Network Working Group C. Jennings
Internet-Draft S. Nandakumar
Intended status: Standards Track R. Barnes
Expires: 23 April 2025 Cisco
20 October 2024
End-to-End Secure Objects for Media over QUIC Transport
draft-jennings-moq-secure-objects-01
Abstract
This document describes an end-to-end authenticated encryption scheme
for application objects intended to be delivered over Media over QUIC
Transport (MOQT). We reuse the SFrame scheme for authenticated
encryption of media objects, while suppressing data that would be
redundant between SFrame and MOQT, for an efficient on-the-wire
representation.
About This Document
This note is to be removed before publishing as an RFC.
The latest revision of this draft can be found at
draft-jennings-moq-secure-objects.html">https://suhashere.github.io/moq-secure-objects/#go.draft-jennings-
draft-jennings-moq-secure-objects.html">moq-secure-objects.html. Status information for this document may be
found at https://datatracker.ietf.org/doc/draft-jennings-moq-secure-
objects/.
Discussion of this document takes place on the Media over QUIC
Working Group mailing list (mailto:moq@ietf.org), which is archived
at https://mailarchive.ietf.org/arch/browse/moq/. Subscribe at
https://www.ietf.org/mailman/listinfo/moq/.
Source for this draft and an issue tracker can be found at
https://github.com/suhasHere/moq-secure-objects.
Status of This Memo
This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet-
Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
This Internet-Draft will expire on 23 April 2025.
Copyright Notice
Copyright (c) 2024 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents (https://trustee.ietf.org/
license-info) in effect on the date of publication of this document.
Please review these documents carefully, as they describe your rights
and restrictions with respect to this document. Code Components
extracted from this document must include Revised BSD License text as
described in Section 4.e of the Trust Legal Provisions and are
provided without warranty as described in the Revised BSD License.
Table of Contents
1. Introduction
2. Terminology
3. MOQT Object Model Recap
4. Secure Objects
4.1. Setup Assumptions
4.2. Secure Object Format
4.3. Encryption Schema
4.4. Metadata Authentication
4.5. Nonce Formation
4.6. Key Derivation
4.7. Encryption
4.8. Decryption
5. Security Considerations
6. IANA Considerations
7. References
7.1. Normative References
7.2. Informative References
Appendix A. Acknowledgements
Authors' Addresses
1. Introduction
Media Over QUIC Transport (MOQT) is a protocol that is optimized for
the QUIC protocol, either directly or via WebTransport, for the
dissemination of delivery of low latency media
[I-D.ietf-moq-transport]. MOQT defines a publish/subscribe media
delivery layer across set of participating relays for supporting wide
range of use-cases with different resiliency and latency (live,
interactive) needs without compromising the scalability and cost
effectiveness associated with content delivery networks. It supports
sending media objects through sets of relays nodes.
Typically a MOQ Relay doesn't need to access the media content, thus
allowing the media to be "end-to-end" encrypted so that it cannot be
decrypted by the relays. However for a relay to participate
effectively in the media delivery, it needs to access naming
information of a MOQT object to carryout the required store and
forward functions.
As such, two layers of security are required:
1. Hop-by-hop (HBH) security between two MOQT relays
2. End-to-end (E2E) security from the Publisher of an MOQT object to
End Subscribers
The HBH security is provided by TLS in the QUIC connection that MOQT
runs over. MOQT support different E2EE protection as well as
allowing for E2EE security.
This document defines a scheme for E2E authenticated encryption of
MOQT objects. This scheme is based on the SFrame mechanism for
authenticated encryption of media objects [I-D.ietf-sframe-enc].
However, a secondary goal of this design is to minimize the amount of
additional data the encryptions requires for each object. This is
particularly important for very low bit rate audio applications where
the encryption overhead can increase overall bandwidth usage by a
significant percentage. To minimize the overhead added by end-to-end
encryption, certain fields that would be redundant between MOQT and
SFrame are not transmitted.
2. Terminology
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
"OPTIONAL" in this document are to be interpreted as described in BCP
14 [RFC2119] [RFC8174] when, and only when, they appear in all
capitals, as shown here.
This document re
E2EE: End to End Encryption
HBH: Hop By Hop
Producer: Software that creates and encrypts MoQ Objects.
