Trusted Execution Environment Provisioning (TEEP) ProtocolArm Ltd.AbsamTirol6067Austriahannes.tschofenig@arm.comBroadcom350 Ellis StMountain ViewCA94043USAmingliang.pei@broadcom.comAmazonUSdavewhee@amazon.comMicrosoftUSdthaler@microsoft.comAISTJPakira.tsukamoto@aist.go.jp
Security
TEEPTrusted Execution EnvironmentThis document specifies a protocol that installs, updates, and deletes
Trusted Components in a device with a Trusted Execution
Environment (TEE). This specification defines an interoperable
protocol for managing the lifecycle of Trusted Components.The Trusted Execution Environment (TEE) concept has been designed to
separate a regular operating system, also referred as a Rich Execution
Environment (REE), from security-sensitive applications. In a TEE
ecosystem, device vendors may use different operating systems in the
REE and may use different types of TEEs. When Trusted Component Developers or
Device Administrators use Trusted Application Managers (TAMs) to
install, update, and delete Trusted Applications and their dependencies on a wide range
of devices with potentially different TEEs then an interoperability
need arises.This document specifies the protocol for communicating between a TAM
and a TEEP Agent.The Trusted Execution Environment Provisioning (TEEP) architecture
document provides design
guidance and introduces the
necessary 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 when, and only when, they
appear in all capitals, as shown here.This specification re-uses the terminology defined in .As explained in Section 4.4 of that document, the TEEP protocol treats
each Trusted Application (TA), any dependencies the TA has, and personalization data as separate
components that are expressed in SUIT manifests, and a SUIT manifest
might contain or reference multiple binaries (see
for more details).As such, the term Trusted Component (TC) in this document refers to a
set of binaries expressed in a SUIT manifest, to be installed in
a TEE. Note that a Trusted Component may include one or more TAs
and/or configuration data and keys needed by a TA to operate correctly.Each Trusted Component is uniquely identified by a SUIT Component Identifier
(see Section 8.7.2.2).Attestation related terms, such as Evidence and Attestation Results,
are as defined in .The TEEP protocol consists of messages exchanged between a TAM
and a TEEP Agent.
The messages are encoded in CBOR and designed to provide end-to-end security.
TEEP protocol messages are signed by the endpoints, i.e., the TAM and the
TEEP Agent, but Trusted
Applications may also be encrypted and signed by a Trusted Component Developer or
Device Administrator.
The TEEP protocol not only uses
CBOR but also the respective security wrapper, namely COSE . Furthermore, for software updates the SUIT
manifest format is used, and
for attestation the Entity Attestation Token (EAT)
format is supported although other attestation formats are also permitted.This specification defines five messages: QueryRequest, QueryResponse,
Update, Success, and Error.A TAM queries a device’s current state with a QueryRequest message.
A TEEP Agent will, after authenticating and authorizing the request, report
attestation information, list all Trusted Components, and provide information about supported
algorithms and extensions in a QueryResponse message. An error message is
returned if the request
could not be processed. A TAM will process the QueryResponse message and
determine
whether to initiate subsequent message exchanges to install, update, or delete Trusted
Applications.With the Update message a TAM can instruct a TEEP Agent to install and/or
delete one or more Trusted Components.
The TEEP Agent will process the message, determine whether the TAM is authorized
and whether the
Trusted Component has been signed by an authorized Trusted Component Signer.
A Success message is returned when the operation has been completed successfully,
or an Error message
otherwise.TEEP messages are protected by the COSE_Sign1 structure.
The TEEP protocol messages are described in CDDL format below.To create a TEEP message, the following steps are performed.Create a TEEP message according to the description below and populate
it with the respective content. TEEP messages sent by TAMs (QueryRequest
and Update) can include a “token”.
The TAM can decide, in any implementation-specific way, whether to include a token
in a message. The first usage of a token
generated by a TAM MUST be randomly created.
Subsequent token values MUST be different for each subsequent message
created by a TAM.Create a COSE Header containing the desired set of Header
Parameters. The COSE Header MUST be valid per the specification.Create a COSE_Sign1 object
using the TEEP message as the COSE_Sign1 Payload; all
steps specified in for creating a
COSE_Sign1 object MUST be followed.When TEEP message is received (see the ProcessTeepMessage conceptual API
defined in section 6.2.1),
the following validation steps are performed. If any of
the listed steps fail, then the TEEP message MUST be rejected.Verify that the received message is a valid CBOR object.Verify that the message contains a COSE_Sign1 structure.Verify that the resulting COSE Header includes only parameters
and values whose syntax and semantics are both understood and
supported or that are specified as being ignored when not
understood.Follow the steps specified in Section 4 of (“Signing Objects”) for
validating a COSE_Sign1 object. The COSE_Sign1 payload is the content
of the TEEP message.Verify that the TEEP message is a valid CBOR map and verify the fields of
the
TEEP message according to this specification.A QueryRequest message is used by the TAM to learn
information from the TEEP Agent, such as
the features supported by the TEEP Agent, including
ciphersuites and protocol versions. Additionally,
the TAM can selectively request data items from the
TEEP Agent via the request parameter. Currently,
the following features are supported:Request for attestation information,Listing supported extensions,Querying installed Trusted Components, andListing supported SUIT commands.Like other TEEP messages, the QueryRequest message is
signed, and the relevant CDDL snippet is shown below.
The complete CDDL structure is shown in Appendix C.The message has the following fields:
The value of (1) corresponds to a QueryRequest message sent from the TAM to
the TEEP Agent.
The value in the token parameter is used to match responses to requests.
