title: Encrypted Payloads in SUIT Manifests abbrev: Encrypted Payloads in SUIT Manifests docname: draft-ietf-suit-firmware-encryption-22 category: std
ipr: trust200902 area: Security workgroup: SUIT keyword: Internet-Draft
stand_alone: yes pi: rfcedstyle: yes toc: yes tocindent: yes sortrefs: yes symrefs: yes strict: yes comments: yes inline: yes text-list-symbols: -o*+ docmapping: yes toc_levels: 4
name: Hannes Tschofenig
org: University of Applied Sciences Bonn-Rhein-Sieg
abbrev: H-BRS
email: [email protected]
-
ins: R. Housley name: Russ Housley organization: Vigil Security, LLC abbrev: Vigil Security email: [email protected] -
ins: B. Moran name: Brendan Moran organization: Arm Limited email: [email protected]
-
ins: D. Brown name: David Brown organization: Linaro email: [email protected]
-
ins: K. Takayama name: Ken Takayama organization: SECOM CO., LTD. email: [email protected]
normative: RFC2119: RFC3394: RFC9052: RFC9053: RFC8174: I-D.ietf-suit-manifest: RFC9459: I-D.ietf-suit-trust-domains:
informative: RFC9019: RFC9124: RFC5869: RFC8937: RFC5652: RFC5280: RFC5869: iana-suit: author: org: Internet Assigned Numbers Authority title: IANA SUIT Manifest Registry date: 2023 target: TBD ROP: author: org: Wikipedia title: Return-Oriented Programming date: 06.03.2023 target: https://en.wikipedia.org/wiki/Return-oriented_programming SP800-56: author: org: NIST title: Recommendation for Pair-Wise Key Establishment Schemes Using Discrete Logarithm Cryptography, NIST Special Publication 800-56A Revision 3 date: April 2018 target: http://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.800-56Ar3.pdf
--- abstract
This document specifies techniques for encrypting software, firmware, machine learning models, and personalization data by utilizing the IETF SUIT manifest. Key agreement is provided by ephemeral-static (ES) Diffie-Hellman (DH) and AES Key Wrap (AES-KW). ES-DH uses public key cryptography while AES-KW uses a pre-shared key. Encryption of the plaintext is accomplished with conventional symmetric key cryptography.
--- middle
Vulnerabilities in Internet of Things (IoT) devices have highlighted the need for a reliable and secure firmware update mechanism, especially for constrained devices. To protect firmware images, the SUIT manifest format was developed {{I-D.ietf-suit-manifest}}. A manifest is a bundle of metadata about the firmware for an IoT device, where to find the firmware, and the devices to which it applies. {{RFC9124}} outlines the necessary information a SUIT manifest has to provide. In addition to protecting against modification via digital signatures or message authentication codes, the format can also offer confidentiality.
Encryption prevents third parties, including attackers, from accessing the payload. Attackers often require detailed knowledge of a binary, such as a firmware image, to launch successful attacks. For instance, return-oriented programming (ROP) {{ROP}} requires access to the binary, and encryption makes writing exploits significantly more difficult. Beyond ensuring the confidentiality of the binary itself, protecting the confidentiality of the source code will also be necessary to prevent reverse engineering and reproduction of the firmware.
While the initial motivation for this document was firmware encryption, the use of SUIT manifests has expanded to cover other scenarios requiring integrity and confidentiality protection, such as:
- software packages,
- personalization data,
- configuration data, and
- machine learning models.
Thus, the term "payload" is used generically to refer to all these objects.
The payload is encrypted using a symmetric content encryption key, which can be established through various mechanisms. This document defines two content key distribution methods for use with the IETF SUIT manifest:
- Ephemeral-Static (ES) Diffie-Hellman (DH), and
- AES Key Wrap (AES-KW).
The first method relies on asymmetric cryptography, while the second uses symmetric cryptography.
Our design aims to reduce the number of content key distribution methods for payload encryption, thereby increasing interoperability between different SUIT manifest parser implementations.
The goal of this specification is to protect payloads both during end-to-end transport (from the distribution system to the device) and at rest when stored on the device. Constrained devices often employ eXecute In Place (XIP), a method of executing code directly from flash memory rather than loading it into RAM. Many of these devices lack hardware-based, on-the-fly decryption for code stored in flash memory, which may require decrypting and storing firmware images in on-chip flash before execution. However, we expect hardware-based, on-the-fly decryption to become more common in the future, enhancing confidentiality at rest.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 {{RFC2119}} {{RFC8174}} when, and only when, they appear in all capitals, as shown here.
This document assumes familiarity with the IETF SUIT manifest {{I-D.ietf-suit-manifest}}, the SUIT information model {{RFC9124}}, and the SUIT architecture {{RFC9019}}.
The following abbreviations are used in this document:
- Key Wrap (KW), defined in {{RFC3394}} (for use with AES)
- Key-Encryption Key (KEK) {{RFC3394}}
- Content-Encryption Key (CEK) {{RFC5652}}
- Ephemeral-Static (ES) Diffie-Hellman (DH) {{RFC9052}}
- Authenticated Encryption with Associated Data (AEAD)
- Execute in Place (XIP)
The terms sender and recipient have the following meaning:
- Sender: Entity that sends an encrypted payload.
- Recipient: Entity that receives an encrypted payload.
Additionally, we introduce the term "distribution system" (or distributor) to refer to an entity that knows the recipients of payloads. It is important to note that the distribution system is far more than a file server. For use of encryption, the distribution system either knows the public key of the recipient (for ES-DH), or the KEK (for AES-KW).
The author, which is responsible for creating the payload, does not know the recipients. The author may, for example, be a developer building a firmware image.
The author and the distribution system are logical roles. In some deployments these roles are separated in different physical entities and in others they are co-located.
{{RFC9019}} outlines the architecture for distributing payloads and manifests from an author to devices. However, it does not cover payload encryption in detail. This document extends that architecture to support encryption, as illustrated in {{arch-fig}}.
To encrypt a payload, it is essential to know the recipient. For AES-KW, the Key Encryption Key (KEK) must be known, and for ES-DH, the sender needs access to the recipient's public key. This public key and its associated parameters may be found in the recipient's X.509 certificate {{RFC5280}}. For authentication and integrity protection, recipients must be provisioned with a trust anchor when the manifest is protected by a digital signature. If a MAC is used for manifest protection, a symmetric key must be shared between the recipient and the sender.
With encryption, the author cannot simply create and sign a manifest for the payload, as the recipients are often unknown. Therefore, the author must collaborate with the distribution system. The degree of this collaboration is discussed below.
