enforcing
   particular traffic flows to prevent topology attacks.  While these
   rules and policies may utilize some features provided by BPSec, their
   definition is beyond the scope of this specification.

8.2.4.  Message Injection

   Olive is also able to generate new bundles and transmit them into the
   delay-tolerant network at will.  These bundles may be either 1)
   copies or slight modifications of previously observed bundles (i.e.,
   a replay attack) or 2) entirely new bundles generated based on the
   Bundle Protocol, BPSec, or other bundle-related protocols.  With
   these attacks, Olive's objectives may vary, but may be targeting
   either the Bundle Protocol or application-layer protocols conveyed by
   the Bundle Protocol.  The target could also be the storage and
   computing capabilities of the nodes running the bundle or
   application-layer protocols (e.g., a denial of service to flood on
   the storage of the store-and-forward mechanism or a computation that
   would process the bundles and perhaps prevent other activities).

   BPSec relies on cipher suite capabilities to prevent replay or forged
   message attacks.  A BCB used with appropriate cryptographic
   mechanisms may provide replay protection under certain circumstances.
   Alternatively, application data itself may be augmented to include
   mechanisms to assert data uniqueness and then be protected with a
   BIB, a BCB, or both along with other block data.  In such a case, the
   receiving node would be able to validate the uniqueness of the data.

   For example, a BIB may be used to validate the integrity of a
   bundle's primary block, which includes a timestamp and lifetime for
   the bundle.  If a bundle is replayed outside of its lifetime, then
   the replay attack will fail as the bundle will be discarded.
   Similarly, additional blocks, such as the Bundle Age, may be signed
   and validated to identify replay attacks.  Finally, security context
   parameters within BIBs and BCBs may include anti-replay mechanisms
   such as session identifiers, nonces, and dynamic passwords as
   supported by network characteristics.

9.  Security Context Considerations

9.1.  Mandating Security Contexts

   Because of the diversity of networking scenarios and node
   capabilities that may utilize BPSec, there is a risk that a single
   security context mandated for every possible BPSec implementation is
   not feasible.  For example, a security context appropriate for a
   resource-constrained node with limited connectivity may be
   inappropriate for use in a well-resourced, well-connected node.

   This does not mean that the use of BPSec in a particular network is
   meant to happen without security contexts for interoperability and
   default behavior.  Network designers must identify the minimal set of
   security contexts necessary for functions in their network.  For
   example, a default set of security contexts could be created for use
   over the terrestrial Internet, and they could be required by any
   BPSec implementation communicating over the terrestrial Internet.

   To ensure interoperability among various implementations, all BPSec
   implementations MUST support at least the current, mandatory security
   context(s) defined in IETF Standards Track RFCs.  As of this writing,
   that BP mandatory security context is specified in [RFC9173], but the
   mandatory security context(s) might change over time in accordance
   with usual IETF processes.  Such changes are likely to occur in the
   future if/when flaws are discovered in the applicable cryptographic
   algorithms, for example.

   Additionally, BPSec implementations need to support the security
   contexts that are required by the BP networks in which they are

   If a node serves as a gateway between two or more networks, the BPSec
   implementation at that node needs to support the union of security
   contexts mandated in those networks.

   BPSec has been designed to allow for a diversity of security contexts
   and for new contexts to be defined over time.  The use of different
   security contexts does not change the BPSec protocol itself, and the
   definition of new security contexts MUST adhere to the requirements
   of such contexts as presented in this section and generally in this

   Implementers should monitor the state of security context
   specifications to check for future updates and replacement.

9.2.  Identification and Configuration

   Security blocks uniquely identify the security context to be used in
   the processing of their security services.  The security context for
   a security block MUST be uniquely identifiable and MAY use parameters
   for customization.

