| Network Working Group | C. Jennings |
| Internet-Draft | P. Jones |
| Intended status: Standards Track | R. Barnes |
| Expires: April 20, 2019 | Cisco Systems |
| A. Roach | |
| Mozilla | |
| October 17, 2018 |
SRTP Double Encryption Procedures
draft-ietf-perc-double-10
In some conferencing scenarios, it is desirable for an intermediary to be able to manipulate some parameters in Real Time Protocol (RTP) packets, while still providing strong end-to-end security guarantees. This document defines a cryptographic transform for the Secure Real Time Protocol (SRTP) that uses two separate but related cryptographic operations to provide hop-by-hop and end-to-end security guarantees. Both the end-to-end and hop-by-hop cryptographic algorithms can utilize an authenticated encryption with associated data scheme or take advantage of future SRTP transforms with different properties.
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 April 20, 2019.
Copyright (c) 2018 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
Cloud conferencing systems that are based on switched conferencing have a central Media Distributor device that receives media from endpoints and distributes it to other endpoints, but does not need to interpret or change the media content. For these systems, it is desirable to have one cryptographic key from the sending endpoint to the receiving endpoint that can encrypt and authenticate the media end-to-end while still allowing certain information in the header of a Real Time Protocol (RTP) packet to be changed by the Media Distributor. At the same time, a separate cryptographic key provides integrity and optional confidentiality for the media flowing between the Media Distributor and the endpoints. The framework document [I-D.ietf-perc-private-media-framework] describes this concept in more detail.
This specification defines a transform for the Secure Real Time Protocol (SRTP) that uses the AES-GCM algorithm [RFC7714] to provide encryption and integrity for an RTP packet for the end-to-end cryptographic key as well as a hop-by-hop cryptographic encryption and integrity between the endpoint and the Media Distributor. The Media Distributor decrypts and checks integrity of the hop-by-hop security. The Media Distributor MAY change some of the RTP header information that would impact the end-to-end integrity. In that case, the original value of any RTP header field that is changed is included in a new RTP header extension called the Original Header Block. The new RTP packet is encrypted with the hop-by-hop cryptographic algorithm before it is sent. The receiving endpoint decrypts and checks integrity using the hop-by-hop cryptographic algorithm and then replaces any parameters the Media Distributor changed using the information in the Original Header Block before decrypting and checking the end-to-end integrity.
One can think of the double as a normal SRTP transform for encrypting the RTP in a way where things that only know half of the key, can decrypt and modify part of the RTP packet but not other parts, including the media payload.
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.
Terms used throughout this document include:
This specification uses a cryptographic context with two parts:
The RECOMMENDED cipher for the hop-by-hop and end-to-end algorithm is AES-GCM. Other combinations of SRTP ciphers that support the procedures in this document can be added to the IANA registry.
The keys and salt for these algorithms are generated with the following steps:
If the Media Distributor is to be able to modify header fields but not decrypt the payload, then it must have cryptographic key for the outer algorithm, but not the inner (end-to-end) algorithm. This document does not define how the Media Distributor should be provisioned with this information. One possible way to provide keying material for the outer (hop-by-hop) algorithm is to use [I-D.ietf-perc-dtls-tunnel].
In order to allow the inner and outer keys to be managed independently via the master key, the transforms defined in this document MUST be used with the following pseudo-random function (PRF), which preserves the separation between the two halves of the key. Given a positive integer n representing the desired output length, a master key k_master, and an input x:
PRF_double_n(k_master,x) = PRF_inner_(n/2)(k_master,x) ||
PRF_outer_(n/2)(k_master,x)
PRF_inner_n(k_master,x) = PRF_n(inner(k_master),x)
PRF_outer_n(k_master,x) = PRF_n(outer(k_master),x)
Here PRF_n(k, x) represents the AES_CM PRF KDF [RFC3711] for DOUBLE_AEAD_AES_128_GCM_AEAD_AES_128_GCM algorithm and AES_256_CM_PRF KDF [RFC6188] for DOUBLE_AEAD_AES_256_GCM_AEAD_AES_256_GCM algorithm. inner(key) represents the first half of the key, and outer(key) represents the second half of the key.
The Original Header Block (OHB) contains the original values of any modified RTP header fields. In the encryption process, the OHB is appended to the RTP payload. In the decryption process, the receiving endpoint uses it to reconstruct the original RTP header, so that it can pass the proper AAD value to the inner transform.