Consumer: Software that decrypts MoQ Objects.
3. MOQT Object Model Recap
MOQT defines a publish/subscribe based media delivery protocol, where
in endpoints, called producers, publish objects which are delivered
via participating relays to receiving endpoints, called consumers.
Section 2 of [I-D.ietf-moq-transport] defines hierarchical object
model for application data, comprised of objects, groups and tracks.
Objects defines the basic data element, an addressable unit whose
payload is sequence of bytes. All objects belong to a group,
indicating ordering and potential dependencies. A track contains a
sequence of groups and serves as the entity against which a consumer
issues a subscription request.
Media Over QUIC Application
|
| time
+-- TrackA --+---------+-----+---------+-------+---------+------>
| | Group1 | | Group2 | ... | GroupN |
| +----+----+ +----+----+ +---------+
| | |
| | |
| +----+----+ +----+----+
| | Object0 | | Object0 |
| +---------+ +---------+
| | Object1 | | Object1 |
| +---------+ +---------+
| | Object2 | | Object2 |
| +---------+ +---------+
| ...
| +---------+
| | ObjectN |
| +---------+
|
| time
+-- TrackB --+---------+-----+---------+-------+---------+------>
| Group1 | | Group2 | ... | GroupN |
+----+----+ +----+----+ +----+----+
| | |
| | |
+----+----+ +----+----+ +----+----+
| Object0 | | Object0 | | Object0 |
+---------+ +---------+ +---------+
Figure 1: Structure of an MOQT session
Objects are comprised of two parts: envelope and a payload. The
envelope is never end to end encrypted and is always visible to
relays. The payload portion may be end to end encrypted, in which
case it is only visible to the producer and consumer. The
application is solely responsible for the content of the object
payload.
Tracks are identified by a combination of its TrackNamespace and
TrackName. TrackNamespace and TrackName are treated as a sequence of
binary bytes. Group and Objects are represented as variable length
integers called GroupId and ObjectId respectively.
For purposes of this specification, we define FullTrackName as :
FullTrackName = TrackNamespace | TrackName
where | representations concatenation of byte strings,
and ObjectName is combination of following properties:
ObjectName = (FullTrackName, GroupId, ObjectId)
Two important properties of objects are:
1. ObjectNames are globally unique in a given relay network.
2. The data inside an object (and its size) can never change after
the object is first published. There can never be two objects
with the same name but different data.
One of the ways system keep the object names unique is by using a
fully qualified domain names or UUIDs as part of the TrackNamespace.
4. Secure Objects
Section 7.1.1 of [I-D.ietf-moq-transport] defines fields of a
canonical MOQT Object. The protection scheme defined in this draft
encrypts the Object Payload and authenticates the Track Alias, Group
ID, Object ID, and Object Payload fields, regardless of the on-the-
wire encoding of the objects over QUIC Datagrams or QUIC streams.
4.1. Setup Assumptions
We assume that the application assigns each track a set of (KID,
track_base_key) tuples, where each track_base_key is known only to
authorized producer and consumers for a given track. How these per-
track secrets are established is outside the scope of this
specification. We also assume that the application defines which KID
should be used for a given encryption operation. (For decryption,
the KID is obtained from the object payload.)
It is also up to the application to specify the ciphersuite to be
used for each track's encryption context. Any SFrame ciphersuite can
be used.
4.2. Secure Object Format
The payload of a secure object comprises an AEAD-encrypted object
payload, with a header prepended that specifies the KID (encoded as
QUIC Varint) in use.
SECURE_OBJECT {
Key ID (i),
Encrypted Data (..),
}
4.3. Encryption Schema
MOQT secure object protection relies on an SFrame cipher suite to
define the AEAD encryption algorithm and hash algorithm in use
[RFC9605]. We will refer to the following aspects of the AEAD and
the hash algorithm below:
* AEAD.Encrypt and AEAD.Decrypt - The encryption and decryption
functions for the AEAD. We follow the convention of RFC 5116
[RFC5116] and consider the authentication tag part of the
ciphertext produced by AEAD.Encrypt (as opposed to a separate
field as in SRTP [RFC3711]).