   To reduce the number of security contexts used in a network, security
   context designers should make security contexts customizable through
   the definition of security context parameters.  For example, a single
   security context could be associated with a single cipher suite and
   security context parameters could be used to configure the use of
   this security context with different key lengths and different key
   management options without needing to define separate security
   contexts for each possible option.

   A single security context may be used in the application of more than
   one security service.  This means that a security context identifier
   MAY be used with a BIB, with a BCB, or with any other BPSec-compliant
   security block.  The definition of a security context MUST identify
   which security services may be used with the security context, how
   security context parameters are interpreted as a function of the
   security operation being supported, and which security results are
   produced for each security service.

   Network operators must determine the number, type, and configuration
   of security contexts in a system.  Networks with rapidly changing
   configurations may define relatively few security contexts with each
   context customized with multiple parameters.  For networks with more
   stability, or an increased need for confidentiality, a larger number
   of contexts can be defined with each context supporting few, if any,

   |   Context   | Parameters |               Definition              |
   |     Type    |            |                                       |
   |     Key     | Encrypted  |     AES-GCM-256 cipher suite with     |
   |   Exchange  |  Key, IV   | provided ephemeral key encrypted with |
   |     AES     |            |   a predetermined key encryption key  |
   |             |            |  and cleartext initialization vector. |
   |  Pre-Shared |     IV     |     AES-GCM-256 cipher suite with     |
   |   Key AES   |            |  predetermined key and predetermined  |
   |             |            |          key-rotation policy.         |
   | Out-of-Band |    None    |   AES-GCM-256 cipher suite with all   |
   |     AES     |            |          info predetermined.          |

                    Table 1: Security Context Examples

9.3.  Authorship

   Developers or implementers should consider the diverse performance
   and conditions of networks on which the Bundle Protocol (and,
   therefore, BPSec) will operate.  Specifically, the delay and capacity
   of DTNs can vary substantially.  Developers should consider these
   conditions to better describe the conditions in which those contexts
   will operate or exhibit vulnerability, and selection of these
   contexts for implementation should be made with consideration for
   this reality.  There are key differences that may limit the
   opportunity for a security context to leverage existing cipher suites
   and technologies that have been developed for use in more reliable

   Data Lifetime:  Depending on the application environment, bundles may
      persist on the network for extended periods of time, perhaps even
      years.  Cryptographic algorithms should be selected to ensure
      protection of data against attacks for a length of time reasonable
      for the application.

   One-Way Traffic:  Depending on the application environment, it is
      possible that only a one-way connection may exist between two
      endpoints, or if a two-way connection does exist, the round-trip
      time may be extremely large.  This may limit the utility of
      session key generation mechanisms, such as Diffie-Hellman, as a
      two-way handshake may not be feasible or reliable.

   Opportunistic Access:  Depending on the application environment, a
      given endpoint may not be guaranteed to be accessible within a
      certain amount of time.  This may make asymmetric cryptographic
      architectures that rely on a key distribution center or other
      trust center impractical under certain conditions.

   When developing security contexts for use with BPSec, the following
   information SHOULD be considered for inclusion in these

   Security Context Parameters:  Security contexts MUST define their
      parameter Ids, the data types of those parameters, and their CBOR

   Security Results:  Security contexts MUST define their security
      result Ids, the data types of those results, and their CBOR

   New Canonicalizations:  Security contexts may define new
      canonicalization algorithms as necessary.

   Ciphertext Size:  Security contexts MUST state whether their
      associated cipher suites generate ciphertext (to include any
      authentication information) that is of a different size than the
      input plaintext.

      If a security context does not wish to alter the size of the
      plaintext, it should place overflow bytes and authentication tags
      in security result fields.

   Block Header Information:  Security contexts SHOULD include block
      header information that is considered to be immutable for the
      block.  This information MAY include the block type code, block
      number, CRC type, and CRC field (if present or if missing and
      unlikely to be added later), and possibly certain block processing
      control flags.  Designers should input these fields as additional
      data for integrity protection when these fields are expected to
      remain unchanged over the path the block will take from the
      security source to the security acceptor.  Security contexts
      considering block header information MUST describe expected
      behavior when these fields fail their integrity verification.