The primary purpose of encryption is to protect against adversaries along the path between the distribution system and the device. There is also a risk that adversaries may extract the decrypted firmware image from the device itself. Consequently, the device must be safeguarded against physical attacks, though such countermeasures are typically beyond the scope of IETF specifications.
Note: It is assumed that the communication between the author and the distribution system is secure. For example, the author might upload the manifest and firmware image to the distribution system using a secure HTTPS REST API.
+----------+
| Device | +----------+
| 1 |<--+ | Author |
| | | +----------+
+----------+ | |
| | Payload +
| | Manifest
| v
+----------+ | +--------------+
| Device | | Payload + Manifest | Distribution |
| 2 |<--+------------------------| System |
| | | +--------------+
+----------+ |
|
... |
|
+----------+ |
| Device | |
| n |<--+
| |
+----------+
{: #arch-fig title="Architecture for the distribution of Encrypted Payloads."}
When the author delegates encryption rights to the distributor, two models are possible:
-
Replacing the COSE_Encrypt and Re-signing the Manifest: The distributor replaces the COSE_Encrypt structure in the manifest and then signs the manifest again. However, since the COSE_Encrypt structure is within a signed container, this presents a challenge: replacing COSE_Encrypt alters the digest of the manifest, thereby invalidating the signature. As a result, the distributor must be able to sign the new manifest. If this is the case, the distributor gains the authority to construct and sign manifests, effectively allowing them to sign code and giving them full control over the recipient. Distributors typically perform re-encryption online to manage large numbers of devices efficiently, which prevents air-gapping the signing operations. This has implications for the recommendations in {{Section 4.3.17 of RFC9124}}. Despite these issues, this model represents the current standard practice for IoT firmware updates.
-
Two-Layer Manifest System: The distributor creates a new manifest that overrides the COSE_Encrypt using the dependency system defined in {{I-D.ietf-suit-trust-domains}}. This method introduces additional overhead, including one more signature verification, one extra manifest, and the need for extra mechanisms on the recipient side to handle dependency processing. While this adds complexity, it also enhances security.
These two models offer different threat profiles for the distributor. If the distributor is limited to encryption rights, an attacker who breaches the distributor can only launch a limited attack by encrypting a modified binary. However, recipients will detect the attack during the image digest check and immediately revert to the correct image.
It is RECOMMENDED that distributors adopt the two-layer manifest approach to distribute content encryption keys without re-signing the manifest, despite the added complexity and the increased number of signature verifications required on the recipient side.
This specification introduces a new extension to the SUIT_Parameters structure.
The SUIT_Encryption_Info structure (referred to as suit-parameter-encryption-info in {{parameter-fig}}) contains the content key distribution information. The details of the SUIT_Encryption_Info structure are provided in {{AES-KW}} (for AES-KW) and {{ES-DH}} (for ES-DH).
Once a CEK is available, the steps outlined in {{content-enc}} apply to both content key distribution methods described in this section.
The SUIT_Encryption_Info structure is carried within either the suit-directive-override-parameters or suit-directive-set-parameters used in the "Directive Write" and "Directive Copy" operations. An implementation conforming to this specification must support these two parameters.
Interoperability requirements for content key distribution methods vary: since a device will typically support only one of the two specified methods, the distribution system needs to know which method is supported. Limiting a constrained device to a single content key distribution method also helps reduce code size.
SUIT_Parameters //= (suit-parameter-encryption-info
=> bstr .cbor SUIT_Encryption_Info)
suit-parameter-encryption-info = TBD19
{: #parameter-fig title="CDDL of the SUIT_Parameters Extension."}
RFC Editor's Note (TBD19): The value for the suit-parameter-encryption-info parameter is set to 19, as the proposed value.
This specification extends the following directives:
-
Directive Write (suit-directive-write): Used to decrypt the content specified by suit-parameter-content using suit-parameter-encryption-info.
-
Directive Copy (suit-directive-copy): Used to decrypt the content of the component specified by suit-parameter-source-component using suit-parameter-encryption-info.
Examples of these two directives are provided below, focusing on the essential aspects. A complete example for AES Key Wrap with the Fetch and Copy directives can be found in {{example-AES-KW-copy}}, while an example illustrating the Write directive is shown in {{example-AES-KW-write}}.
{{encryption-info-consumed-with-write}} illustrates the Directive Write. The encrypted payload specified by parameter-content, represented as h'EA1...CED' in the example, is decrypted using the SUIT_Encryption_Info structure referenced by parameter-encryption-info, i.e., h'D86...1F0'. The resulting plaintext payload is then stored in component #0.
/ directive-override-parameters / 20, {
/ parameter-content / 18: h'EA1...CED',
/ parameter-encryption-info / TBD19: h'D86...1F0'
},
/ directive-write / 18, 15
{: #encryption-info-consumed-with-write title="Example showing the extended suit-directive-write."}
RFC Editor's Note (TBD19): The value for the parameter-encryption-info parameter is set to 19, as the proposed value.
{{encryption-info-consumed-with-copy}} illustrates the Directive Copy. In this example the encrypted payload is found at the URI indicated by the parameter-uri, i.e., "http://example.com/encrypted.bin". The encrypted payload will be downloaded and stored in component #1. Then, the information in the SUIT_Encryption_Info structure referred to by parameter-encryption-info, i.e., h'D86...1F0', will be used to decrypt the content in component #1 and the resulting plaintext payload will be stored into component #0.
/ directive-set-component-index / 12, 1,
/ directive-override-parameters / 20, {
/ parameter-uri / 21: "http://example.com/encrypted.bin",
},
/ directive-fetch / 21, 15,
/ directive-set-component-index / 12, 0,
/ directive-override-parameters / 20, {
/ parameter-encryption-info / TBD19: h'D86...1F0',
/ parameter-source-component / 22: 1
},
/ directive-copy / 22, 15
{: #encryption-info-consumed-with-copy title="Example showing the extended suit-directive-copy."}
RFC Editor's Note (TBD19): The value for the suit-parameter-encryption-info parameter is set to 19, as the proposed value.
The payload to be encrypted may be detached and, in that case, it is not covered by the digital signature or the MAC protecting the manifest. (To be more precise, the suit-authentication-wrapper found in the envelope contains a digest of the manifest in the SUIT Digest Container.)
The lack of authentication and integrity protection of the payload is particularly a concern when a cipher without integrity protection is used.
To provide authentication and integrity protection of the payload in the detached payload case a SUIT Digest Container with the hash of the encrypted and/or plaintext payload MUST be included in the manifest. See suit-parameter-image-digest parameter in {{Section 8.4.8.6 of I-D.ietf-suit-manifest}}.