This is particularly useful when a TAM issues multiple concurrent requests
to a TEEP Agent. The token MUST be present if and only if the attestation bit is clear in
the data-item-requested value. The size of the token is at least 8 bytes
(64 bits) and maximum of 64 bytes, which is the same as in an EAT Nonce
Claim (see Section 3.3). The first usage of a token
generated by a TAM MUST be randomly created.
Subsequent token values MUST be different for each request message
to distinguish the correct response from multiple requests.
The token value MUST NOT be used for other purposes, such as a TAM to
identify the devices and/or a device to identify TAMs or Trusted Components.
The TAM SHOULD set an expiration time for each token and MUST ignore any messages with expired tokens.
The TAM MUST expire the token value after receiving the first response
containing the token value and ignore any subsequent messages that have the same token
value.
The data-item-requested parameter indicates what information the TAM requests from the TEEP
Agent in the form of a bitmap.
With this value the TAM requests the TEEP Agent to return an attestation payload,
whether Evidence (e.g., an EAT) or Attestation Results, in the response.
With this value the TAM queries the TEEP Agent for all installed Trusted Components.
With this value the TAM queries the TEEP Agent for supported capabilities
and extensions, which allows a TAM to discover the capabilities of a TEEP
Agent implementation.
With this value the TAM requests the TEEP Agent to return SUIT Reports
in the response.
Further values may be added in the future.
The supported-cipher-suites parameter lists the ciphersuites supported by the TAM. If this parameter is not present, it is to be treated the same as if
it contained all ciphersuites defined in this document that are listed as “MUST”. Details
about the ciphersuite encoding can be found in .
The supported-freshness-mechanisms parameter lists the freshness mechanism(s) supported by the TAM.
Details about the encoding can be found in .
If this parameter is absent, it means only the nonce mechanism is supported.
The challenge field is an optional parameter used for ensuring the freshness of the
attestation payload returned with a QueryResponse message. It MUST be absent if
the attestation bit is clear (since the token is used instead in that case).
When a challenge is
provided in the QueryRequest and an EAT is returned with a QueryResponse message
then the challenge contained in this request MUST be used to generate the EAT,
such as by copying the challenge into the nonce claim found in the EAT if
using the Nonce freshness mechanism. For more details see .
If any format other than EAT is used, it is up to that
format to define the use of the challenge field.
The versions parameter enumerates the TEEP protocol version(s) supported by the TAM.
A value of 0 refers to the current version of the TEEP protocol.
If this field is not present, it is to be treated the same as if
it contained only version 0.The QueryResponse message is the successful response by the TEEP Agent after
receiving a QueryRequest message. As discussed in , it can also be sent
unsolicited if the contents of the QueryRequest are already known and do not vary
per message.Like other TEEP messages, the QueryResponse message is
signed, and the relevant CDDL snippet is shown below.
The complete CDDL structure is shown in Appendix C.The QueryResponse message has the following fields:
The value of (2) corresponds to a QueryResponse message sent from the TEEP Agent
to the TAM.
The value in the token parameter is used to match responses to requests. The
value MUST correspond to the value received with the QueryRequest message
if one was present, and MUST be absent if no token was present in the
QueryRequest.
The selected-cipher-suite parameter indicates the selected ciphersuite. If this
parameter is not present, it is to be treated as if the TEEP Agent accepts
any ciphersuites listed in the QueryRequest, so the TAM can select one.
Details about the ciphersuite encoding can be found in .
The selected-version parameter indicates the TEEP protocol version selected by the
TEEP Agent. The absence of this parameter indicates the same as if it
was present with a value of 0.
The attestation-payload-format parameter indicates the IANA Media Type of the
attestation-payload parameter, where media type parameters are permitted after
the media type. The absence of this parameter indicates that
the format is “application/eat-cwt; profile=https://datatracker.ietf.org/doc/html/draft-ietf-teep-protocol-09” (see
for further discussion).
(RFC-editor: upon RFC publication, replace URI above with
“https://www.rfc-editor.org/info/rfcXXXX” where XXXX is the RFC number
of this document.)
It MUST be present if the evidence parameter
is present and the format is not an EAT in CWT format with the profile
defined below in .
The attestation-payload parameter contains Evidence or Attestation Results. This parameter
MUST be present if the QueryResponse is sent in response to a QueryRequest
with the attestation bit set. If the attestation-payload-format parameter is absent,
the attestation payload contained in this parameter MUST be
an Entity Attestation Token following the encoding
defined in . See for further discussion.
If present, the suit-reports parameter contains a set of “boot” (including
starting an executable in an OS context) time SUIT Reports
as defined in Section 4 of .
If a token parameter was present in the QueryRequest
message the QueryResponse message is in response to,
the suit-report-nonce field MUST be present in the SUIT Report with a
value matching the token parameter in the QueryRequest
message. SUIT Reports can be useful in QueryResponse messages to
pass information to the TAM without depending on a Verifier including
the relevant information in Attestation Results.
The tc-list parameter enumerates the Trusted Components installed on the device
in the form of tc-info objects. This parameter MUST be present if the
QueryResponse is sent in response to a QueryRequest with the
trusted-components bit set.
The requested-tc-list parameter enumerates the Trusted Components that are
not currently installed in the TEE, but which are requested to be installed,
for example by an installer of an Untrusted Application that has a TA
as a dependency, or by a Trusted Application that has another Trusted
Component as a dependency. Requested Trusted Components are expressed in
the form of requested-tc-info objects.
A TEEP Agent can get this information from the RequestTA conceptual API
defined in section 6.2.1.
The unneeded-tc-list parameter enumerates the Trusted Components that are
currently installed in the TEE, but which are no longer needed by any
other application. The TAM can use this information in determining
whether a Trusted Component can be deleted. Each unneeded Trusted Component is identified
by its SUIT Component Identifier.