* AEAD.Nk - The size in bytes of a key for the encryption algorithm
* AEAD.Nn - The size in bytes of a nonce for the encryption
algorithm
* AEAD.Nt - The overhead in bytes of the encryption algorithm
(typically the size of a "tag" that is added to the plaintext)
* AEAD.Nka - For cipher suites using the compound AEAD described in
Section 4.5.1 of [RFC9605], the size in bytes of a key for the
underlying encryption algorithm
* Hash.Nh - The size in bytes of the output of the hash function
4.4. Metadata Authentication
The KID, FullTrackName, Group ID, and Object ID for a given object
are authenticated as part of secure object encryption. This ensures,
for example, that encrypted objects cannot be replayed across tracks.
When protecting or unprotecting a secure object, an endpoint encodes
the key ID, Group ID, Object ID, and FullTrackName in the following
data structure, for input to the AEAD function's AAD argument:
SECURE_OBJECT_AAD {
Key ID (i),
Group ID (i),
Object ID (i),
Track Namespace (tuple),
Track Name (b),
}
4.5. Nonce Formation
The Group ID and Object ID for an object are used to form a 96-bit
counter (CTR) value, which XORed with a salt to form the nonce used
in AEAD encryption. The counter value is formed by encoding the
Group ID and Object ID as QUIC varints, then concatenating these
representations. This scheme MUST NOT be applied to an object where
group ID is larger than 2^62 or the object ID is larger than 2^30.
def encode_varint(x):
if x < 0x40:
return (x, 8)
elif x < 0x4000:
return (0x4000 + x, 16)
elif x < 0x40000000:
return (0x80000000 + x, 32)
elif x < 0x4000000000000000:
return (0xc000000000000000 + x, 64)
def encode_ctr(group_id, object_id):
(group_id, group_bits) = encode_varint(group_id)
(object_id, object_bits) = encode_varint(object_id)
group_shift = 96 - group_bits
object_shift = group_shift - object_bits
return (group_id << group_shift) | (object_id << object_shift)
4.6. Key Derivation
Encryption and decryption use a key and salt derived from the
track_base_key associated with a KID. Given a track_base_key value,
the key and salt are derived using HMAC-based Key Derivation Function
(HKDF) [RFC5869] as follows:
def derive_key_salt(KID, track_base_key):
moq_secret = HKDF-Extract("", track_base_key)
moq_key_label = "MOQ 1.0 Secret key " + KID + cipher_suite
moq_key =
HKDF-Expand(moq_secret, moq_key_label, AEAD.Nk)
moq_salt_label = "MOQ 1.0 Secret salt " + KID + cipher_suite
moq_salt =
HKDF-Expand(moq_secret, moq_salt_label, AEAD.Nn)
return moq_key, moq_salt
In the derivation of moq_secret:
* The + operator represents concatenation of byte strings.
* The KID value is encoded as an 8-byte big-endian integer.
* The cipher_suite value is a 2-byte big-endian integer representing
the cipher suite in use (see [I-D.ietf-sframe-enc]).
The hash function used for HKDF is determined by the cipher suite in
use.
4.7. Encryption
MOQT secure object encryption uses the AEAD encryption algorithm for
the cipher suite in use. The key for the encryption is the moq_key
derived from the track_base_key Section 4.6. The nonce is formed by
first XORing the moq_salt with the current CTR value, and then
encoding the result as a big-endian integer of length AEAD.Nn.
The payload field from the MOQT object is used by the AEAD algorithm
for the plaintext.
The encryptor forms an SecObj header using the KID value provided.
The encryption procedure is as follows:
1. From the MOQT Object obtain MOQT object payload as the plaintext
to encrypt. Get the GroupId and ObjectId from the MOQT object
envelope.
2. Retrieve the moq_key and moq_salt matching the KID.
3. Form the aad input as described in Section 4.4.
4. Form the nonce by as described in Section 4.5.
5. Apply the AEAD encryption function with moq_key, nonce, aad and
object payload as inputs.
6. Add the KID value to KID Object Header Extension.
The final SecureObject is formed from the MOQT transport headers,
then the KID encdoded as QUIC variale length integer[RFC9000],
followed by the output of the encryption.