Once a CEK is available, the steps described in {{content-enc}} are applicable. These steps apply to both content key distribution methods.
The following sub-sections describe two content key distribution methods: AES Key Wrap (AES-KW) and Ephemeral-Static Diffie-Hellman (ES-DH). While many other methods are specified in the literature and supported by COSE, AES-KW and ES-DH were chosen for their widespread use in the market today. They were selected for their maturity, differing security properties, and strong interoperability.
Both content key distribution methods require the CEKs to be randomly generated. The guidelines for random number generation in {{RFC8937}} MUST be followed.
When sending an encrypted payload to multiple recipients, various deployment options are available. The following notation is used to explain these options:
- KEK[R1, S] refers to a KEK shared between recipient R1 and
the sender S.
- CEK[R1, S] refers to a CEK shared between R1 and S.
- CEK[*, S] or KEK[*, S] are used when a single CEK or a single
KEK is shared with all authorized recipients by a given sender
S in a certain context.
- ENC(plaintext, k) refers to the encryption of plaintext with
a key k.
The AES Key Wrap (AES-KW) algorithm, as described in {{RFC3394}}, is used to encrypt a randomly generated content-encryption key (CEK) with a pre-shared key-encryption key (KEK). The COSE conventions for using AES-KW are specified in {{Section 8.5.2 of RFC9052}} and in {{Section 6.2.1 of RFC9053}}. The encrypted CEK is carried within the COSE_recipient structure , which includes the necessary information for AES-KW. The COSE_recipient structure, a substructure of COSE_Encrypt, contains the CEK encrypted by the KEK.
To ensure high security when using AES Key Wrap, it is important that the KEK is of high entropy and that implementations protect the KEK from disclosure. A compromised KEK could expose all data encrypted with it, including binaries and configuration data.
The COSE_Encrypt structure conveys the information needed to encrypt the payload, including details such as the algorithm and IV. Even though the payload may be conveyed as detached content, the encryption information is still embedded in the COSE_Encrypt.ciphertext structure.
There are three deployment options for use with AES Key Wrap for payload encryption:
- If all recipients (typically of the same product family) share the same KEK, a single COSE_recipient structure contains the encrypted CEK. The sender executes the following steps:
1. Fetch KEK[*, S]
2. Generate CEK
3. ENC(CEK, KEK)
4. ENC(payload, CEK)
This deployment option is strongly discouraged. An attacker gaining access to the KEK will be able to encrypt and send payloads to all recipients configured to use this KEK.
- If recipients have different KEKs, then multiple COSE_recipient structures are included but only a single CEK is used. Each COSE_recipient structure contains the CEK encrypted with the KEKs appropriate for a given recipient. The benefit of this approach is that the payload is encrypted only once with a CEK while there is no sharing of the KEK across recipients. Hence, authorized recipients still use their individual KEK to decrypt the CEK and to subsequently obtain the plaintext. The steps taken by the sender are:
1. Generate CEK
2. for i=1 to n
{
2a. Fetch KEK[Ri, S]
2b. ENC(CEK, KEK[Ri, S])
}
3. ENC(payload, CEK)
- The third option is to use different CEKs encrypted with KEKs of authorized recipients. This approach is appropriate when no benefits can be gained from encrypting and transmitting payloads only once. Assume there are n recipients with their unique KEKs - KEK[R1, S], ..., KEK[Rn, S] and unique CEKs. The sender needs to execute the following steps:
1. for i=1 to n
{
1a. Fetch KEK[Ri, S]
1b. Generate CEK[Ri, S]
1c. ENC(CEK[Ri, S], KEK[Ri, S])
1d. ENC(payload, CEK[Ri, S])
2. }
The CDDL for the AES-KW binary is shown in {{cddl-aeskw}}. empty_or_serialized_map and header_map are structures defined in {{RFC9052}}.
{::include-fold cddls/aeskw.cddl}
{: #cddl-aeskw title="CDDL for AES-KW-based Content Key Distribution"}
Note that the AES-KW algorithm, as defined in {{Section 2.2.3.1 of RFC3394}}, does not have public parameters that vary on a per-invocation basis. Hence, the protected header in the COSE_recipient structure is a byte string of zero length.
Ephemeral-Static Diffie-Hellman (ES-DH) is a public key encryption scheme that enables encryption using the recipient's public key. There are several variations of this scheme; this document adopts the version specified in {{Section 8.5.5 of RFC9052}}.
The structure is composed of two layers:
-
Layer 0: Contains content encrypted with a Content Encryption Key (CEK). The content may be provided separately.
-
Layer 1: Uses the AES Key Wrap (AES-KW) algorithm to encrypt the randomly generated CEK with a Key Encryption Key (KEK) derived via ES-DH. The resulting symmetric key is processed through an HKDF-based key derivation function {{RFC5869}}.
This two-layer structure combines ES-DH with AES-KW and HKDF, referred to as ECDH-ES + AES-KW. An example can be found in {{esdh-aesgcm-example}}.
Another variant of the ES-DH algorithm, called ECDH-ES + HKDF, does not utilize AES Key Wrap. However, this version is not covered in this document.
This approach supports only two deployment options, as it assumes that each recipient is always equipped with a device-specific public/private key pair.
- When a sender transmits a payload to multiple recipients, all recipients receive the same encrypted payload, meaning the same CEK is used to encrypt the content. For each recipient, a separate COSE_recipient structure is used, which contains the CEK encrypted with the recipient-specific KEK. To derive the KEK, each COSE_recipient structure includes a COSE_recipient_inner structure that carries the sender's ephemeral key and an identifier for the recipient's public key.
The steps taken by the sender are:
1. Generate CEK
2. for i=1 to n
{
2a. Generate KEK[Ri, S] using ES-DH
2b. ENC(CEK, KEK[Ri, S])
}
3. ENC(payload,CEK)
- The alternative is to encrypt the payload with a unique CEK for each recipient, resulting in multiple manifests. This approach is useful when payloads contain device-specific information. In this case, the encryption operation becomes ENC(payload_i, CEK[Ri, S]) where each recipient Ri receives a unique CEK. Assume that KEK[R1, S],..., KEK[Rn, S] have been generated for the recipients using ES-DH. The sender must then follow these steps:
1. for i=1 to n
{
1a. Generate KEK[Ri, S] using ES-DH
1b. Generate CEK[Ri, S]
1c. ENC(CEK[Ri, S], KEK[Ri, S])
1d. ENC(payload, CEK[Ri, S])
}
The CDDL for the ECDH-ES+AES-KW binary is provided in {{cddl-esdh}}. Only the essential parameters are included. The structures empty_or_serialized_map and header_map are defined in {{RFC9052}}.