A TEEP Agent can get this information from the UnrequestTA conceptual API
defined in section 6.2.1.
The ext-list parameter lists the supported extensions. This document does not
define any extensions. This parameter MUST be present if the
QueryResponse is sent in response to a QueryRequest with the
extensions bit set.The tc-info object has the following fields:
A SUIT Component Identifier.
The suit-manifest-sequence-number value from the SUIT manifest for the Trusted Component,
if a SUIT manifest was used.The requested-tc-info message has the following fields:
A SUIT Component Identifier.
The minimum suit-manifest-sequence-number value from a SUIT manifest for
the Trusted Component. If not present, indicates that any sequence number will do.
If present with a value of true, indicates that the TEEP agent already has
the Trusted Component binary and only needs an Update message with a SUIT manifest
that authorizes installing it. If have-binary is true, the
tc-manifest-sequence-number field MUST be present.Section 7 of lists information that may appear
in Evidence depending on the circumstance. However, the Evidence is
opaque to the TEEP protocol and there are no formal requirements on the contents
of Evidence.TAMs however consume Attestation Results and do need enough information therein to
make decisions on how to remediate a TEE that is out of compliance, or update a TEE
that is requesting an authorized change. To do so, the information in
Section 7 of is often required depending on the policy.
When an Entity
Attestation Token is used, the following claims can be used to meet those
requirements, whether these claims appear in Attestation Results, or in Evidence
for the Verifier to use when generating Attestation Results of some form:RequirementClaimReferenceDevice unique identifierueid section 3.4Vendor of the deviceoemid section 3.6Class of the devicehwmodel section 3.7TEE hardware typechip-version section 3.8TEE hardware versionchip-version section 3.8TEE firmware typesw-name section 3.9TEE firmware versionsw-version section 3.10Freshness proofnonce section 3.3The Update message is used by the TAM to install and/or delete one or more Trusted
Components via the TEEP Agent.Like other TEEP messages, the Update message is
signed, and the relevant CDDL snippet is shown below.
The complete CDDL structure is shown in Appendix C.The Update message has the following fields:
The value of (3) corresponds to an Update message sent from the TAM to
the TEEP Agent. In case of successful processing, a Success
message is returned by the TEEP Agent. In case of an error, an Error message
is returned. Note that the Update message
is used for initial Trusted Component installation as well as for updates
and deletes.
The value in the token field is used to match responses to requests.
The manifest-list field is used to convey one or multiple SUIT manifests
to install. A manifest is
a bundle of metadata about a Trusted Component, such as where to
find the code, the devices to which it applies, and cryptographic
information protecting the manifest. The manifest may also convey personalization
data. Trusted Component binaries and personalization data can be signed and encrypted
by the same Trusted Component Signer. Other combinations are, however, possible as well. For example,
it is also possible for the TAM to sign and encrypt the personalization data
and to let the Trusted Component Developer sign and/or encrypt the Trusted Component binary.Note that an Update message carrying one or more SUIT manifests will inherently
involve multiple signatures, one by the TAM in the TEEP message and one from
a Trusted Component Signer inside each manifest. This is intentional as they
are for different purposes.The TAM is what authorizes
apps to be installed, updated, and deleted on a given TEE and so the TEEP
signature is checked by the TEEP Agent at protocol message processing time.
(This same TEEP security wrapper is also used on messages like QueryRequest
so that Agents only send potentially sensitive data such as Evidence to
trusted TAMs.)The Trusted Component signer on the other hand is what authorizes the
Trusted Component to actually run, so the manifest signature could be
checked at install time or load (or run) time or both, and this checking is
done by the TEE independent of whether TEEP is used or some other update
mechanism.
See section 5 of for further discussion.The Update Message has a SUIT_Envelope containing SUIT manifests. Following are some examples of using SUIT manifests in the Update Message.In this example, a SUIT Manifest has a URI pointing to a Trusted Component Binary.A Trusted Component Developer creates a new Trusted Component Binary and hosts it at a Trusted Component Developer’s URI. Then the Trusted Component Developer generates an associated SUIT manifest with the filename “tc-uuid.suit” that contains the URI. The filename “tc-uuid.suit” is used in Example 3 later.The TAM receives the latest SUIT manifest from the Trusted Component Developer, and
the URI it contains will not be changeable by the TAM since the SUIT manifest is signed by the Trusted Component Developer.Pros:The Trusted Component Developer can ensure that the intact Trusted Component Binary is downloaded by devicesThe TAM does not have to send large Update messages containing the Trusted Component BinaryCons:The Trusted Component Developer must host the Trusted Component Binary serverThe device must fetch the Trusted Component Binary in another connection after receiving an Update messageFor the full SUIT Manifest example binary, see .In this example, the SUIT manifest contains the entire Trusted Component Binary using the integrated-payload (see Section 7.6).A Trusted Component Developer delegates to the TAM the task of delivering the Trusted Component Binary in the SUIT manifest. The Trusted Component Developer creates a SUIT manifest and embeds the Trusted Component Binary, which is referenced in the URI parameter with identifier “#tc”. The Trusted Component Developer provides the SUIT manifest to the TAM.The TAM serves the SUIT manifest containing the Trusted Component Binary to the device in an Update message.Pros:The device can obtain the Trusted Component Binary and its SUIT manifest together in one Update messageThe Trusted Component Developer does not have to host a server to deliver the Trusted Component Binary directly to devicesCons:The TAM must host the Trusted Component Binary itself, rather than delegating such storage to the Trusted Component DeveloperThe TAM must deliver Trusted Component Binaries in Update messages, which result in increased Update message sizeFor the full SUIT Manifest example binary, see .In this example, Personalization Data is associated with the Trusted Component Binary “tc-uuid.suit” from Example 1.The Trusted Component Developer places Personalization Data in a file named “config.json” and hosts it on an HTTPS server. The Trusted Component Developer then creates a SUIT manifest with the URI, specifying which Trusted Component Binary it correlates to in the parameter ‘dependency-resolution’, and signs the SUIT manifest.The TAM delivers the SUIT manifest of the Personalization Data which depends on the Trusted Component Binary from Example 1.For the full SUIT Manifest example binary, see .This subsection shows an example deleting the Trusted Component Binary in the TEEP Device.A Trusted Component Developer can also generate SUIT Manifest which unlinks the installed Trusted Component. The TAM deliver it when the TAM want to uninstall the component.The directive-unlink (see Section-6.5.4) is located in the manifest to delete the Trusted Component. Note that in case other Trusted Components depend on it, i.e. the reference count is not zero, the TEEP Device SHOULD NOT delete it immediately.For the full SUIT Manifest example binary, see Appendix E. SUIT Example 4The Success message is used by the TEEP Agent to return a success in
response to an Update message.Like other TEEP messages, the Success message is
signed, and the relevant CDDL snippet is shown below.