The OHB can reflect modifications to the following fields in an RTP header: the payload type, the sequence number, and the marker bit. All other fields in the RTP header MUST remain unmodified; since the OHB cannot reflect their original values, the receiver will be unable to verify the E2E integrity of the packet.
The OHB has the following syntax (in ABNF [RFC5234]):
OCTET = %x00-FF PT = OCTET SEQ = 2OCTET Config = OCTET OHB = [ PT ] [ SEQ ] Config
If present, the PT and SEQ parts of the OHB contain the original payload type and sequence number fields, respectively. The final "config" octet of the OHB specifies whether these fields are present, and the original value of the marker bit (if necessary):
+-+-+-+-+-+-+-+-+ |R R R R B M P Q| +-+-+-+-+-+-+-+-+
In particular, an all-zero OHB config octet (0x00) indicates that there have been no modifications from the original header.
As implied by the use of the word "double" above, this transform applies AES-GCM to the SRTP packet twice. This allows media distributors to be able to modify some header fields while allowing endpoints to verify the end-to-end integrity and confidentiality of a packet.
The first, "inner" application of AES-GCM encrypts the SRTP payload and integrity-protects a version of the SRTP header with extensions truncated. Omitting extensions from the inner integrity check means that they can be modified by a media distributor holding only the "outer" key.
The second, "outer" application of AES-GCM encrypts the ciphertext produced by the inner encryption (i.e., the encrypted payload and authentication tag), plus an OHB that expresses any changes made between the inner and outer transforms.
A media distributor that has the outer key but not the inner key may modify the header fields that can be included in the OHB by decrypting, modifying, and re-encrypting the packet.
To encrypt a packet, the endpoint encrypts the packet using the inner (end-to-end) cryptographic key and then encrypts using the outer (hop-by-hop) cryptographic key. The encryption also supports a mode for repair packets that only does the outer (hop-by-hop) encryption. The processes is as follows:
When using EKT [I-D.ietf-perc-srtp-ekt-diet], the EKT Field comes after the SRTP packet exactly like using EKT with any other SRTP transform.
The Media Distributor has the part of the key for the outer (hop-by-hop) cryptographic algorithm, but it does not have the part of the key for the (end-to-end) cryptographic algorithm. The cryptographic algorithm and key used to decrypt a packet and any encrypted RTP header extensions would be the same as those used in the endpoint's outer algorithm and key.
In order to modify a packet, the Media Distributor decrypts the received packet, modifies the packet, updates the OHB with any modifications not already present in the OHB, and re-encrypts the packet using the the outer (hop-by-hop) cryptographic key before transmitting.
In order to avoid nonce reuse, the cryptographic contexts used in step 1 and step 5 MUST use different, independent master keys and master salts.
Note that if multiple MDs modify the same packet, then the first MD to alter a given header field is the one that adds it to the OHB. If a subsequent MD changes the value of a header field that has already been changed, then the original value will already be in the OHB, so no update to the OHB is required.
A Media Distributor that decrypts, modifies, and re-encrypts packets in this way MUST use an independent key for each recipient, SHOULD use an independent salt for each recipient, and MUST NOT re-encrypt the packet using the sender's keys. If the Media Distributor decrypts and re-encrypts with the same key and salt, it will result in the reuse of a (key, nonce) pair, undermining the security of GCM.
To decrypt a packet, the endpoint first decrypts and verifies using the outer (hop-by-hop) cryptographic key, then uses the OHB to reconstruct the original packet, which it decrypts and verifies with the inner (end-to-end) cryptographic key.
Once the packet has been successfully decrypted, the application needs to be careful about which information it uses to get the correct behavior. The application MUST use only the information found in the synthetic SRTP packet and MUST NOT use the other data that was in the outer SRTP packet with the following exceptions:
The PT and sequence number from the inner SRTP packet can be used for collection of various statistics.
If the RTP header of the outer packet contains extensions, they MAY be used. However, because extensions are not protected end-to-end, implementations SHOULD reject an RTP packet containing headers that would require end-to-end protection.
Unlike RTP, which is encrypted both hop-by-hop and end-to-end using two separate cryptographic keys, RTCP is encrypted using only the outer (hop-by-hop) cryptographic key. The procedures for RTCP encryption are specified in [RFC3711] and this document introduces no additional steps.