{::include-fold cddls/esdh_aeskw.cddl}
{: #cddl-esdh title="CDDL for ES-DH-based Content Key Distribution"}
See {{content-enc}} for a description on how to encrypt the payload.
The context information structure ensures that the derived keying material is "bound" to the specific context of the transaction. This specification reuses the structure defined in {{Section 5.2 of RFC9053}}, with modifications to fit the current use case.
The following elements are bound to the context:
- the protocol employing the key-derivation method,
- information about the utilized AES Key Wrap algorithm, and the key length.
- the protected header field, which contains the content key encryption algorithm.
The sender and recipient identities are left empty.
The following fields in {{cddl-context-info}} require an explanation:
-
The COSE_KDF_Context.AlgorithmID field MUST contain the identifier for the AES Key Wrap algorithm being used. This specification uses the following values: A128KW (value -3), A192KW (value -4), or A256KW (value -5)
-
The COSE_KDF_Context.SuppPubInfo.keyDataLength field MUST specify the key length, in bits, corresponding to the algorithm in the AlgorithmID field. For A128KW the value is 128, for A192KW the value is 192, and for A256KW the value 256.
-
The COSE_KDF_Context.SuppPubInfo.other field captures the protocol that uses the ES-DH content key distribution algorithm. It MUST be set to the constant string "SUIT Payload Encryption".
-
The COSE_KDF_Context.SuppPubInfo.protected field MUST contain the serialized content of the recipient_header_map_esdh field, which contains (among other elements) the identifier of the content key distribution method.
{::include-fold cddls/kdf-context.cddl}
{: #cddl-context-info title="CDDL for COSE_KDF_Context Structure"}
The HKDF-based key derivation function MAY contain a salt value, as described in {{Section 5.1 of RFC9053}}. This optional value influences the key generation process, though this specification does not require the use of a salt. If the salt is public and included in the message, the "salt" algorithm header parameter MUST be used. The salt adds extra randomness to the KDF context. When the salt is transmitted via the "salt" algorithm header parameter, the receiver MUST be capable of processing it and MUST pass it into the key derivation function. For more details on salt usage, refer to {{RFC5869}} and NIST SP800-56 {{SP800-56}}.
Profiles of this specification MAY define an extended version of the context information structure or MAY employ a different context information structure.
This section summarizes the steps involved in content encryption, applicable to both content key distribution methods.
When using AEAD ciphers, such as AES-GCM or ChaCha20/Poly1305, the COSE specification requires a consistent byte stream to create the authenticated data structure. This structure is illustrated in {{cddl-enc-aeskw}} and defined in {{Section 5.3 of RFC9052}}.
Enc_structure = [
context : "Encrypt",
protected : empty_or_serialized_map,
external_aad : bstr
]
{: #cddl-enc-aeskw title="CDDL for Enc_structure Data Structure"}
This Enc_structure must be populated as follows:
-
The protected field in the Enc_structure from {{cddl-enc-aeskw}} refers to the content of the protected field in the COSE_Encrypt structure.
-
The value of external_aad MUST be set to a zero-length byte string, represented as h'' in diagnostic notation and encoded as 0x40.
Some ciphers, such as AES-CTR and AES-CBC, provide confidentiality without integrity protection (see {{RFC9459}}). For these ciphers, the Enc_structure shown in {{cddl-enc-aeskw}} cannot be used, as the Additional Authenticated Data (AAD) byte string is only applicable to AEAD ciphers. Therefore, the AAD structure is not passed to the API for these ciphers, and the protected header in the SUIT_Encryption_Info structure MUST be a zero-length byte string.
AES-CTR and AES-CBC are discussed in separate sub-sections below and {{aes-ctr-fig}} and {{aes-cbc-fig}} use the following abbreviations:
- Pi = Plaintext blocks
- Ci = Ciphertext blocks
- E = Encryption function
- k = Symmetric key
- ⊕ = XOR operation
AES-GCM is an AEAD cipher and provides confidentiality and integrity protection.
Examples in this section use the following parameters:
- Algorithm for payload encryption: AES-GCM-128
- k: h'15F785B5C931414411B4B71373A9C0F7'
- IV: h'F14AAB9D81D51F7AD943FE87AF4F70CD'
- Plaintext: "This is a real firmware image."
- in hex: 546869732069732061207265616C206669726D7761726520696D6167652E
This example uses the following parameters:
- Algorithm id for key wrap: A128KW
- KEK COSE_Key (Secret Key):
- kty: Symmetric
- k: 'aaaaaaaaaaaaaaaa'
- kid: 'kid-1'
The COSE_Encrypt structure, in hex format, is (with a line break inserted):
{::include-fold examples/suit-encryption-info-aes-kw-aes-gcm.hex}
The resulting COSE_Encrypt structure in a diagnostic format is shown in {{aeskw-aesgcm-example}}.
{::include-fold examples/suit-encryption-info-aes-kw-aes-gcm.diag}
{: #aeskw-aesgcm-example title="COSE_Encrypt Example for AES Key Wrap"}
The encrypted payload (with a line feed added) was:
{::include-fold examples/encrypted-payload-aes-kw-aes-gcm.hex}
This example uses the following parameters:
- Algorithm for content key distribution: ECDH-ES + A128KW
- KEK COSE_Key (Receiver's Private Key):
- kty: EC2
- crv: P-256
- x: h'5886CD61DD875862E5AAA820E7A15274C968A9BC96048DDCACE32F50C3651BA3'
- y: h'9EED8125E932CD60C0EAD3650D0A485CF726D378D1B016ED4298B2961E258F1B'
- d: h'60FE6DD6D85D5740A5349B6F91267EEAC5BA81B8CB53EE249E4B4EB102C476B3'
- kid: 'kid-2'
- KDF Context
- Algorithm ID: -3 (A128KW)
- SuppPubInfo
- keyDataLength: 128
- protected: { / alg / 1: -29 / ECDH-ES+A128KW / }
- other: 'SUIT Payload Encryption'
The COSE_Encrypt structure, in hex format, is (with a line break inserted):
{::include-fold examples/suit-encryption-info-es-ecdh-aes-gcm.hex}
The resulting COSE_Encrypt structure in a diagnostic format is shown in {{esdh-aesgcm-example}}.