The complete CDDL structure is shown in Appendix C.The Success message has the following fields:
The value of (5) corresponds to corresponds to a Success message sent from the TEEP Agent to the
TAM.
The value in the token parameter is used to match responses to requests.
It MUST match the value of the token parameter in the Update
message the Success is in response to, if one was present. If none was
present, the token MUST be absent in the Success message.
The msg parameter contains optional diagnostics information encoded in
UTF-8 using Net-Unicode form with max 128 bytes
returned by the TEEP Agent.
If present, the suit-reports parameter contains a set of SUIT Reports
as defined in Section 4 of .
If a token parameter was present in the Update
message the Success message is in response to,
the suit-report-nonce field MUST be present in the SUIT Report with a
value matching the token parameter in the Update
message.The Error message is used by the TEEP Agent to return an error in
response to an Update message.Like other TEEP messages, the Error message is
signed, and the relevant CDDL snippet is shown below.
The complete CDDL structure is shown in Appendix C.The Error message has the following fields:
The value of (6) corresponds to an Error message sent from the TEEP Agent to the TAM.
The value in the token parameter is used to match responses to requests.
It MUST match the value of the token parameter in the Update
message the Success is in response to, if one was present. If none was
present, the token MUST be absent in the Error message.
The err-msg parameter is human-readable diagnostic text that MUST be encoded
using UTF-8 using Net-Unicode form with max 128 bytes.
The supported-cipher-suites parameter lists the ciphersuite(s) supported by the TEEP Agent.
Details about the ciphersuite encoding can be found in .
This otherwise optional parameter MUST be returned if err-code is ERR_UNSUPPORTED_CIPHER_SUITES.
The supported-freshness-mechanisms parameter lists the freshness mechanism(s) supported by the TEEP Agent.
Details about the encoding can be found in .
This otherwise optional parameter MUST be returned if err-code is ERR_UNSUPPORTED_FRESHNESS_MECHANISMS.
The versions parameter enumerates the TEEP protocol version(s) supported by the TEEP
Agent. This otherwise optional parameter MUST be returned if err-code is ERR_UNSUPPORTED_MSG_VERSION.
If present, the suit-reports parameter contains a set of SUIT Reports
as defined in Section 4 of . If
a token parameter was present in the Update message the Error message is in response to,
the suit-report-nonce field MUST be present in the SUIT Report with a
value matching the token parameter in the Update
message.
The err-code parameter contains one of the
error codes listed below). Only selected values are applicable
to each message.This specification defines the following initial error messages:
The TEEP
request contained incorrect fields or fields that are inconsistent with
other fields.
For diagnosis purposes it is RECOMMMENDED to identify the failure reason
in the error message.
A TAM receiving this error might refuse to communicate further with
the TEEP Agent for some period of time until it has reason to believe
it is worth trying again, but it should take care not to give up on
communication. In contrast, ERR_TEMPORARY_ERROR is an indication
that a more agressive retry is warranted.
The TEEP Agent does not support an extension included in the request
message.
For diagnosis purposes it is RECOMMMENDED to identify the unsupported
extension in the error message.
A TAM receiving this error might retry the request without using extensions.
The TEEP Agent does not
support any freshness algorithm mechanisms in the request message.
A TAM receiving this error might retry the request using a different
set of supported freshness mechanisms in the request message.
The TEEP Agent does not
support the TEEP protocol version indicated in the request message.
A TAM receiving this error might retry the request using a different
TEEP protocol version.
The TEEP Agent does not
support any ciphersuites indicated in the request message.
A TAM receiving this error might retry the request using a different
set of supported ciphersuites in the request message.
Processing of a certificate failed. For diagnosis purposes it is
RECOMMMENDED to include information about the failing certificate
in the error message. For example, the certificate was of an
unsupported type, or the certificate was revoked by its signer.
A TAM receiving this error might attempt to use an alternate certificate.
A certificate has expired or is not currently
valid.
A TAM receiving this error might attempt to renew its certificate
before using it again.
A miscellaneous
temporary error, such as a memory allocation failure, occurred while processing the request message.
A TAM receiving this error might retry the same request at a later point
in time.
The TEEP Agent encountered one or more manifest processing failures.
If the suit-reports parameter is present, it contains the failure details.
A TAM receiving this error might still attempt to install or update
other components that do not depend on the failed manifest.New error codes should be added sparingly, not for every implementation
error. That is the intent of the err-msg field, which can be used to
provide details meaningful to humans. New error codes should only be
added if the TAM is expected to do something behaviorally different upon
receipt of the error message, rather than just logging the event.