+-----------------+------------------+-----------------+
| MOQT Object | SecObj Header | SecObj |
| Header | (KID Extension) | Ciphertext |
+-----------------+------------------+-----------------+
Below shows pseudocode for the encryption process.
def encrypt(full_track_name, kid, object):
# Identify the appropriate encryption context
ctx = context_for_track(full_track_name)
moq_key, moq_salt = ctx.key_store[kid]
# Compute the required CTR parameter
ctr = encode_ctr(object.group_id, object.object_id)
# Assemble the AAD value
aad = encode_aad(kid, ctr, full_track_name)
# Perform the AEAD encryption
nonce = xor(moq_salt, ctr)
encrypted_payload = AEAD.encrypt(moq_key, nonce, aad, object.payload)
# Assemble the secure object payload
(encoded_kid, _) = encode_varint(kid)
object.payload = encoded_kid + encrypted_payload
4.8. Decryption
For decrypting, the KID field in the secure object payload is used to
find the right key and salt for the encrypted object, the nonce field
is obtained from the GroupId and ObjectId fields of the MOQT object
header.
The decryption procedure is as follows:
1. Parse the SecureObject to obtain KID, the ciphertext
corresponding to MOQT object payload and the GroupID and ObjectId
from the MOQT object envelope.
2. Retrieve the moq_key and moq_salt matching the KID.
3. Form the aad input as described in Section 4.4.
4. Form the nonce by as described in Section 4.5.
5. Apply the AEAD decryption function with moq_key, nonce, aad and
ciphertext as inputs.
Below shows psuedocode for the decrpytion process
def decrypt(full_track_name, object):
# Parse the secure object payload to obtain key ID and ciphertext
(kid, kid_byte_len) = parse_varint(object.payload)
ciphertext = object.payload[kid_byte_len:]
# Identify the appropriate encryption context for the full track name
# and the key ID
ctx = context_for_track(full_track_name)
moq_key, moq_salt = ctx.key_store[kid]
# Compute the required CTR parameter
ctr = encode_ctr(object.group_id, object.object_id)
# Assemble the AAD value
aad = encode_aad(kid, ctr, full_track_name)
# Perform the AEAD decryption
nonce = xor(moq_salt, ctr)
object.payload = AEAD.decrypt(moq_key, nonce, aad, ciphertext)
If a ciphertext fails to decrypt because there is no key available
for the KID value presented, the client MAY buffer the ciphertext and
retry decryption once a key with that KID is received. If a
ciphertext fails to decrypt for any other reason, the client MUST
discard the ciphertext. Invalid ciphertexts SHOULD be discarded in a
way that is indistinguishable (to an external observer) from having
processed a valid ciphertext. In other words, the decryption
operation should take the same amount of time regardless of whether
decryption succeeds or fails.
5. Security Considerations
The cryptographic computations described in this document are exactly
those performed in the SFrame encryption scheme defined in
[I-D.ietf-sframe-enc], The scheme in this document is effectively a
"virtualized" version of SFrame:
* The CTR value used in nonce formation is not carried in the object
payload, but instead synthesized from the group ID and object ID.
* The AAD for the AEAD operation is not sent on the wire (as with
the SFrame Header), but constructed locally by the encrypting and
decrypting endpoints.
* The format of the AAD is different:
- The SFrame Header is constructed using QUIC-style varints,
instead of the variable-length integer scheme defined in
SFrame.
- The group ID and object ID are sent directly, not as the packed
CTR value.
* The metadata input in to SFrame operations is defined to be the
FullTrackName value for the object.
* The labels used in key derivation reflect MOQ usage, not generic
SFrame.
The security considerations discussed in the SFrame specification
thus also apply here.
The SFrame specification lists several things that an application
needs to account for in order to use SFrame securely, which are all
accounted for here:
1. *Header value uniqueness:* Uniqueness of CTR values follows from
the uniqueness of MOQT (group ID, object ID) pairs. We only use
one KID value, but instead use distinct SFrame contexts with
distinct keys per track. This assures that the same
(track_base_key, KID, CTR) tuple is never used twice.
2. *Key management:* We delegate this to the MOQT application, with
subject to the assumptions described in Section 4.1.