{::include-fold examples/suit-encryption-info-es-ecdh-aes-gcm.diag}
{: #esdh-aesgcm-example title="COSE_Encrypt Example for ES-DH"}
The encrypted payload (with a line feed added) was:
{::include-fold examples/encrypted-payload-es-ecdh-aes-gcm.hex}
AES-CTR is a non-AEAD cipher that provides confidentiality but lacks integrity protection. Unlike AES-CBC, AES-CTR uses an IV per block, as shown in {{aes-ctr-fig}}. Hence, when an image is encrypted using AES-CTR-128 or AES-CTR-256, the IV MUST start with zero (0) and MUST be incremented by one for each 16-byte plaintext block within the entire slot.
In our example, we assume the slot size of a specific flash controller on an IoT device is 64 KiB, the sector size 4096 bytes (4 KiB) and an AES plaintext block size of 16 bytes, the IVs range from 0 to 255 in the first sector, and 16 * 256 IVs are required for the remaining sectors in the slot.
IV1 IV2
| |
| |
| |
+-------+ +-------+
| | | |
| | | |
k--| E | k--| E |
| | | |
+-------+ +-------+
| |
P1---⊕ P2---⊕
| |
| |
C1 C2
{: #aes-ctr-fig title="AES-CTR Operation"}
Note: The abbreviations shown in {{aes-ctr-fig}} are described in {{content-enc}}.
Examples in this section use the following parameters:
- Algorithm for payload encryption: AES-CTR-128
- k: h'261DE6165070FB8951EC5D7B92A065FE'
- IV: h'DAE613B2E0DC55F4322BE38BDBA9DC68'
- Plaintext: "This is a real firmware image."
- in hex: 546869732069732061207265616C206669726D7761726520696D6167652E
This example uses the following parameters:
- Algorithm id for key wrap: A128KW
- KEK COSE_Key (Secret Key):
- kty: Symmetric
- k: 'aaaaaaaaaaaaaaaa'
- kid: 'kid-1'
The COSE_Encrypt structure, in hex format, is (with a line break inserted):
{::include-fold examples/suit-encryption-info-aes-kw-aes-ctr.hex}
The resulting COSE_Encrypt structure in a diagnostic format is shown in {{aeskw-aesctr-example}}.
{::include-fold examples/suit-encryption-info-aes-kw-aes-ctr.diag}
{: #aeskw-aesctr-example title="COSE_Encrypt Example for AES Key Wrap"}
The encrypted payload (with a line feed added) was:
{::include-fold examples/encrypted-payload-aes-kw-aes-ctr.hex}
This example uses the following parameters:
- Algorithm for content key distribution: ECDH-ES + A128KW
- KEK COSE_Key (Receiver's Private Key):
- kty: EC2
- crv: P-256
- x: h'5886CD61DD875862E5AAA820E7A15274C968A9BC96048DDCACE32F50C3651BA3'
- y: h'9EED8125E932CD60C0EAD3650D0A485CF726D378D1B016ED4298B2961E258F1B'
- d: h'60FE6DD6D85D5740A5349B6F91267EEAC5BA81B8CB53EE249E4B4EB102C476B3'
- kid: 'kid-2'
- KDF Context
- Algorithm ID: -3 (A128KW)
- SuppPubInfo
- keyDataLength: 128
- protected: { / alg / 1: -29 / ECDH-ES+A128KW / }
- other: 'SUIT Payload Encryption'
The COSE_Encrypt structure, in hex format, is (with a line break inserted):
{::include-fold examples/suit-encryption-info-es-ecdh-aes-ctr.hex}
The resulting COSE_Encrypt structure in a diagnostic format is shown in {{esdh-aesctr-example}}.
{::include-fold examples/suit-encryption-info-es-ecdh-aes-ctr.diag}
{: #esdh-aesctr-example title="COSE_Encrypt Example for ES-DH"}
The encrypted payload (with a line feed added) was:
{::include-fold examples/encrypted-payload-es-ecdh-aes-ctr.hex}
AES-CBC is a non-AEAD cipher that provides confidentiality but does not offer integrity protection. In AES-CBC, a single IV is used to encrypt the firmware belonging to a single sector, as individual AES blocks are chained together, as illustrated in {{aes-cbc-fig}}. The numbering of sectors in a slot start with zero (0) and increase by one with every sector till the end of the slot is reached. The IV follows this numbering.
For example, let us assume the slot size of a specific flash controller on an IoT device is 64 KiB, the sector size 4096 bytes (4 KiB) and AES-128-CBC uses an AES-block size of 128 bit (16 bytes). Hence, sector 0 needs 4096/16=256 AES-128-CBC operations using IV 0. If the firmware image occupies the entire slot, it will contain 16 sectors, corresponding to IVs ranging from 0 to 15.
P1 P2
| |
IV---⊕ +--------⊕
| | |
| | |
+-------+ | +-------+
| | | | |
| | | | |
k--| E | | k--| E |
| | | | |
+-------+ | +-------+
| | |
+-----+ |
| |
| |
C1 C2
{: #aes-cbc-fig title="AES-CBC Operation"}
Note: The abbreviations shown in {{aes-cbc-fig}} are described in {{content-enc}}.
Examples in this section use the following parameters:
- Algorithm for payload encryption: AES-CBC-128
- k: h'627FCF0EA82C967D5ED8981EB325F303'
- IV: h'93702C81590F845D9EC866CCAC767BD1'
- Plaintext: "This is a real firmware image."
- in hex: 546869732069732061207265616C206669726D7761726520696D6167652E
This example uses the following parameters:
- Algorithm id for key wrap: A128KW
- KEK COSE_Key (Secret Key):
- kty: Symmetric
- k: 'aaaaaaaaaaaaaaaa'
- kid: 'kid-1'
The COSE_Encrypt structure, in hex format, is (with a line break inserted):
{::include-fold examples/suit-encryption-info-aes-kw-aes-cbc.hex}
The resulting COSE_Encrypt structure in a diagnostic format is shown in {{aeskw-aescbc-example}}.