Hence, each error code is responsible for saying what the
behavioral difference is expected to be.The TEEP protocol operates between a TEEP Agent and a TAM. While
the TEEP protocol does not require use of EAT, use of EAT is encouraged and
explicitly defines a way to carry an Entity Attestation Token
in a QueryResponse.As discussed in , the content of Evidence is opaque to the TEEP
architecture, but the content of Attestation Results is not, where Attestation
Results flow between a Verifier and a TAM (as the Relying Party).
Although Attestation Results required by a TAM are separable from the TEEP protocol
per se, this section is included as part of the requirements for building
a compliant TAM that uses EATs for Attestation Results.Section 7 of defines the requirement for
Entity Attestation Token profiles. This section defines an EAT profile
for use with TEEP.profile-label: The profile-label for this specification is the URIhttps://datatracker.ietf.org/doc/html/draft-ietf-teep-protocol-09.
(RFC-editor: upon RFC publication, replace string with
“https://www.rfc-editor.org/info/rfcXXXX” where XXXX is the RFC number
of this document.)Use of JSON, CBOR, or both: CBOR only.CBOR Map and Array Encoding: Only definite length arrays and maps.CBOR String Encoding: Only definite-length strings are allowed.CBOR Preferred Serialization: Encoders must use preferred serialization,
and decoders need not accept non-preferred serialization.COSE/JOSE Protection: See .Detached EAT Bundle Support: DEB use is permitted.Verification Key Identification: COSE Key ID (kid) is used, where
the key ID is the hash of a public key (where the public key may be
used as a raw public key, or in a certificate).Endorsement Identification: Optional, but semantics are the same
as in Verification Key Identification.Freshness: See .Required Claims: None.Prohibited Claims: None.Additional Claims: Optional claims are those listed in .Refined Claim Definition: None.CBOR Tags: CBOR Tags are not used.Manifests and Software Evidence Claims: The sw-name claim for a Trusted
Component holds the URI of the SUIT manifest for that component.In COSE, arrays and maps use strings, negative integers, and unsigned
integers as their keys. Integers are used for compactness of
encoding. Since the word “key” is mainly used in its other meaning, as a
cryptographic key, this specification uses the term “label” for this usage
as a map key.This specification uses the following mapping:NameLabelsupported-cipher-suites1challenge2versions3selected-cipher-suite5selected-version6attestation-payload7tc-list8ext-list9manifest-list10msg11err-msg12attestation-payload-format13requested-tc-list14unneeded-tc-list15component-id16tc-manifest-sequence-number17have-binary18suit-reports19token20supported-freshness-mechanisms21Behavior is specified in terms of the conceptual APIs defined in
section 6.2.1 of .When the ProcessConnect API is invoked, the TAM sends a QueryRequest message.When the ProcessTeepMessage API is invoked, the TAM first does validation
as specified in , and drops the message if it is not valid.
Otherwise, it proceeds as follows.If the message includes a token, it can be used to
match the response to a request previously sent by the TAM.
The TAM MUST expire the token value after receiving the first response
from the device that has a valid signature and ignore any subsequent messages that have the same token
value. The token value MUST NOT be used for other purposes, such as a TAM to
identify the devices and/or a device to identify TAMs or Trusted Components.If a QueryResponse message is received, the TAM verifies the presence of any parameters
required based on the data-items-requested in the QueryRequest, and also validates that
the nonce in any SUIT Report matches the token send in the QueryRequest message if a token
was present. If these requirements are not met, the TAM drops the message. It may also do
additional implementation specific actions such as logging the results. If the requirements
are met, processing continues as follows.If a QueryResponse message is received that contains that contains Evidence, the Evidence
is passed to an attestation Verifier (see )
to determine whether the Agent is in a trustworthy state. Once the TAM receives Attestation
Results, processing continues as follows.Based on the results of attestation (if any), any SUIT Reports,
and the lists of installed, requested,
and unneeded Trusted Components reported in the QueryResponse, the TAM
determines, in any implementation specific manner, which Trusted Components
need to be installed, updated, or deleted, if any.
If any Trusted Components need to be installed, updated, or deleted,
the TAM sends an Update message containing SUIT Manifests with command
sequences to do the relevant installs, updates, or deletes.
It is important to note that the TEEP Agent’s
Update Procedure requires resolving and installing any dependencies
indicated in the manifest, which may take some time, and the resulting Success
or Error message is generated only after completing the Update Procedure.
Hence, depending on the freshness mechanism in use, the TAM may need to
store data (e.g., a nonce) for some time.If a Success or Error message is received containing one or more SUIT Reports, the TAM also validates that
the nonce in any SUIT Report matches the token sent in the Update message,
and drops the message if it does not match. Otherwise, the TAM handles
the update in any implementation specific way, such as updating any locally
cached information about the state of the TEEP Agent, or logging the results.If any other Error message is received, the TAM can handle it in any implementation
specific way, but provides recommendations for such handling.When the RequestTA API is invoked, the TEEP Agent first checks whether the
requested TA is already installed. If it is already installed, the
TEEP Agent passes no data back to the caller. Otherwise,
if the TEEP Agent chooses to initiate the process of requesting the indicated
TA, it determines (in any implementation specific way) the TAM URI based on
any TAM URI provided by the RequestTA caller and any local configuration,
and passes back the TAM URI to connect to. It MAY also pass back a
QueryResponse message if all of the following conditions are true:The last QueryRequest message received from that TAM contained no token or challenge,The ProcessError API was not invoked for that TAM since the last QueryResponse
message was received from it, andThe public key or certificate of the TAM is cached and not expired.When the RequestPolicyCheck API is invoked, the TEEP Agent decides
whether to initiate communication with any trusted TAMs (e.g., it might
choose to do so for a given TAM unless it detects that it has already
communicated with that TAM recently). If so, it passes back a TAM URI
to connect to. If the TEEP Agent has multiple TAMs it needs to connect
with, it just passes back one, with the expectation that
RequestPolicyCheck API will be invoked to retrieve each one successively
until there are no more and it can pass back no data at that time.