   Handling CRC Fields:  Security contexts may include algorithms that
      alter the contexts of their security target block, such as the
      case when encrypting the block-type-specific data of a target
      block as part of a BCB confidentiality service.  Security context
      specifications SHOULD address how preexisting CRC type and CRC
      value fields be handled.  For example, a BCB security context
      could remove the plaintext CRC value from its target upon
      encryption and replace or recalculate the value upon decryption.

10.  Defining Other Security Blocks

   Other Security Blocks (OSBs) may be defined and used in addition to
   the security blocks identified in this specification.  BIB, BCB, and
   any future OSBs can coexist within a bundle and can be considered in
   conformance with BPSec if all of the following requirements are met
   by any future identified security blocks.

   *  OSBs MUST NOT reuse any enumerations identified in this
      specification, to include the block type codes for BIB and BCB.

   *  An OSB definition MUST state whether it can be the target of a BIB
      or a BCB.  The definition MUST also state whether the OSB can
      target a BIB or a BCB.

   *  An OSB definition MUST provide a deterministic processing order in
      the event that a bundle is received containing BIBs, BCBs, and
      OSBs.  This processing order MUST NOT alter the BIB and BCB
      processing orders identified in this specification.

   *  An OSB definition MUST provide a canonicalization algorithm if the
      default algorithm for non-primary-block canonicalization cannot be
      used to generate a deterministic input for a cipher suite.  This
      requirement can be waived if the OSB is defined so as to never be
      the security target of a BIB or a BCB.

   *  An OSB definition MUST NOT require any behavior of a BPSec BPA
      that is in conflict with the behavior identified in this
      specification.  In particular, the security processing
      requirements imposed by this specification must be consistent
      across all BPSec BPAs in a network.

   *  The behavior of an OSB when dealing with fragmentation must be
      specified and MUST NOT lead to ambiguous processing states.  In
      particular, an OSB definition should address how to receive and
      process an OSB in a bundle fragment that may or may not also
      contain its security target.  An OSB definition should also
      address whether an OSB may be added to a bundle marked as a

   Additionally, policy considerations for the management, monitoring,
   and configuration associated with blocks SHOULD be included in any
   OSB definition.

   NOTE: The burden of showing compliance with processing rules is
   placed upon the specifications defining new security blocks, and the
   identification of such blocks shall not, alone, require maintenance
   of this specification.

11.  IANA Considerations

   This specification includes fields that require registries managed by

11.1.  Bundle Block Types

   This specification allocates two block types from the existing
   "Bundle Block Types" registry defined in [RFC6255].

             | Value |      Description      |   Reference   |
             |   11  |    Block Integrity    | This document |
             |   12  | Block Confidentiality | This document |

                Table 2: Additional Entries for the "Bundle
                           Block Types" Registry

   The "Bundle Block Types" registry notes whether a block type is meant
   for use in BP version 6, BP version 7 (BPv7), or both.  The two block
   types defined in this specification are meant for use with BPv7.

11.2.  Bundle Status Report Reason Codes

   This specification allocates five reason codes from the existing
   "Bundle Status Report Reason Codes" registry defined in [RFC6255].

   | BP Version | Value |        Description         |   Reference    |
   |     7      |   12  | Missing security operation | This document, |
   |            |       |                            |  Section 7.1   |
   |     7      |   13  | Unknown security operation | This document, |
   |            |       |                            |  Section 7.1   |
   |     7      |   14  |    Unexpected security     | This document, |
   |            |       |         operation          |  Section 7.1   |
   |     7      |   15  | Failed security operation  | This document, |
   |            |       |                            |  Section 7.1   |
   |     7      |   16  |    Conflicting security    | This document, |
   |            |       |         operation          |  Section 7.1   |

     Table 3: Additional Entries for the "Bundle Status Report Reason
                             Codes" Registry

11.3.  Security Context Identifiers

   BPSec has a Security Context Identifier field for which IANA has
   created a new registry named "BPSec Security Context Identifiers".
   Initial values for this registry are given below.