Media distributors sometimes interact with RTP media packets sent by endpoints, e.g., to provide recovery or receive commands via DTMF. When media packets are encrypted end-to-end, these procedures require modification.
Repair mechanisms, in general, will need to perform recovery on encrypted packets (double-encrypted when using this transform). When the recovery mechanism calls for the recovery packet itself to be encrypted, it is encrypted with only the outer, HBH key. This allows a media distributor to generate recovery packets without having access to the inner, E2E keys. However, it also results in recovery packets being triple-encrypted, twice for the base transform, and once for the recovery protection.
When using RTX [RFC4588] with double, the cached payloads MUST be the double-encrypted packets, i.e., the bits that are sent over the wire to the other side. When encrypting a retransmission packet, it MUST be encrypted the packet in repair mode (i.e., with only the HBH key).
A typical RTX receiver would decrypt the packet, undo the RTX transformation, then process the resulting packet normally by using the steps in Section 5.3.
When using RED [RFC2198] with double, the primary encoding MAY contain RTP header extensions and CSRC identifiers but non primary encodings cannot.
The sender takes encrypted payload from the cached packets to form the RED payload. Any header extensions from the primary encoding are copied to the RTP packet that will carry the RED payload and the other RTP header information such as SSRC, SEQ, CSRC, etc are set to the same as the primary payload. The RED RTP packet is then encrypted in repair mode and sent.
The receiver decrypts the payload to find the encrypted RED payload. Note a media relay can do this decryption as the packet was sent in repair mode that only needs the hop-by-hop key. The RTP headers and header extensions along with the primary payload and PT from inside the RED payload (for the primary encoding) are used to form the encrypted primary RTP packet which can then be decrypted with double.
The RTP headers (but not header extensions or CSRC) along with PT from inside the RED payload corresponding to the redundant encoding are used to from the non primary payloads. The time offset and packet rate information in the RED data MUST be used to adjust the sequence number in the RTP header. At this point the non primary packets can be decrypted with double.
Note that Flex FEC [I-D.ietf-payload-flexible-fec-scheme] is a superset of the capabilities of RED. For most applications, FlexFEC is a better choice than RED.
When using Flex FEC [I-D.ietf-payload-flexible-fec-scheme] with double, repair packets MUST be constructed by first double-encrypting the packet, then performing FEC. Processing of repair packets proceeds in the opposite order, performing FEC recovery and then decrypting. This ensures that the original media is not revealed to the Media Distributor but at the same time allows the Media Distributor to repair media. When encrypting a packet that contains the Flex FEC data, which is already encrypted, it MUST be encrypted with only the outer, HBH transform.
The algorithm recommended in [I-D.ietf-rtcweb-fec] for repair of video is Flex FEC [I-D.ietf-payload-flexible-fec-scheme]. Note that for interoperability with WebRTC, [I-D.ietf-rtcweb-fec] recommends not using additional FEC only m-line in SDP for the repair packets.
When DTMF is sent using the mechanism in [RFC4733], it is end-to-end encrypted and the relay can not read it, so it cannot be used to control the relay. Other out of band methods to control the relay need to be used instead.
This specification recommends and defines AES-GCM as both the inner and outer cryptographic algorithms, identified as DOUBLE_AEAD_AES_128_GCM_AEAD_AES_128_GCM and DOUBLE_AEAD_AES_256_GCM_AEAD_AES_256_GCM. These algorithm provide for authenticated encryption and will consume additional processing time double-encrypting for hop-by-hop and end-to-end. However, the approach is secure and simple, and is thus viewed as an acceptable trade-off in processing efficiency.
Note that names for the cryptographic transforms are of the form DOUBLE_(inner algorithm)_(outer algorithm).
While this document only defines a profile based on AES-GCM, it is possible for future documents to define further profiles with different inner and outer algorithms in this same framework. For example, if a new SRTP transform was defined that encrypts some or all of the RTP header, it would be reasonable for systems to have the option of using that for the outer algorithm. Similarly, if a new transform was defined that provided only integrity, that would also be reasonable to use for the outer transform as the payload data is already encrypted by the inner transform.