Thus, once a TAM URI is returned, the TEEP Agent can remember that it has
already initiated communication with that TAM.When the ProcessError API is invoked, the TEEP Agent can handle it in
any implementation specific way, such as logging the error or
using the information in future choices of TAM URI.When the ProcessTeepMessage API is invoked, the Agent first does validation
as specified in , and drops the message if it is not valid.
Otherwise, processing continues as follows based on the type of message.When a QueryRequest message is received, the Agent responds with a
QueryResponse message if all fields were understood, or an Error message
if any error was encountered.When an Update message is received, the Agent attempts to update
the Trusted Components specified in the SUIT manifests
by following the Update Procedure specified
in , and responds with a Success message if
all SUIT manifests were successfully installed, or an Error message
if any error was encountered.
It is important to note that the
Update Procedure requires resolving and installing any dependencies
indicated in the manifest, which may take some time, and the Success
or Error message is generated only after completing the Update Procedure.The TEEP protocol uses COSE for protection of TEEP messages.
After a QueryResponse is received, the selected cryptographic algorithm is used in subsequent TEEP messages (Install, Success, and Error).
To negotiate cryptographic mechanisms and algorithms, the TEEP protocol defines the following ciphersuite structure.The ciphersuite structure is used to present the combination of mechanisms and cryptographic algorithms.
Each ciphersuite value corresponds with a COSE-type defined in Section 2 of .Cryptographic algorithm values are defined in the COSE Algorithms registry .
A TAM MUST support both of the following ciphersuites. A TEEP Agent MUST support at least
one of the two but can choose which one. For example, a TEEP Agent might
choose a given ciphersuite if it has hardware support for it.A TAM or TEEP Agent MUST also support the following algorithms:A TAM or TEEP Agent MAY also support one or more of the following algorithms:Any ciphersuites without confidentiality protection can only be added if the
associated specification includes a discussion of security considerations and
applicability, since manifests may carry sensitive information. For example,
Section 6 of permits implementations that
terminate transport security inside the TEE and if the transport security
provides confidentiality then additional encryption might not be needed in
the manifest for some use cases. For most use cases, however, manifest
confidentiality will be needed to protect sensitive fields from the TAM as
discussed in Section 9.8 of .A freshness mechanism determines how a TAM can tell whether an attestation payload provided
in a Query Response is fresh. There are multiple ways this can be done
as discussed in Section 10 of .Each freshness mechanism is identified with an integer value, which corresponds to
an IANA registered freshness mechanism (see .
This document defines the following freshness mechanisms:ValueFreshness mechanism1Nonce2Timestamp3Epoch IDIn the Nonce mechanism, the attestation payload MUST include a nonce provided
in the QueryRequest challenge. In other mechanisms, a timestamp
or epoch ID determined via mechanisms outside the TEEP protocol is
used, and the challenge is only needed in the QueryRequest message
if a challenge is needed in generating the attestation payload for reasons other
than freshness.If a TAM supports multiple freshness mechanisms that require different challenge
formats, the QueryRequest message can currently only send one such challenge.
This situation is expected to be rare, but should it occur, the TAM can
choose to prioritize one of them and exclude the other from the
supported-freshness-mechanisms in the QueryRequest, and resend the QueryRequest
with the other mechanism if an ERR_UNSUPPORTED_FRESHNESS_MECHANISMS Error
is received that indicates the TEEP Agent supports the other mechanism.This section summarizes the security considerations discussed in this
specification:
TEEP protocol messages exchanged between the TAM and the TEEP Agent
are protected using COSE. This specification relies on the
cryptographic algorithms provided by COSE. Public key based
authentication is used by the TEEP Agent to authenticate the TAM
and vice versa.
A TAM relies on signed Attestation Results provided by a Verifier,
either obtained directly using a mechanism outside the TEEP protocol
(by using some mechanism to pass Evidence obtained in the attestation payload of
a QueryResponse, and getting back the Attestation Results), or indirectly
via the TEEP Agent forwarding the Attestation Results in the attestation
payload of a QueryResponse. See the security considerations of the
specific mechanism in use (e.g., EAT) for more discussion.
Depending on
the properties of the attestation mechanism, it is possible to
uniquely identify a device based on information in the
attestation payload or in the certificate used to sign the
attestation payload. This uniqueness may raise privacy concerns. To lower the
privacy implications the TEEP Agent MUST present its
attestation payload only to an authenticated and authorized TAM and when using
EATS, it SHOULD use encryption as discussed in , since
confidentiality is not provided by the TEEP protocol itself and
the transport protocol under the TEEP protocol might be implemented
outside of any TEE. If any mechanism other than EATs is used, it is
up to that mechanism to specify how privacy is provided.
Each Trusted Component binary is signed by a Trusted Component Signer. It is the responsibility of the
TAM to relay only verified Trusted Components from authorized Trusted Component Signers. Delivery of
a Trusted Component to the TEEP Agent is then the responsibility of the TAM,
using the security mechanisms provided by the TEEP
protocol. To protect the Trusted Component binary, the SUIT manifest format is used and
it offers a variety of security features, including digitial
signatures and symmetric encryption.