   The registration policy for this registry is Specification Required
   (see [RFC8126]).

   The value range: signed 16-bit integer.

                  | Value | Description |   Reference   |
                  |  < 0  |   Reserved  | This document |
                  |   0   |   Reserved  | This document |

                      Table 4: "BPSec Security Context
                            Identifier" Registry

   Negative security context identifiers are reserved for local/site-
   specific uses.  The use of 0 as a security context identifier is for
   nonoperational testing purposes only.

12.  References

12.1.  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,

   [RFC3552]  Rescorla, E. and B. Korver, "Guidelines for Writing RFC
              Text on Security Considerations", BCP 72, RFC 3552,
              DOI 10.17487/RFC3552, July 2003,

   [RFC6255]  Blanchet, M., "Delay-Tolerant Networking Bundle Protocol
              IANA Registries", RFC 6255, DOI 10.17487/RFC6255, May
              2011, <https://www.rfc-editor.org/info/rfc6255>.

   [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>.

   [RFC8949]  Bormann, C. and P. Hoffman, "Concise Binary Object
              Representation (CBOR)", STD 94, RFC 8949,
              DOI 10.17487/RFC8949, December 2020,

   [RFC9171]  Burleigh, S., Fall, K., and E. Birrane, III, "Bundle
              Protocol Version 7", RFC 9171, DOI 10.17487/RFC9171,
              January 2022, <https://www.rfc-editor.org/info/rfc9171>.

   [RFC9173]  Birrane, III, E., White, A., and S. Heiner, "Default
              Security Contexts for Bundle Protocol Security (BPSec)",
              RFC 9173, DOI 10.17487/RFC9173, January 2022,

12.2.  Informative References

   [RFC4838]  Cerf, V., Burleigh, S., Hooke, A., Torgerson, L., Durst,
              R., Scott, K., Fall, K., and H. Weiss, "Delay-Tolerant
              Networking Architecture", RFC 4838, DOI 10.17487/RFC4838,
              April 2007, <https://www.rfc-editor.org/info/rfc4838>.

   [RFC6257]  Symington, S., Farrell, S., Weiss, H., and P. Lovell,
              "Bundle Security Protocol Specification", RFC 6257,
              DOI 10.17487/RFC6257, May 2011,

   [RFC8126]  Cotton, M., Leiba, B., and T. Narten, "Guidelines for
              Writing an IANA Considerations Section in RFCs", BCP 26,
              RFC 8126, DOI 10.17487/RFC8126, June 2017,


   The following participants contributed technical material, use cases,
   and useful thoughts on the overall approach to this security
   specification: Scott Burleigh of the IPNGROUP, Angela Hennessy of the
   Laboratory for Telecommunications Sciences, Amy Alford and Cherita
   Corbett of the Johns Hopkins University Applied Physics Laboratory
   (JHU/APL), and Angela Dalton of AMD Research.

   Additionally, Benjamin Kaduk of Akamai Technologies provided a
   detailed technical review that resulted in a stronger and more
   precise specification.

Authors' Addresses

   Edward J. Birrane, III
   The Johns Hopkins University Applied Physics Laboratory
   11100 Johns Hopkins Rd.
   Laurel, MD 20723
   United States of America

   Phone: +1 443 778 7423
   Email: Edward.Birrane@jhuapl.edu

   Kenneth McKeever
   The Johns Hopkins University Applied Physics Laboratory
   11100 Johns Hopkins Rd.
   Laurel, MD 20723
   United States of America

   Phone: +1 443 778 2237
   Email: Ken.McKeever@jhuapl.edu