{::include-fold examples/suit-encryption-info-aes-kw-aes-cbc.diag}
{: #aeskw-aescbc-example title="COSE_Encrypt Example for AES Key Wrap"}
The encrypted payload (with a line feed added) was:
{::include-fold examples/encrypted-payload-aes-kw-aes-cbc.hex}
This example uses the following parameters:
- Algorithm for content key distribution: ECDH-ES + A128KW
- KEK COSE_Key (Receiver's Private Key):
- kty: EC2
- crv: P-256
- x: h'5886CD61DD875862E5AAA820E7A15274C968A9BC96048DDCACE32F50C3651BA3'
- y: h'9EED8125E932CD60C0EAD3650D0A485CF726D378D1B016ED4298B2961E258F1B'
- d: h'60FE6DD6D85D5740A5349B6F91267EEAC5BA81B8CB53EE249E4B4EB102C476B3'
- kid: 'kid-2'
- KDF Context
- Algorithm ID: -3 (A128KW)
- SuppPubInfo
- keyDataLength: 128
- protected: { / alg / 1: -29 / ECDH-ES+A128KW / }
- other: 'SUIT Payload Encryption'
The COSE_Encrypt structure, in hex format, is (with a line break inserted):
{::include-fold examples/suit-encryption-info-es-ecdh-aes-cbc.hex}
The resulting COSE_Encrypt structure in a diagnostic format is shown in {{esdh-aescbc-example}}.
{::include-fold examples/suit-encryption-info-es-ecdh-aes-cbc.diag}
{: #esdh-aescbc-example title="COSE_Encrypt Example for ES-DH"}
The encrypted payload (with a line feed added) was:
{::include-fold examples/encrypted-payload-es-ecdh-aes-cbc.hex}
In addition to suit-condition-image-match (see {{Section 8.4.9.2 of I-D.ietf-suit-manifest}}), AEAD algorithms used for content encryption provides another way to validate the integrity of components. This section provides a guideline to construct secure but not redundant SUIT Manifest for encrypted payloads.
This sub-section explains three ways to validate the integrity of payloads.
The suit-condition-image-match on the plaintext payload is used after decryption. An example command sequence is shown in {{figure-image-match-after-decryption}}.
/ directive-set-component-index / 12, 1,
/ directive-override-parameters / 20, {
/ parameter-uri / 21: "http://example.com/encrypted.bin"
},
/ directive-fetch / 21, 15,
/ directive-set-component-index / 12, 0,
/ directive-override-parameters / 20, {
/ parameter-image-digest / 3: << {
/ algorithm-id: / -16 / SHA256 /,
/ digest-bytes: / h'3B1...92A' / digest of plaintext payload /
} >>,
/ parameter-image-size / 14: 30 / size of plaintext payload /,
/ parameter-encryption-info / TBD19: h'369...50F',
/ parameter-source-component / 22: 1
},
/ directive-copy / 22, 15,
/ condition-image-match / 3, 15 / check decrypted payload integrity /,
{: #figure-image-match-after-decryption title="Check Image Match After Decryption"}
RFC Editor's Note (TBD19): The value for the suit-parameter-encryption-info parameter is set to 19, as the proposed value.
The suit-condition-image-match can also be applied on encrypted payloads before decryption takes place. An example command sequence is shown in {{figure-image-match-before-decryption}}.
This option mitigates battery exhaustion attacks discussed in {{sec-cons}}.
/ directive-set-component-index / 12, 1,
/ directive-override-parameters / 20, {
/ parameter-image-digest / 3: << {
/ algorithm-id: / -16 / SHA256 /,
/ digest-bytes: / h'8B4...D34' / digest of encrypted payload /
} >>,
/ parameter-image-size / 14: 30 / size of encrypted payload /,
/ parameter-uri / 21: "http://example.com/encrypted.bin"
},
/ directive-fetch / 21, 15,
/ condition-image-match / 3, 15 / check decrypted payload integrity /,
/ directive-set-component-index / 12, 0,
/ directive-override-parameters / 20, {
/ parameter-encryption-info / TBD19: h'D86...1F0',
/ parameter-source-component / 22: 1
},
/ directive-copy / 22, 15,
{: #figure-image-match-before-decryption title="Check Image Match Before Decryption"}
RFC Editor's Note (TBD19): The value for the suit-parameter-encryption-info parameter is set to 19, as the proposed value.
AEAD algorithms, such as AES-GCM and ChaCha20/Poly1305, verify the integrity of the encrypted concent.
This subsection offers guidelines for validating the integrity of payloads within the SUIT manifest. The decision tree in {{payload-integrity-decision-tree}} illustrates the process for establishing payload integrity.
+------------------------------------------------+
| Q1. Payload Delivery |
+-+--------------------------------------------+-+
| |
| in Content others |
| v
| +--------------------------------+
| | Q2. Mitigate Battery |
| | Exhaustion Attacks |
| +-+----------------------------+-+
| | |
| | No Yes |
| v |
| +-----------------+ |
| | Q3. AEAD cipher | |
| +-+-------------+-+ |
| | | |
| | Yes No | |
v v v v
.+------+. .-----+-----. .----------+.
| NOT | | AFTER | | BEFORE |
| Required | | Decryption | | Decryption |
'--------' '-----------' '-----------'
{: #payload-integrity-decision-tree title="Decision Tree: Validating the Payload"}
There are three questions to ask:
-
Q1. How does the recipient receive the encrypted payload? If the encrypted payload is part of an integrated payload, its integrity is already validated by the suit-authentication-wrapper. Hence, no additional integrity check is necessary.
-
Q2. Does the sender wish to mitigate battery exhaustion attacks? If so, the encrypted payload must be validated before decryption.
-
Q3. Is the payload encrypted with an AEAD cipher? If yes, no additional integrity check is required, as the recipient verifies the payload's integrity during decryption. If no, integrity validation can occur either before or after decryption; however, validating integrity before decryption is RECOMMENDED.
Embedded devices come in many forms, and the market is both large and fragmented. As a result, some implementations and deployments may adopt firmware update procedures that differ from the descriptions provided here. On a positive note, the SUIT manifest accommodates various deployment scenarios, thanks to the "scripting" functionality offered by its commands.
This section specifically addresses firmware images on microcontrollers and does not pertain to generic software, configuration data, or machine learning models. The differences arise from two main aspects:
-
Use of Flash Memory: Flash memory in microcontrollers is a type of non-volatile memory that erases data in larger units called blocks, pages, or sectors, and rewrites data at the byte level (often 4 bytes) or larger units. Furthermore, flash memory is segmented into different regions, storing the bootloader, various versions of firmware images (in designated slots), and configuration data. An example layout of a microcontroller flash area is illustrated in {{image-layout}}.
-
Microcontroller Design: Code on microcontrollers typically cannot be executed from arbitrary locations in flash memory without additional software development and design efforts. Consequently, developers often compile firmware so that the bootloader can execute code from a specific location in flash memory, commonly referred to as the "primary slot."
Once the encrypted firmware image is transferred to the device, it is usually stored in a dedicated area known as the "secondary slot."
During the next boot, the bootloader detects the new firmware image and begins decrypting it sector by sector, swapping it with the image located in the primary slot. This method of swapping the newly downloaded image with the previously valid one requires two slots, allowing for a rollback if the new firmware fails to boot, thereby enhancing the robustness of the firmware update process.
The swap occurs only after verifying the signature on the plaintext. It is important to note that the plaintext firmware image is available in the primary slot only after the swap is completed, unless "dummy decrypt" is used to compute the hash over the plaintext prior to executing the decryption during the swap. In this context, dummy decryption refers to decrypting the firmware image in the secondary slot sector by sector while computing a rolling hash over the resulting plaintext (also sector by sector) without performing the swap operation. Although performance optimizations, such as conveying hashes for each sector in the manifest rather than a hash of the entire firmware image, are possible, these optimizations are not detailed in this specification.
Without hardware-based, on-the-fly decryption, the image in the primary slot is available in cleartext and may need to be re-encrypted before copying it to the secondary slot. This step might be necessary if the secondary slot has different access permissions or is located in off-chip flash memory, which tends to be more vulnerable to physical attacks.
+--------------------------------------------------+
| Bootloader |
+--------------------------------------------------+
| Primary Slot |
| (sector 1)|
|..................................................|
| |
| (sector 2)|
|..................................................|
| |
| (sector 3)|
|..................................................|
| |
| (sector 4)|
+--------------------------------------------------+
| Secondary Slot |
| (sector 1)|
|..................................................|
| |
| (sector 2)|
|..................................................|
| |
| (sector 3)|
|..................................................|
| |
| (sector 4)|
+--------------------------------------------------+
| Swap Area |
| |
+--------------------------------------------------+
| Configuration Data |
+--------------------------------------------------+
{: #image-layout title="Example Flash Area Layout"}
The ability to resume an interrupted firmware update is often essential for unattended devices, including low-end, constrained IoT devices. To meet this requirement, a firmware image must be divided into sectors, with each sector encrypted individually using a cipher that does not increase the size of the resulting ciphertext (i.e., by avoiding the addition of an authentication tag after each encrypted block).
If an update is aborted while the bootloader is decrypting the newly received image and swapping the sectors, the bootloader can restart from where it left off. This technique enhances robustness and performance.
For this purpose, ciphers without integrity protection are employed to encrypt the firmware image. It is crucial that integrity protection for the firmware image is provided, and the suit-parameter-image-digest, defined in {{Section 8.4.8.6 of I-D.ietf-suit-manifest}}, MUST be utilized.
{{RFC9459}} specifies the AES Counter (AES-CTR) mode and AES Cipher Block Chaining (AES-CBC) ciphers, both of which do not provide integrity protection. These ciphers are suitable for firmware encryption in IoT devices. However, for many other scenarios involving software packages, configuration information, or personalization data, the use of AEAD ciphers is RECOMMENDED.
The following subsections offer additional information on the selection of initialization vectors (IVs) for use with AES-CBC and AES-CTR in the context of firmware encryption. An IV MUST NOT be reused when the same key is employed. In this application, the IVs are not random but are instead based on the slot/sector combination in flash memory. The discussion assumes that the block size of AES is significantly smaller than the sector size. Typically, flash memory sectors are measured in KiB, necessitating the decryption of multiple AES blocks to complete the decryption of an entire sector.
The following manifests illustrate how to deliver an encrypted payload along with its encryption information to devices.
In the AES-KW examples, HMAC-256 MACs are included, utilizing the following secret key:
'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa'
(616161... in hex, and its length is 32)
ES-DH examples are signed using the following ECDSA secp256r1 key:
-----BEGIN PRIVATE KEY-----
MIGHAgEAMBMGByqGSM49AgEGCCqGSM49AwEHBG0wawIBAQQgApZYjZCUGLM50VBC
CjYStX+09jGmnyJPrpDLTz/hiXOhRANCAASEloEarguqq9JhVxie7NomvqqL8Rtv
P+bitWWchdvArTsfKktsCYExwKNtrNHXi9OB3N+wnAUtszmR23M4tKiW
-----END PRIVATE KEY-----
The corresponding public key can be used to verify these examples:
-----BEGIN PUBLIC KEY-----
MFkwEwYHKoZIzj0CAQYIKoZIzj0DAQcDQgAEhJaBGq4LqqvSYVcYnuzaJr6qi/Eb
bz/m4rVlnIXbwK07HypLbAmBMcCjbazR14vTgdzfsJwFLbM5kdtzOLSolg==
-----END PUBLIC KEY-----
Each example uses SHA-256 as the digest function.
The following SUIT manifest instructs a parser to authenticate the manifest using COSE_Mac0 with HMAC256. It also directs the parser to write and decrypt the encrypted payload into a component using the suit-directive-write directive.
The SUIT manifest in diagnostic notation (with line breaks added for clarity) is displayed below:
{::include-fold examples/suit-manifest-aes-kw-content.diag.signed}
In hex format, the SUIT manifest is:
{::include-fold examples/suit-manifest-aes-kw-content.hex.signed}
The following SUIT manifest instructs a parser to fetch and store the encrypted payload. Subsequently, the payload is decrypted and copied into another component using the suit-directive-copy directive. This approach is particularly effective for constrained devices with execute-in-place (XIP) flash memory.
The SUIT manifest in diagnostic notation (with line breaks added for clarity) is displayed below:
{::include-fold examples/suit-manifest-aes-kw.diag.signed}
The default storage area is defined by the component identifier (see {{Section 8.4.5.1 of I-D.ietf-suit-manifest}}). In this example, the component identifier for component #0 is ['plaintext-firmware'] and the file path "/plaintext-firmware" is the expected location.
While parsing the manifest, the behavior of SUIT manifest processor would be
- [L2-L17] authenticates the manifest part on [L18-L68]
- [L22-L25] gets two component identifiers; ['plaintext-firmware'] for component #0, and ['encrypted-firmware'] for component # 1 respectively
- [L29] sets current component index # 1 (the lasting directives target ['encrypted-firmware'])
- [L33-L34] sets source uri parameter "https://example.com/encrypted-firmware"
- [L36] fetches content from source uri into ['encrypted-firmware']
- [L39] sets current component index # 0 (the lasting directives target ['plaintext-firmware'])
- [L42-L62] sets SUIT encryption info parameter
- [L63-L64] sets source component index parameter # 1
- [L66] decrypts component # 1 (source component index) and stores the result into component # 0 (current component index)
The following attributes and features from the SUIT manifest specification are used:
| Attribute Name | Abb. | Manifest Ref. |
|---|---|---|
| component identifier | CI | Sec. 8.4.5.1 |
| (destination) component index | dst-CI | Sec. 8.4.10.1 |
| (destination) component slot OPTIONAL param | dst-CS | Sec. 8.4.8.8 |
| (source) uri OPTIONAL parameter | src-URI | Sec. 8.4.8.10 |
| source component index OPTIONAL parameter | src-CI | Sec. 8.4.8.11 |
The resulting state of SUIT manifest processor is shown in the following table:
| Abbreviation | Plaintext | Ciphertext |
|---|---|---|
| CI | ['plaintext-firmware'] | ['encrypted-firmware'] |
| dst-CI | 0 | 1 |
| dst-CS | N/A | N/A |
| src-URI | N/A | "https://example.com/encrypted-firmware" |
| src-CI | 1 | N/A |
In hex format, the SUIT manifest shown above is:
{::include-fold examples/suit-manifest-aes-kw.hex.signed}
The previous example does not utilize storage slots. However, it is possible to implement this functionality for devices that support slots in flash memory. In the enhanced example below, we reference the slots using [h'00'] and [h'01']. In this context, the component identifier [h'00'] designates component slot #0.
{::include-fold examples/suit-manifest-aes-kw-slot.diag.signed}
The following SUIT manifest instructs a parser to authenticate the manifest using COSE_Sign1 with ES256. It also directs the parser to write and decrypt the encrypted payload into a component via the suit-directive-write directive.
The SUIT manifest in diagnostic notation (formatted with line breaks for clarity) is presented below:
{::include-fold examples/suit-manifest-es-ecdh-content.diag.signed}
In hex format, the SUIT manifest is this:
{::include-fold examples/suit-manifest-es-ecdh-content.hex.signed}
The following SUIT manifest requests a parser to resolve the dependency.
The dependent manifest is signed with another key:
-----BEGIN EC PRIVATE KEY-----
MHcCAQEEIIQa67e56m8CYL5zVaJFiLl30j0qxb8ray2DeUMqH+qYoAoGCCqGSM49
AwEHoUQDQgAEDpCKqPBm2x8ITgw2UsY5Ur2Z8qW9si+eATZ6rQOrpot32hvYrE8M
tJC6IQZIv3mrFk1JrTVR1x0xSydJ7kLSmg==
-----END EC PRIVATE KEY-----
The dependency manifest is embedded as an integrated-dependency and referred to by the "#dependency-manifest" URI.
The SUIT manifest in diagnostic notation (with line breaks added for readability) is shown here:
{::include-fold examples/suit-manifest-es-ecdh-dependency.diag.signed}
In hex format, the SUIT manifest is this:
{::include-fold examples/suit-manifest-es-ecdh-dependency.hex.signed}
The algorithms outlined in this document assume that the party responsible for payload encryption:
- shares a key-encryption key (KEK) with the recipient (for use with the AES Key Wrap scheme), or
- possesses the recipient's public key (for use with ES-DH).
Both scenarios necessitate initial communication to distribute these keys among the involved parties. This interaction can be facilitated by a device management protocol, as described in {{RFC9019}}, or may occur earlier in the device lifecycle, such as during manufacturing or commissioning. In addition to the keying material, key identifiers and algorithm information must also be provisioned. This specification does not impose any requirements on the structure of the key identifier.
In certain situations, third-party companies analyze binaries for known security vulnerabilities. However, encrypted payloads hinder this type of analysis. Consequently, these third-party companies must either be granted access to the plaintext binary before encryption or be authorized recipients of the encrypted payloads. In either case, it is essential to explicitly consider these third parties in the software supply chain when binary analysis is required.
This entire document focuses on security.
It is considered best security practice to use different keys for different purposes. For instance, the key-encryption key (KEK) utilized in an AES-KW-based content key distribution method for encryption should be distinct from the long-term symmetric key employed for authentication in a communication security protocol.
To further minimize the attack surface, it may be advantageous to use different long-term keys for encrypting various types of payloads. For example, KEK_1 could be used with an AES-KW content key distribution method to encrypt a firmware image, while KEK_2 would encrypt configuration data.
A significant portion of this document addresses content key distribution, employing two methods: AES Key Wrap (AES-KW) and Ephemeral-Static Diffie-Hellman (ES-DH). The table below summarizes the main properties related to their deployment:
| Number of Long-Term Keys |
Number of Content Encryption Keys (CEKs) |
Use Case | Recommended? |
|---|---|---|---|
| Same key for all devices |
Single CEK per payload shared with all devies |
Legacy Usage |
No, bad practice |
| One key per device |
Single CEK per payload shared with all devies |
Efficient Payload Distribution |
Yes |
| One Key per device |
One CEK per payload encryption transaction per device |
Point-to- Point Payload Distribution |
Yes |
The use of firmware encryption in IoT devices introduces the risk of a battery exhaustion attack. This attack exploits the high energy cost of flash memory operations. To execute this attack, the adversary must be able to swap detached payloads and trick the device into processing an incorrect payload. Payload swapping is feasible only if there is no communication security protocol between the device and the distribution system or if the distribution system itself has been compromised.
While the security features provided by the manifest can detect this attack and prevent the device from booting with an incorrectly supplied payload, the energy-intensive flash operations will have already occurred. As a result, these operations can diminish the lifespan of the devices, making battery-powered IoT devices particularly susceptible to such attacks. For further discussion on IoT devices using flash memory, see {{flash}}.
Including the digest of the encrypted payload in the manifest enables the device to detect a battery exhaustion attack before energy-consuming decryption and flash memory copy or swap operations take place.
IANA is asked to add the following value to the SUIT Parameters registry established by {{Section 11.5 of I-D.ietf-suit-manifest}}:
Label Name Reference
-----------------------------------------
TBD19 Encryption Info Section 4
RFC Editor's Note (TBD19): The value for the Encryption Info parameter is set to 19, as the proposed value.
--- back
The following CDDL must be appended to the SUIT Manifest CDDL. The SUIT CDDL is defined in Appendix A of {{I-D.ietf-suit-manifest}}
{::include-fold draft-ietf-suit-firmware-encryption.cddl}
{: numbered="no"}
We would like to thank Henk Birkholz for his feedback on the CDDL description in this document. Additionally, we would like to thank Michael Richardson, Dick Brooks, Øyvind Rønningstad, Dave Thaler, Laurence Lundblade, Christian Amsüss, Ruud Derwig, Martin Thomson. Kris Kwiatkowski, Suresh Krishnan and Carsten Bormann for their review feedback.
We would like to thank the IESG, in particular Deb Cooley, Éric Vyncke and Roman Danyliw, for their help to improve the quality of this document.