A Trusted Component Signer or TAM can supply personalization data along with a Trusted Component.
This data is also protected by a SUIT manifest.
Personalization data signed and encrypted by a Trusted Component Signer other than
the TAM is opaque to the TAM.
As discussed in section 6 of ,
the TEEP protocol typically relies on a TEEP Broker to relay messages
between the TAM and the TEEP Agent. When the TEEP Broker is
compromised it can drop messages, delay the delivery of messages,
and replay messages but it cannot modify those messages. (A replay
would be, however, detected by the TEEP Agent.) A compromised TEEP
Broker could reorder messages in an attempt to install an old
version of a Trusted Component. Information in the manifest ensures that TEEP
Agents are protected against such downgrade attacks based on
features offered by the manifest itself.
A TAM is responsible for vetting a Trusted Component and
before distributing them to TEEP Agents.
It is RECOMMENDED to provide a way to
update the trust anchor store used by the TEE, for example using
a firmware update mechanism. Thus, if a Trusted Component
Signer is later compromised, the TAM can update the trust anchor
store used by the TEE, for example using a firmware update mechanism.
The CA issuing certificates to a TEE or a Trusted Component Signer might get compromised.
It is RECOMMENDED to provide a way to
update the trust anchor store used by the TEE, for example using
a firmware update mechanism. If the CA issuing certificates to
devices gets compromised then these devices might be rejected by a
TAM, if revocation is available to the TAM.
The integrity and the accuracy of the
clock within the TEE determines the ability to determine an expired
TAM certificate, if certificates are used.
As discussed above, certificate validity checks rely on comparing
validity dates to the current time, which relies on having a trusted
source of time, such as . A compromised time source could
thus be used to subvert such validity checks.IANA is requested to assign a media type for
application/teep+cbor.
application
teep+cbor
none
none
Same as encoding considerations of
application/cbor.
See Security Considerations Section of this document.
Same as interoperability
considerations of application/cbor as specified in .
This document.
TEEP protocol implementations
N/A
N/A
N/A
N/A
N/A
teep@ietf.org
COMMON
none
See the “Authors’ Addresses” section of this document
IETFIANA is also requested to create a new registry for freshness mechanisms.Name of registry: TEEP Freshness MechanismsPolicy: Specification Required Additional requirements: The specification must document relevant security considerations.Initial values:ValueFreshness mechanismSpecification1NonceRFC TBD 2TimestampRFC TBD 3Epoch IDRFC TBD (RFC Editor: please replace TBD above with the number assigned to this document.)CBOR Object Signing and Encryption (COSE)Concise Binary Object Representation (CBOR) is a data format designed for small code size and small message size. There is a need for the ability to have basic security services defined for this data format. This document defines the CBOR Object Signing and Encryption (COSE) protocol. This specification describes how to create and process signatures, message authentication codes, and encryption using CBOR for serialization. This specification additionally describes how to represent cryptographic keys using CBOR.UTF-8, a transformation format of ISO 10646ISO/IEC 10646-1 defines a large character set called the Universal Character Set (UCS) which encompasses most of the world's writing systems. The originally proposed encodings of the UCS, however, were not compatible with many current applications and protocols, and this has led to the development of UTF-8, the object of this memo. UTF-8 has the characteristic of preserving the full US-ASCII range, providing compatibility with file systems, parsers and other software that rely on US-ASCII values but are transparent to other values. This memo obsoletes and replaces RFC 2279.Unicode Format for Network InterchangeThe Internet today is in need of a standardized form for the transmission of internationalized "text" information, paralleling the specifications for the use of ASCII that date from the early days of the ARPANET. This document specifies that format, using UTF-8 with normalization and specific line-ending sequences. [STANDARDS-TRACK]Internet X.509 Public Key Infrastructure Certificate and Certificate Revocation List (CRL) ProfileThis memo profiles the X.509 v3 certificate and X.509 v2 certificate revocation list (CRL) for use in the Internet. An overview of this approach and model is provided as an introduction. The X.509 v3 certificate format is described in detail, with additional information regarding the format and semantics of Internet name forms. Standard certificate extensions are described and two Internet-specific extensions are defined. A set of required certificate extensions is specified. The X.509 v2 CRL format is described in detail along with standard and Internet-specific extensions. An algorithm for X.509 certification path validation is described. An ASN.1 module and examples are provided in the appendices. [STANDARDS-TRACK]Concise Binary Object Representation (CBOR)The Concise Binary Object Representation (CBOR) is a data format whose design goals include the possibility of extremely small code size, fairly small message size, and extensibility without the need for version negotiation. These design goals make it different from earlier binary serializations such as ASN.1 and MessagePack.Remote Attestation Procedures ArchitectureFraunhofer SITMicrosoftSandelman Software WorksIntel CorporationHuawei Technologies In network protocol exchanges it is often useful for one end of a
communication to know whether the other end is in an intended
operating state. This document provides an architectural overview of
the entities involved that make such tests possible through the
process of generating, conveying, and evaluating evidentiary claims.
An attempt is made to provide for a model that is neutral toward
processor architectures, the content of claims, and protocols.
The Entity Attestation Token (EAT)Security Theory LLCQualcomm Technologies Inc.Qualcomm Technologies Inc. An Entity Attestation Token (EAT) provides an attested claims set
that describes state and characteristics of an entity, a device like
a phone, IoT device, network equipment or such. This claims set is
used by a relying party, server or service to determine how much it
wishes to trust the entity.
An EAT is either a CBOR Web Token (CWT) or JSON Web Token (JWT) with
attestation-oriented claims. To a large degree, all this document
does is extend CWT and JWT.