3. *Anti-replay:* Replay is not possible within the MOQT framework
because of the uniqueness constraints on object IDs and objects,
and because the group ID and object ID are cryptographically
bound to the secure object payload.
4. *Metadata:* The analogue of the SFrame metadata input is defined
in Section 4.4.
*NOTE:* It is not clear to me that the anti-replay point actually
holds up here, but that is probably just due to the limitations of
my understanding of MOQT. How is a receiver or relay supposed to
be have if its next upstream hop sends it multiple values with the
same track name, group ID, and object ID?
Any of the SFrame ciphersuites defined in the relevant IANA registry
can be used to protect MOQT objects. The caution against short tags
in Section 7.5 of [I-D.ietf-sframe-enc] still applies here, but the
MOQT environment provides some safeguards that make it safer to use
short tags, namely:
* MOQT has hop-by-hop protections provided by the underlying QUIC
layer, so a brute-force attack could only be mounted by a relay.
* MOQT tracks have predictable object arrival rates, so a receiver
can interpret a large deviation from this rate as a sign of an
attack.
* The the binding of the secure object payload to other MOQT
parameters (as metadata), together with MOQT's uniqueness
properties ensure that a valid secure object payload cannot be
replayed in a different context.
6. IANA Considerations
This document defines a new MOQT Object extension header for carrying
KID value, under the MOQ Extension Headers registry.
+======+=============================+
| Type | Value |
+======+=============================+
| 0x1 | KID Value - see Section 4.2 |
+------+-----------------------------+
Table 1
7. References
7.1. Normative References
[I-D.ietf-moq-transport]
Curley, L., Pugin, K., Nandakumar, S., Vasiliev, V., and
I. Swett, "Media over QUIC Transport", Work in Progress,
Internet-Draft, draft-ietf-moq-transport-06, 19 September
2024, <https://datatracker.ietf.org/doc/html/draft-ietf-
moq-transport-06>.
[I-D.ietf-sframe-enc]
Omara, E., Uberti, J., Murillo, S. G., Barnes, R., and Y.
Fablet, "Secure Frame (SFrame)", Work in Progress,
Internet-Draft, draft-ietf-sframe-enc-09, 4 April 2024,
<https://datatracker.ietf.org/doc/html/draft-ietf-sframe-
enc-09>.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119,
DOI 10.17487/RFC2119, March 1997,
<https://www.rfc-editor.org/rfc/rfc2119>.
[RFC5116] McGrew, D., "An Interface and Algorithms for Authenticated
Encryption", RFC 5116, DOI 10.17487/RFC5116, January 2008,
<https://www.rfc-editor.org/rfc/rfc5116>.
[RFC5869] Krawczyk, H. and P. Eronen, "HMAC-based Extract-and-Expand
Key Derivation Function (HKDF)", RFC 5869,
DOI 10.17487/RFC5869, May 2010,
<https://www.rfc-editor.org/rfc/rfc5869>.
[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/rfc/rfc8174>.
[RFC9000] Iyengar, J., Ed. and M. Thomson, Ed., "QUIC: A UDP-Based
Multiplexed and Secure Transport", RFC 9000,
DOI 10.17487/RFC9000, May 2021,
<https://www.rfc-editor.org/rfc/rfc9000>.
[RFC9605] Omara, E., Uberti, J., Murillo, S. G., Barnes, R., Ed.,
and Y. Fablet, "Secure Frame (SFrame): Lightweight
Authenticated Encryption for Real-Time Media", RFC 9605,
DOI 10.17487/RFC9605, August 2024,
<https://www.rfc-editor.org/rfc/rfc9605>.
7.2. Informative References
[RFC3711] Baugher, M., McGrew, D., Naslund, M., Carrara, E., and K.
Norrman, "The Secure Real-time Transport Protocol (SRTP)",
RFC 3711, DOI 10.17487/RFC3711, March 2004,
<https://www.rfc-editor.org/rfc/rfc3711>.
Appendix A. Acknowledgements
TODO
Authors' Addresses
Cullen Jennings
Cisco
Email: fluffy@cisco.com
Suhas Nandakumar
Cisco
Email: snandaku@cisco.com
Richard Barnes
Cisco
Email: rlb@ipv.sx