The AES-GCM cryptographic algorithm introduces an additional 16 octets to the length of the packet. When using AES-GCM for both the inner and outer cryptographic algorithms, the total additional length is 32 octets. If no other header extensions are present in the packet and the OHB is introduced, that will consume an additional 8 octets. If other extensions are already present, the OHB will consume up to 4 additional octets. Packets in repair mode will carry additional repair data, further increasing their size.
This SRTP transform provides protection against two classes of attacker: An network attacker that knows neither the inner nor outer keys, and a malicious MD that knows the outer key. Obviously, it provides no protections against an attacker that holds both the inner and outer keys.
The protections with regard to the network are the same as with the normal SRTP AES-GCM transforms.
With regard to a malicious MD, the recipient can verify the integrity of the base header fields and confidentiality and integrity of the payload. The recipient has no assurance, however, of the integrity of the header extensions in the packet.
The main innovation of this transform relative to other SRTP transforms is that it allows a partly-trusted MD to decrypt, modify, and re-encrypt a packet. When this is done, the cryptographic contexts used for decryption and re-encryption MUST use different, independent master keys and master salts. If the same context is used, the nonce formation rules for SRTP will cause the same key and nonce to be used with two different plaintexts, which substantially degrades the security of AES-GCM.
In other words, from the perspective of the MD, re-encrypting packets using this protocol will involve the same cryptographic operations as if it had established independent AES-GCM crypto contexts with the sender and the receiver. If the MD doesn't modify any header fields, then an MD that supports AES-GCM could be unused unmodified.
We request IANA to add the following values to defines a DTLS-SRTP "SRTP Protection Profile" defined in [RFC5764].
| Value | Profile | Reference |
|---|---|---|
| {0x00, 0x09} | DOUBLE_AEAD_AES_128_GCM_AEAD_AES_128_GCM | RFCXXXX |
| {0x00, 0x0A} | DOUBLE_AEAD_AES_256_GCM_AEAD_AES_256_GCM | RFCXXXX |
Note to IANA: Please assign value RFCXXXX and update table to point at this RFC for these values.
The SRTP transform parameters for each of these protection are:
DOUBLE_AEAD_AES_128_GCM_AEAD_AES_128_GCM
cipher: AES_128_GCM then AES_128_GCM
cipher_key_length: 256 bits
cipher_salt_length: 192 bits
aead_auth_tag_length: 256 bits
auth_function: NULL
auth_key_length: N/A
auth_tag_length: N/A
maximum lifetime: at most 2^31 SRTCP packets and
at most 2^48 SRTP packets
DOUBLE_AEAD_AES_256_GCM_AEAD_AES_256_GCM
cipher: AES_256_GCM then AES_256_GCM
cipher_key_length: 512 bits
cipher_salt_length: 192 bits
aead_auth_tag_length: 256 bits
auth_function: NULL
auth_key_length: N/A
auth_tag_length: N/A
maximum lifetime: at most 2^31 SRTCP packets and
at most 2^48 SRTP packets
The first half of the key and salt is used for the inner (end-to-end) algorithm and the second half is used for the outer (hop-by-hop) algorithm.
Thank you for reviews and improvements to this specification from Alex Gouaillard, David Benham, Magnus Westerlund, Nils Ohlmeier, Paul Jones, Roni Even, and Suhas Nandakumar. In addition, thank you to Sergio Garcia Murillo proposed the change of transporting the OHB information in the RTP payload instead of the RTP header.
The following figure shows a double encrypted SRTP packet. The sides indicate the parts of the packet that are encrypted and authenticated by the hop-by-hop and end-to-end operations.
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<++
|V=2|P|X| CC |M| PT | sequence number | IO
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ IO
| timestamp | IO
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ IO
| synchronization source (SSRC) identifier | IO
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ IO
| contributing source (CSRC) identifiers | IO
| .... | IO
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<+O
| RTP extension (OPTIONAL) ... | |O
+>+>+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<+O
O I | payload ... | IO
O I | +-------------------------------+ IO
O I | | RTP padding | RTP pad count | IO
O +>+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<+O
O | | E2E authentication tag | |O
O | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |O
O | | OHB ... | |O
+>| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |+
| | | HBH authentication tag | ||
| | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ ||
| | ||
| +- E2E Encrypted Portion E2E Authenticated Portion ---+|
| |
+--- HBH Encrypted Portion HBH Authenticated Portion ----+