A Concise Binary Object Representation (CBOR)-based Serialization Format for the Software Updates for Internet of Things (SUIT) ManifestArm LimitedArm LimitedFraunhofer SITInria This specification describes the format of a manifest. A manifest is
a bundle of metadata about code/data obtained by a recipient (chiefly
the firmware for an IoT device), where to find the that code/data,
the devices to which it applies, and cryptographic information
protecting the manifest. Software updates and Trusted Invocation
both tend to use sequences of common operations, so the manifest
encodes those sequences of operations, rather than declaring the
metadata.
SUIT Manifest Extensions for Multiple Trust DomainsArm Limited This specification describes extensions to the SUIT manifest format
(as defined in [I-D.ietf-suit-manifest]) for use in deployments with
multiple trust domains. A device has more than one trust domain when
it uses different trust anchors for different purposes or components
in the context of firmware update.
Secure Reporting of Update StatusArm LimitedFraunhofer SIT The Software Update for the Internet of Things (SUIT) manifest
provides a way for many different update and boot workflows to be
described by a common format. However, this does not provide a
feedback mechanism for developers in the event that an update or boot
fails.
This specification describes a lightweight feedback mechanism that
allows a developer in possession of a manifest to reconstruct the
decisions made and actions performed by a manifest processor.
COSE AlgorithmsIANAKey words for use in RFCs to Indicate Requirement LevelsIn many standards track documents several words are used to signify the requirements in the specification. These words are often capitalized. This document defines these words as they should be interpreted in IETF documents. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements.Ambiguity of Uppercase vs Lowercase in RFC 2119 Key WordsRFC 2119 specifies common key words that may be used in protocol specifications. This document aims to reduce the ambiguity by clarifying that only UPPERCASE usage of the key words have the defined special meanings.Trusted Execution Environment Provisioning (TEEP) ArchitectureBroadcomArm LimitedMicrosoftAmazon A Trusted Execution Environment (TEE) is an environment that enforces
that any code within that environment cannot be tampered with, and
that any data used by such code cannot be read or tampered with by
any code outside that environment. This architecture document
motivates the design and standardization of a protocol for managing
the lifecycle of trusted applications running inside such a TEE.
EAT Media TypesSecurity Theory LLCFraunhofer Institute for Secure Information Technologyarm Payloads used in Remote Attestation Procedures may require an
associated media type for their conveyance, for example when used in
RESTful APIs.
This memo defines media types to be used for Entity Attestation
Tokens (EAT).
Concise Data Definition Language (CDDL): A Notational Convention to Express Concise Binary Object Representation (CBOR) and JSON Data StructuresThis document proposes a notational convention to express Concise Binary Object Representation (CBOR) data structures (RFC 7049). Its main goal is to provide an easy and unambiguous way to express structures for protocol messages and data formats that use CBOR or JSON.Guidelines for Writing an IANA Considerations Section in RFCsMany protocols make use of points of extensibility that use constants to identify various protocol parameters. To ensure that the values in these fields do not have conflicting uses and to promote interoperability, their allocations are often coordinated by a central record keeper. For IETF protocols, that role is filled by the Internet Assigned Numbers Authority (IANA).To make assignments in a given registry prudently, guidance describing the conditions under which new values should be assigned, as well as when and how modifications to existing values can be made, is needed. This document defines a framework for the documentation of these guidelines by specification authors, in order to assure that the provided guidance for the IANA Considerations is clear and addresses the various issues that are likely in the operation of a registry.This is the third edition of this document; it obsoletes RFC 5226.Network Time Security for the Network Time ProtocolThis memo specifies Network Time Security (NTS), a mechanism for using Transport Layer Security (TLS) and Authenticated Encryption with Associated Data (AEAD) to provide cryptographic security for the client-server mode of the Network Time Protocol (NTP). NTS is structured as a suite of two loosely coupled sub-protocols. The first (NTS Key Establishment (NTS-KE)) handles initial authentication and key establishment over TLS. The second (NTS Extension Fields for NTPv4) handles encryption and authentication during NTP time synchronization via extension fields in the NTP packets, and holds all required state only on the client via opaque cookies.We would like to thank Brian Witten (Symantec), Tyler Kim (Solacia), Nick Cook (Arm), and Minho Yoo (IoTrust) for their contributions
to the Open Trust Protocol (OTrP), which influenced the design of this specification.We would like to thank Eve Schooler for the suggestion of the protocol name.We would like to thank Kohei Isobe (TRASIO/SECOM), Ken Takayama (SECOM)
Kuniyasu Suzaki (TRASIO/AIST), Tsukasa Oi (TRASIO), and Yuichi Takita (SECOM)
for their valuable implementation feedback.We would also like to thank Carsten Bormann and Henk Birkholz for their help with the CDDL.Valid TEEP messages MUST adhere to the following CDDL data definitions,
except that SUIT_Envelope and SUIT_Component_Identifier are
specified in .This section includes some examples with the following assumptions:The device will have two TCs with the following SUIT Component Identifiers:
[ 0x000102030405060708090a0b0c0d0e0f ][ 0x100102030405060708090a0b0c0d0e0f ]SUIT manifest-list is set empty only for example purposes (see Appendix E
for actual manifest examples)This is shown below in CBOR diagnostic form. Only the payload signed by
COSE is shown.This section shows some examples of SUIT manifests described in .The examples are signed using the following ECDSA secp256r1 key with SHA256 as the digest function.COSE_Sign1 Cryptographic Key:The corresponding public key can be used to verify these examples:This section shows some examples of SUIT reports.SUIT Reports have no records if no conditions have failed.
The URI in this example is the reference URI provided in the SUIT manifest.where the dependency-resolution refers to:and the suit-record-section-offset refers to: