Network Device Remote Integrity Verification

There are many aspects to consider in fielding a trusted computing device, from operating systems to applications. A mechanism to prove that a device installed at a customer’s site is authentic (i.e., not counterfeit) and has been configured with authorized software, all as part of a trusted supply chain, is one of those aspects that’s easily overlooked. Attestation is defined here as the process of creating, conveying and appraising assertions about Platform trustworthiness characteristics, including supply chain trust, identity, platform provenance, software configuration, hardware configuration, platform composition, compliance to test suites, functional and assurance evaluations, etc. The goal of attestation is simply to assure an administrator that the software that was launched when the device was last started is the same as the software that the device vendor initially shipped. Within the Trusted Computing Group context, attestation is the process by which an independent Verifier can obtain cryptographic proof as to the identity of the device in question, evidence of the integrity of software loaded on that device when it started up, and then verify that what’s there is what’s supposed to be there. For networking equipment, a verifier capability can be embedded in a Network Management Station (NMS), a posture collection server, or other network analytics tool (such as a software asset management solution, or a threat detection and mitigation tool, etc.). While informally referred to as attestation, this document focuses on a subset defined here as Remote Integrity Verification (RIV). RIV takes a network equipment centric perspective that includes a set of protocols and procedures for determining whether a particular device was launched with untampered software, starting from Roots of Trust. While there are many ways to accomplish attestation, RIV sets out a specific set of protocols and tools that work in environments commonly found in Networking Equipment. RIV does not cover other platform characteristics that could be attested, although it does provide evidence of a secure infrastructure to increase the level of trust in other platform characteristics attested by other means (e.g., by Entity Attestation Tokens . This profile outlines the RIV problem, and then identifies elements that are necessary to get the complete attestation procedure working in a scalable solution using commercial products, focusing primarily on software integrity verification using the Trusted Platform Module (TPM) to ensure a trustworthy result.

The remainder of this document is organized into several sections: The remainder of this section covers goals and requirements, plus a top-level overview The Solution Overview section outlines how RIV works The Standards Components section links components of RIV to normative standards. Supporting material is in an appendix at the end.

This document is non-normative; the document does not define protocols, but rather identifies existing protocols that can be used together to achieve the goals above, and in some cases, highlights gaps in existing protocols.

Network operators benefit from a trustworthy attestation mechanism that provides assurance that their network comprises authentic equipment, and has loaded software free of known vulnerabilities and unauthorized tampering. In line with the overall goal of assuring integrity, attestation can be used for asset management, vulnerability and compliance assessment, plus configuration management. As a part of a trusted supply chain, the RIV attestation workflow outlined in this document is intended to meet the following high-level goals: Provable Device Identity - The ability to identify a device using a cryptographic identifier is a critical prerequisite to software inventory attestation. Software Inventory - A key goal is to identify the software release installed on the device, and to provide evidence of its integrity. Verifiability - Verification of software and configuration of the device shows that the software that’s supposed to be installed on there actually has been launched. Verification against reference manifests signed by the supplier of the software provides assurance that the software is free of unauthorized modification. In addition, RIV is designed to operate in a centralized environment, such as with a central authority that manages and configures a number of network devices, or ‘peer-to-peer’, where network devices independently verify one another to establish a trust relationship. (See Section XX below, and also )

Attestation requires two interlocking services from the device to be verified: Platform Identity, the mechanism providing trusted identity, can reassure network managers that the specific devices they ordered from authorized manufacturers for attachment to their network are those that were installed, and that they continue to be present in their network. As part of the mechanism for Platform Identity, cryptographic proof of the identity of the manufacturer is also provided. Software Measurement is the mechanism that reports the state of mutable software components on the device, and can assure network managers that they have known, untampered software configured to run in their network. In this context, RIV is a procedure that assures a network operator that the equipment in their network can be reliably identified, and that untampered software of a known version is installed on each endpoint. Equipment in the network includes devices that make up the network itself, such as routers, switches and firewalls, but could also include conventional connected devices like servers and laptops. (See Section XX, Scope) RIV includes several major processes: Creation of Evidence is the process whereby an endpoint generates cryptographic proof (evidence) of claims about platform properties. In particular, the platform identity and its software configuration are of critical importance Platform Identification refers to the mechanism assuring the attestation relying party (typically a network administrator) of the identity of devices that make up their network, and that their manufacturers are known. Reliable identify is clearly a prerequisite to software attestation to prevent spoofing attacks (See XX Security Considerations). Software used to boot a platform can be described as a chain of measurements, started by a Root of Trust for Measurement, that normally ends when the system software is loaded. A measurement signifies the identity, integrity and version of each software component registered with the TPM, so that the subsequent appraisal stage can determine if the software installed is authentic, up-to-date, and free of tampering. Conveyance of Evidence is the process of reliably transporting evidence from the point in a connected device where a measurement is stored to an appraiser/verifier, e.g. a management station. The transport is typically carried out via a management network. The channel must provide integrity and authenticity, and, in some use cases, may also require confidentiality. Finally, Appraisal of Evidence is the process of verifying the evidence received by a verifier/appraiser from a device, and using verified evidence to inform decision making. In this context, verification means comparing the device measurements reported as evidence with the configuration expected by the system administrator. This step depends on a way to express what should be there, coded as Reference Integrity Measurements (aka Golden Measurements), representing the intended configured state of the connected device. An implementation of RIV requires three technologies Identity: Platform identity can be based on IEEE 802.1AR Device Identity , coupled with careful supply-chain management by the manufacturer. The DevID certificate contains a statement by the manufacturer that establishes the identity of the device as it left the factory. Some applications with a more-complex post-manufacture supply chain (e.g. Value Added Resellers), or with different privacy concerns, may want to use an alternate mechanism for platform authentication based on TCG Platform Certificates . RIV currently relies on asymmetric keying for identity; alternate approaches might use symmetric keys. Platform Attestation provides evidence of configuration of software elements present in the device. This form of attestation can be implemented with TPM PCR, Quote and Log mechanisms, which provide an authenticated mechanism to report what software was started on the device through the boot cycle. Reference Integrity Measurements must be conveyed from the software authority (often the manufacturer for embedded systems) to the system in which verification will take place

Remote Integrity Verification must address the “Lying Endpoint” problem, in which malicious software on an endpoint may subvert the intended function, and also prevent the endpoint from reporting its compromised status. (See Section XX for further Security Considerations) The RIV attestation solution must meet a number of requirements to make it simple to deploy at scale. Easy to Use - RIV should work “out of the box” as far as possible, that is, with the fewest possible steps needed at the end-user’s site. Eliminate complicated databases or provisioning steps that would have to be executed by the owner of a new device. Network equipment is often required to “self-configure”, to reliably reach out without manual intervention to prove its identity and operating posture, then download its own configuration. See for an example of Secure Zero Touch Provisioning. Multi-Vendor - This solution identifies standards-based interfaces that allow attestation to work with attestation-capable devices and verifiers supplied by different vendors in one network.

Remote Attestation is a very general problem that could apply to most network-connected computing devices. However, this document includes several assumptions that limit the scope to Network Equipment (e.g. routers, switches and firewalls): This solution is for use in non-privacy-preserving applications (for example, networking, Industrial IoT), avoiding the need for a Privacy Certificate Authority for attestation keys or TCG Platform Certificates This document assumes network protocols that are common in networking equipment, but not generally used in other applications. Other applications (such as Industrial IoT) would replace YANG/Netconf with a protocol suite more commonly used in that environment. Similarly, the same information flow would be used in Network Endpoint Assessment , mapped to NEA protocols. The approach outlined in this document assumes the use of TPM 1.2 or TPM 2.0. Other roots of trust could be used with the same information flow, although different data structures would likely be called for.

Run-Time Attestation: Run-time attestation of Linux or other multi-threaded operating system processes considerably expands the scope of the problem. Many researchers are working on that problem, but this document defers the run-time attestation problem. Multi-Vendor Embedded Systems: Additional coordination would be needed for devices that themselves comprise hardware and software from multiple vendors, integrated by the end user. Although currently out of scope for RIV, Trusted Computing Group specifications do encompass attestation of multi-vendor endpoint computing devices. Processor Sleep Modes: Network equipment typically does not “sleep”, so sleep and hibernate modes are not considered. Although out of scope for RIV, Trusted Computing Group specifications do encompass sleep and hibernate states. Virtualization and Containerization: These technologies are increasingly used in Network equipment, but are not considered in this revision of the document.

RIV Attestation is a process for determining the identity of software running on a specifically-identified device. Remote Attestation is broken into two phases, shown in Figure 1: During system startup, each distinct software object is “measured”. Its identity, hash (i.e. cryptographic digest) and version information is recorded in a log. Hashes are also extended, or cryptographically folded, into the TPM, in a way that can be used to validate the log entries. The measurement process generally follows the Chain of Trust model used in Measured Boot, where each stage of the system measures the next one, and extends its measurement into the TPM, before launching it. Once the device is running and has operational network connectivity, a separate, trusted server (called a Verifier in this document) can interrogate the network device to retrieve the logs and a copy of the digests collected by hashing each software object, signed by an attestation private key known only to the TPM. The result is that the Verifier can verify the device’s identity by checking the certificate containing the TPM’s attestation public key, and can validate the software that was launched by comparing digests in the log with known-good values, and verifying their correctness by comparing with the signed digests from the TPM. It should be noted that attestation and identity are inextricably linked; signed evidence that a particular version of software was loaded is of little value without cryptographic proof of the identity of the device producing the evidence.

| Loader |-->| Kernel |--->|Userland | | | +--------+ +--------+ +--------+ +---------+ | | | | | | | | | | | | +------------+-----------+-+ | | Step 1 | | | V | | +--------+ | | | TPM | | | +--------+ | | Router | | +--------------------------------|----------------------+ | | Step 2 | +-----------+ +--->| Verifier | +-----------+ Reset---------------flow-of-time-during-boot--...-------> ]]> In Step 1, measurements are “extended” into the TPM as processes start. In Step 2, signed PCR digests are retrieved from the TPM for off-box analysis after the system is operational.

TPM attestation is strongly focused around Platform Configuration Registers (PCRs), but those registers are only vehicles for certifying attestation evidence. The evidence itself is conveyed in log files (xref) which give the name and hash of each object to be attested. These hashes are also extended into PCRs, where they can be retrieved in the form of a quote signed by a key known only to the TPM (xref). The use of multiple PCRs serves only to provide some independence between different classes of object, so that one class of objects can be updated without changing the extended hash for other classes. Although PCRs can be used for any purpose, this section outlines the objects that are conventionally attested with a TPM. In general, PCRs are organized to independently attest three classes of object: Code, i.e., instructions to be executed by a CPU. To maintain a chain of trust, it’s important to measure every bit of code that’s run outside the root of trust. Configuration - Many devices offer numerous options controlled by non-volatile configuration variables which can impact the device’s security posture. These settings may have vendor defaults, but often can be changed by administrators, who may want to verify via attestation that the settings they intend are still in place. Credentials - Administrators may wish to verify via attestation that keys (and other credentials) outside the Root of Trust have not be subject to unauthorized tampering. (By definition, keys inside the root of trust can’t be verified independently) The TCG PC Client Platform Firmware Profile Specification gives considerable detail on what is to be measured during the boot phase of a platform boot using a UEFI BOIS (www.uefi.org), but the goal is simply to measure every bit of code executed in the process of starting the device, along with any configuration information related to security posture. Table XX summarizes the functions that are measured, and how they are allocated to PCRs. It’s important to note that each PCR may contain results from dozens (or even thousands) of individual measurements.

TPM 1.2 and TPM 2.0 have a variety of rules separating the functions of identity and attestation, allowing for use-cases where software configuration must be attested, but privacy must be maintained. To accommodate these rules, enforced inside the TPM, in an environment where device privacy is not normally a requirement, the TCG Guidance for Securing Network Equipment suggests using separate keys for Identity (i.e., DevID) and Attestation (i.e., signing a quote of the contents of the PCRs), but provisioning an Initial Attestation Key (IAK) and x.509 certificate that parallels the IDevID, with the same device ID information as the IDevID certificate (i.e., the same Subject Name and Subject Alt Name, even though the key pairs are different). This allows a quote from the device, signed by the IAK, to be linked directly to the device that provided it, by examining the corresponding IAK certificate. Inclusion of an IAK by a vendor does not preclude a mechanism whereby an Administrator can define Local Attestation Keys (LAKs) if desired.

RIV workflow for networking equipment is organized around a simple use-case, where a network operator wishes to verify the integrity of software installed in specific, fielded devices. This use-case implies several components: A Device (e.g. a router or other embedded device, also known as an Attester) somewhere and the network operator wants to examine its boot state. A Verifier (which might be a network management station) somewhere separate from the Device that will retrieve the information and analyze it to pass judgment on the security posture of the device. A Relying Party, which has access to the Verifier to request attestation and to act on results. Interaction between the Relying Party and the Verifier is considered out of scope for RIV. Signed Reference Integrity Manifests (RIMs) (containing Reference Integrity Measurements, or “golden measurements”) can either be created by the device manufacturer and shipped along with the device as part of its software image, or alternatively, could be obtained a number of other ways (direct to the verifier from the manufacturer, from a third party, from the owner’s observation of what’s thought to be a “known good system”, etc.). Retrieving RIMs from the device itself allows attestation to be done in systems which may not have access to the public internet, or by other devices that are not management stations per-se (e.g., a peer device). If reference measurements are obtained from multiple sources, the Verifier may need to evaluate the relative level of trust to be placed in each source in case of a discrepancy. These components are illustrated in Figure 2. A more-detailed taxonomy of terms is given in

| Mngmt | Party | | Manufacturer | | attestation)| Step 2 | Station)| | | or other | | | | | | | authority) | | | | | | +---------------+ +-------------+ +---------+--------+ | /\ | Step 0 | ----------------------------------------------- ]]> In Step 0, The Asserter (the device manufacturer) provides a Software Image accompanied by one or more Reference Integrity Manifests (RIMs) to the Attester (the device under attestation) signed by the asserter. In Step 1, the Verifier (Network Management Station), on behalf of a Relying Party, requests Identity, Measurement Values (and possibly RIMs) from the Attester. In Step 2, the Attester responds to the request by providing a DevID, quotes (measured values), and optionally RIMs, signed by the Attester. See for more narrowly defined terms related to Attestation

This document makes the following simplifying assumptions to reduce complexity: The product to be attested is shipped with an IEEE 802.1AR DevID and an Initial Attestation Key (IAK) with certificate. The IAK cert contains the same identity information as the DevID (specifically, the same Subject Name and Subject Alt Name, signed by the manufacturer), but it’s a type of key that can be used to sign a TPM Quote. This convention is described in TCG Guidance for Securing Network Equipment . For network equipment, which is generally non-privacy-sensitive, shipping a device with both an IDevID and an IAK already provisioned substantially simplifies initial startup. Privacy-sensitive applications may use the TCG Platform Certificate and additional procedures to install identity credentials on the platform after manufacture. (See Section 2.3.1 below for the Platform Certificate alternative) The product is equipped with a Root of Trust for Measurement, Root of Trust for Storage and Root of Trust for Reporting (as defined in ) that are capable of conforming to the TCG Trusted Attestation Protocol (TAP) Information Model . The vendor will ship Reference Integrity Measurements (i.e., known-good measurements) in the form of signed CoSWID tags , , as described in TCG Reference Integrity Measurement Manifest Information Model .

Some situations may have privacy-sensitive requirements that preclude shipping every device with an Initial Device ID installed. In these cases, the IDevID can be installed remotely using the TCG Platform Certificate . Some administrators may want to install their own identity credentials to certify device identity and attestation results. IEEE 802.1AR allows for both Initial Device Identity credentials, installed by the manufacturer, (analogous to a physical serial number plate), or Local Device Identity credentials installed by the administrator of the device (analogous to the physical Asset Tag used by many enterprises to identify their property). TCG TPM 2.0 Keys documents and specifies parallel Initial and Local Attestation Keys (IAK and LAK), and contains figures showing the relationship between IDevID, LDevID, IAK and LAK keys. Device administrators are free to use any number of criteria to judge authenticity of a device before installing local identity keys, as part of an on-boarding process. The TCG TPM 2.0 Keys document also outlines procedures for creating Local Attestation Keys and Local Device IDs (LDevIDs) rooted in the manufacturer’s IDevID as a check to reduce the chances that counterfeit devices are installed in the network. Note that many networking devices are expected to self-configure (aka Zero Touch Provisioning). Current standardized zero-touch mechanisms such as assume that identity keys are already in place before network on-boarding can start, and as such, are compatible with IDevID and IAK keys installed by the manufacturer, but not with LDevID and LAK keys, which would have to be installed by the administrator.

The Platform Attribute Credential can also be used to convey additional information about a platform from the manufacturer or other entities in the supply chain. While outside the scope of RIV, the Platform Attribute Credential can deliver information such as lists of serial numbers for components embedded in a device or security assertions related to the platform, signed by the manufacturer, system integrator or value-added-reseller.

The measurements needed for attestation require that the device being attested is equipped with a Root of Trust for Measurement, i.e., some trustworthy mechanism that can compute the first measurement in the chain of trust required to attest that each stage of system startup is verified, and a Root of Trust for Reporting to report the results , . While there are many complex aspects of a Root of Trust, two aspects that are important in the case of attestation are: The first measurement computed by the Root of Trust for Measurement, and stored in the TPM’s Root of Trust for Storage, is presumed to be correct. There must not be a way to reset the RTS without re-entering the RTM code. The first measurement must be computed by code that is implicitly trusted; if that first measurement can be subverted, none of the remaining measurements can be trusted. (See )

Much of attestation focuses on collecting and transmitting evidence in the form of PCR measurements and attestation logs. But the critical part of the process is enabling the verifier to decide whether the measurements are “the right ones” or not. While it must be up to network administrators to decide what they want on their networks, the software supplier should supply the Reference Integrity Measurements that may be used by a verifier to determine if evidence shows known good, known bad or unknown software configurations. In general, there are two kinds of reference measurements: Measurements of early system startup (e.g., BIOS, boot loader, OS kernel) are essentially single threaded, and executed exactly once, in a known sequence, before any results could be reported. In this case, while the method for computing the hash and extending relevant PCRs may be complicated, the net result is that the software (more likely, firmware) vendor will have one known good PCR value that “should” be present in the relevant PCRs after the box has booted. In this case, the signed reference measurement could simply list the expected hashes for the given version. However, a RIM that contains the intermediate hashes can be useful in debugging cases where the expected final hash is not the one reported. Measurements taken later in operation of the system, once an OS has started (for example, Linux IMA), may be more complex, with unpredictable “final” PCR values. In this case, the Verifier must have enough information to reconstruct the expected PCR values from logs and signed reference measurements from a trusted authority. In both cases, the expected values can be expressed as signed SWID or CoSWID tags, but the SWID structure in the second case is somewhat more complex, as reconstruction of the extended hash in a PCR may involve thousands of files and other objects. The TCG has published an information model defining elements of reference integrity manifests under the title TCG Reference Integrity Manifest Information Model . This information model outlines how SWID tags should be structured to allow attestation, and defines “bundles” of SWID tags that may be needed to describe a complete software release. The RIM contains some metadata relating to the software release it belongs to, plus hashes for each individual file or other object that could be attested. TCG has also published the PC Client Reference Integrity Measurement specification , which focuses on a SWID-compatible format suitable for expressing expected measurement values in the specific case of a UEFI-compatible BIOS, where the SWID focus on files and file systems is not a direct fit. While the PC Client RIM is not directly applicable to network equipment, many vendors do use a conventional UEFI BIOS to launch their network OS.

Quotes from a TPM can provide evidence of the state of a device up to the time the evidence was recorded, but to make sense of the quote in most cases an event log of what software modules contributed which values to the quote during startup must also be provided. The log must contain enough information to demonstrate its integrity by allowing exact reconstruction of the digest conveyed in the signed quote (e.g., PCR values). TCG has defined several event log formats: UEFI BIOS event log (TCG EFI Platform Specification for TPM Family 1.1 or 1.2, Section 7 ) Canonical Event Log There is also a Legacy BIOS event log, although this document is less relevant as UEFI has largely replaced the Legacy BIOS (TCG PC Client Specific Implementation Specification for Conventional BIOS, Section 11.3) It should be noted that a given device might use more than one event log format (e.g., a UEFI log during initial boot, switching to Canonical Log when the host OS launches). The TCG SNMP Attestation MIB will support any record-oriented log format, including the three TCG-defined formats, but it currently leaves figuring out which log(s) are in what format up to the Verifier.

Initial work at IETF defines remote attestation as follows: The Reference Interaction Model for Challenge-Response-based Remote Attestation is based on the standard roles defined in : Attester: The role that designates the subject of the remote attestation. A system entity that is the provider of evidence takes on the role of an Attester. Verifier: The role that designates the system entity and that is the appraiser of evidence provided by the Attester. A system entity that is the consumer of evidence takes on the role of a Verifier. The following diagram illustrates a common information flow between a Verifier and an Attester, specified in :

assertions | | | signAttestationEvidence(authSecID, assertions, nonce) | | => signedAttestationEvidence | | | | signedAttestationEvidence ----------------------------------> | | | | verifyAttestationEvidence(signedAttestatEvidence, refassertions) | attestationResult <= | | | ]]> The RIV approach outlined in this document aligns with the RATS reference model.

The overall flow for an attestation session is shown in Figure 4. In this diagram: Step 0, obtaining the signed reference measurements, may happen in two ways: Step 0A below shows a verifier obtaining reference measurements directly from a software configuration authority, whether it’s the vendor or another authority chosen by the system administrator. The reference measurements are signed by the Asserter (i.e., the software configuration authority). - Or - Step 0B, the reference measurements, signed by the Asserter, may be distributed as part of software installation, long before the attestation session begins. Software installation is usually vendor-dependent, so there are no standards involved in this step. However, the verifier can use the same protocol to obtain the reference measurements from the device as it would have used with an external reference authority In Step 1, the Verifier initiates an attestation session by opening a TLS session, validated using the DevID to prove that the connection is attesting the right box. In Step 2, measured values are retrieved from the Attester’s TPM using a YANG or SNMP interface that implements the TCG TAP model (e.g. YANG Module for Basic Challenge-Response-based Remote Attestation Procedures ). In Step 3, the Attester also delivers a copy of the signed reference measurements, using Software Inventory YANG module based on Software Identifiers . These steps yield enough information for the Verifier to verify measurements against reference values. Of course, in all cases, the signatures protecting quotes and RIMs must be checked before the contents are used.

| Verifier| | Asserter | | (Device |<------>| (Network| |(Configuration |------->| under | Step 2 | Mngmt | | Authority) | Step 0A| attestation)| | Station)| | | | |------->| | +---------------+ +-------------+ Step 3 +---------+ | /|\ | | ---------------------------------------------- Step 0B ]]> Either CoSWID-encoded reference measurements are signed by a trusted authority and retrieved directly prior to attestation (as shown in Step 0A), or CoSWID-encoded reference measurements are signed by the device manufacturer, installed on the device by a proprietary installer, and delivered during attestation (as shown in Step 0B). In Step 1, the Verifier initiates a connection for attestation. The Attester’s identity is validated using DevID with TLS. In Step 2, a nonce, quotes (measured values) and measurement log are conveyed via TAP with a protocol-specific binding (e.g. SNMP). Logs are sent in the Canonical Log Format In Step 3, CoSWID-encoded reference measurements are retrieved from the Attester using the YANG (. . The following components are used: TPM Keys are configured according to , , or Measurements of firmware and bootable modules are taken according to TCG PC Client and Linux IMA Device Identity is managed by IEEE 802.1AR certificates , with keys protected by TPMs. Attestation logs are formatted according to the Canonical Event Log format Quotes are retrieved according to TCG TAP Information Model . While the TAP IM gives a protocol-independent description of the data elements involved, it’s important to note that quotes from the TPM are signed inside the TPM, so must be retrieved in a way that does not invalidate the signature, as specified in , to preserve the trust model. (See Security Considerations). Reference Integrity Measurements are encoded as CoSWID tags, as defined in the TCG RIM document , compatible with NIST IR 8060 and the IETF CoSWID draft . Reference measurements are signed by the device manufacturer.

Placeholder section for information flow between peer devices such as routers, to establish a known trust relationship. Hint: it’s the same flow, but with devices at either end of a link issuing attestation challenges as to the peer device, and each responding to challenges as a device under attestation. The devices need to carry their own signed reference measurements (RIMs) so that each device has everything needed for attestation, without having to resort to a central authority.

Retrieval of identity and attestation state uses one protocol stack, while retrieval of Reference Measurements uses a different set of protocols. Figure 5 shows the components involved.

| (Management Station) | | | | | | +-----------------------+ | +-------------------------+ | -------------------- -------------------- | | ---------------------------------- --------------------------------- |Reference Integrity Measurements| | Attestation | ---------------------------------- --------------------------------- ******************************************************************** * IETF Attestation Reference Interaction Diagram * ******************************************************************** ....................... ....................... . Reference Integrity . . TAP (PTS2.0) Info . . Manifest . . Model and Canonical . . . . Log Format . ....................... ....................... ************************* .............. ********************** * YANG SWID Module * . TCG . * YANG Attestation * * I-D.birkholz-yang-swid* . Attestation. * Module * * * . MIB . * I-D.ietf-rats- * * * . . * yang-tpm-charra * ************************* .............. ********************** ************************* ************ ************************ * XML, JSON, CBOR (etc) * * UDP * * XML, JSON, CBOR (etc)* ************************* ************ ************************ ************************* ************************ * RESTCONF/NETCONF * * RESTCONF/NETCONF * ************************ ************************* ************************* ************************ * TLS, SSH * * TLS, SSH * ************************* ************************ ]]> IETF documents are captured in boxes surrounded by asterisks. TCG documents are shown in boxes surrounded by dots. The IETF Attestation Reference Interaction Diagram, Reference Integrity Manifest, TAP Information Model and Canonical Log Format, and both YANG modules are works in progress. Information Model layers describe abstract data objects that can be requested, and the corresponding response SNMP is still widely used, but the industry is transitioning to YANG, so in some cases, both will be required. TLS Authentication with TPM has been shown to work; SSH authentication using TPM-protected keys is not as easily done [as of 2019]

Networking Equipment such as routers, switches and firewalls has a key role to play in guarding the privacy of individuals using the network: * Packets passing through the device must not be sent to unauthorized destinations. For example * Routers often act as Policy Enforcement Points, where individual subscribers may be checked for authorization to access a network. Subscriber login information must not be released to unauthorized parties. * Networking Equipment is often called upon to block access to protected resources, or from unauthorized users. * Routing information, such as the identity of a router’s peers, must not be leaked to unauthorized neighbors. * If configured, encryption and decryption of traffic must be carried out reliably, while protecting keys and credentials. Functions that protect privacy are implemented as part of each layer of hardware and software that makes up the networking device. In light of these requirements for protecting the privacy of users of the network, the Network Equipment must identify itself, and its boot configuration and measured device state (for example, PCR values), to the Equipment’s Administrator, so there’s no uncertainty as to what function each device and configuration is configured to carry out . This allows the administrator to ensure that the network provides individual and peer privacy guarantees. RIV specifically addresses the collection information from enterprise network devices by an enterprise network. As such, privacy is a fundamental concern for those deploying this solution, given EU GDPR, California CCPA, and many other privacy regulations. The enterprise should implement and enforce their duty of care. See for more context on privacy in networking devices

Attestation results from the RIV procedure are subject to a number of attacks: Keys may be compromised A counterfeit device may attempt to impersonate (spoof) a known authentic device Man-in-the-middle attacks may be used by a counterfeit device to attempt to deliver responses that originate in an actual authentic device Replay attacks may be attempted by a compromised device Trustworthiness of RIV attestation depends strongly on the validity of keys used for identity and attestation reports. RIV takes full advantage of TPM capabilities to ensure that results can be trusted. Two sets of keys are relevant to RIV attestation A DevID key is used to certify the identity of the device in which the TPM is installed. An Attestation Key (AK) key signs attestation reports, (called ‘quotes’ in TCG documents), used to provide evidence for integrity of the software on the device. TPM practices require that these keys be different, as a way of ensuring that a general-purpose signing key cannot be used to spoof an attestation quote. In each case, the private half of the key is known only to the TPM, and cannot be retrieved externally, even by a trusted party. To ensure that’s the case, specification-compliant private/public key-pairs are generated inside the TPM, where they’re never exposed, and cannot be extracted (See ). Keeping keys safe is just part of attestation security; knowing which keys are bound to the device in question is just as important. While there are many ways to manage keys in a TPM (See ), RIV includes support for “zero touch” provisioning (also known as zero-touch onboarding) of fielded devices (e.g. Secure ZTP, }), where keys which have predictable trust properties are provisioned by the device vendor. Device identity in RIV is based on IEEE 802.1AR DevID. This specification provides several elements A DevID requires a unique key pair for each device, accompanied by an x.509 certificate The private portion of the DevID key is to be stored in the device, in a manner that provides confidentiality (Section 6.2.5 ) The x.509 certificate contains several components The public part of the unique DevID key assigned to that device An identifying string that’s unique to the manufacturer of the device. This is normally the serial number of the unit, which might also be printed on label on the device. The certificate must be signed by a key traceable to the manufacturer’s root key. With these elements, the device’s manufacturer and serial number can be identified by analyzing the DevID certificate plus the chain of intermediate certs leading back to the manufacturer’s root certificate. As is conventional in TLS connections, a nonce must be signed by the device in response to a challenge, proving possession of its DevID private key. RIV uses the DevID to validate a TLS connection to the device as the attestation session begins. Security of this process derives from TLS security, with the DevID providing proof that the TLS session terminates on the intended device. . Evidence of software integrity is delivered in the form of a quote signed by the TPM itself. Because the contents of the quote are signed inside the TPM, any external modification (including reformatting to a different data format) will be detected as tampering. A critical feature of the YANG model described in is the ability to carry TPM data structures in their native format, without requiring any changes to the structures as they were signed and delivered by the TPM. While alternate methods of conveying TPM quotes could add an additional layer of signing using external keys, the important part is to preserve the TPM signing, so that tampering anywhere in the path between the TPM itself and the Verifier can be detected. Prevention of spoofing attacks against attestation systems is also important. There are two cases to consider: * The entire device could be spoofed, that is, when the Verifier goes to verify a specific device, it might be redirected to a different device. Use of the 802.1AR identity in the TPM ensures that the Verifier’s TLS session is in fact terminating on the right device. * A compromised device could respond with a spoofed attestation result, that is, a compromised OS could return a fabricated quote. Protection against spoofed quotes from a device with valid identity is a bit more complex. An identity key must be available to sign any kind of nonce or hash offered by the verifier, and consequently, could be used to sign a fabricated quote. To block spoofed attestation result, the quote generated inside the TPM must by signed by a key that’s different from the DevID, called an Attestation Key (AK). Given separate Attestation and DevID keys, the binding between the AK and the same device must also be proven to prevent a man-in-the-middle attack (e.g. the ‘Asokan Attack’ ). This is accomplished in RIV through use of an AK certificate with the same elements as the DevID (i.e., same manufacturer’s serial number, signed by the same manufacturer’s key), but containing the device’s unique AK public key instead of the DevID public key. [this will require an OID that says the key is known by the CA to be an Attestation key] These two keys and certificates are used together: The DevID is used to validate a TLS connection terminating on the device with a known serial number. The AK is used to sign attestation quotes, providing proof that the attestation evidence comes from the same device. Replay attacks, where results of a previous attestation are submitted in response to subsequent requests, are usually prevented by inclusion of a nonce in the request to the TPM for a quote. Each request from the Verifier includes a new random number (a nonce). The resulting quote signed by the TPM contains the same nonce, allowing the verifier to determine freshness, i.e., that the resulting quote was generated in response to the verifier’s specific request. Time-Based Uni-directional Attestation provides an alternate mechanism to verify freshness without requiring a request/response cycle. Requiring results of attestation of the operating software to be signed by a key known only to the TPM also removes the need to trust the device’s operating software (beyond the first measurement; see below); any changes to the quote, generated and signed by the TPM itself, made by malicious device software, or in the path back to the verifier, will invalidate the signature on the quote. Although RIV recommends that device manufacturers pre-provision devices with easily-verified DevID and AK certs, use of those credentials is not mandatory. IEEE 802.1AR incorporates the idea of an Initial Device ID (IDevID), provisioned by the manufacturer, and a Local Device ID (LDevID) provisioned by the owner of the device. RIV extends that concept by defining an Initial Attestation Key (IAK) and Local Attestation Key (LAK) with the same properties. Device owners can use any method to provision the Local credentials. TCG doc shows how the initial Attestation keys can be used to certify LDevID and LAK keys. Use of the LDevID and LAK allows the device owner to use a uniform identity structure across device types from multiple manufacturers (in the same way that an “Asset Tag” is used by many enterprises use to identify devices they own). TCG doc also contains guidance on provisioning identity keys in TPM 2.0. But device owners can use any other mechanism they want to assure themselves that Local identity certificates are inserted into the intended device, including physical inspection and programming in a secure location, if they prefer to avoid placing trust in the manufacturer-provided keys. Clearly, Local keys can’t be used for secure Zero Touch provisioning; installation of the Local keys can only be done by some process that runs before the device is configured for network operation. On the other end of the device life cycle, provision should be made to wipe Local keys when a device is decommissioned, to indicate that the device is no longer owned by the enterprise. The manufacturer’s Initial identity keys must be preserved, as they contain no information that’s not already printed on the device’s serial number plate. In addition to trustworthy provisioning of keys, RIV depends on other trust anchors. (See for definitions of Roots of Trust.) Secure identity depends on mechanisms to prevent per-device secret keys from being compromised. The TPM provides this capability as a Root of Trust for Storage Attestation depends on an unbroken chain of measurements, starting from the very first measurement. That first measurement is made by code called the Root of Trust for Measurement, typically done by trusted firmware stored in boot flash. Mechanisms for maintaining the trustworthiness of the RTM are out of scope for RIV, but could include immutable firmware, signed updates, or a vendor-specific hardware verification technique. RIV assumes some level of physical defense for the device. If a TPM that has already been programmed with an authentic DevID is stolen and inserted into a counterfeit device, attestation of that counterfeit device may become indistinguishable from an authentic device. RIV also depends on reliable reference measurements, as expressed by the RIM . The definition of trust procedures for RIMs is out of scope for RIV, and the device owner is free to use any policy to validate a set of reference measurements. RIMs may be conveyed out-of-band or in-band, as part of the attestation process (see Section 3.2). But for embedded devices, where software is usually shipped as a self-contained package, RIMs signed by the manufacturer and delivered in-band may be more convenient for the device owner.

TCG technologies can play an important part in the implementation of Remote Integrity Verification. Standards for many of the components needed for implementation of RIV already exist: Platform identity can be based on IEEE 802.1AR Device identity, coupled with careful supply-chain management by the manufacturer. Complex supply chains can be certified using TCG Platform Certificates The TCG TAP mechanism can be used to retrieve attestation evidence. Work is needed on a YANG model for this protocol. Reference Measurements must be conveyed from the software authority (e.g., the manufacturer) to the system in which verification will take place. IETF CoSWID work forms the basis for this, but new work is needed to create an information model and YANG implementation.

There have been demonstrations of attestation using TPMs for years, accompanied by compelling security reasons for adopting attestation. Despite this, the technology has not been widely adopted, in part, due to the difficulties in deploying TPM-based attestation. Some of those difficulties are: Standardizing device identity. Creating and using unique device identifiers is difficult, especially in a privacy-sensitive environment. But attestation is of limited value if the operator is unable to determine which devices pass attestation validation tests, and which fail. This problem is substantially simplified for infrastructure devices like network equipment, where identity can be explicitly coded using IEEE 802.1AR, but doing so relies on adoption of 802.1AR by manufacturers and hardware system integrators. Standardizing attestation representations and conveyance. Interoperable remote attestation has a fundamental dependence on vendors agreeing to a limited set of network protocols commonly used in existing network equipment for communicating attestation data. Network device vendors will be slow to adopt the changes necessary to implement remote attestation without a fully-realized plan for deployment. Interoperability. Networking equipment operates in a fundamentally multi-vendor environment, putting additional emphasis on the need for standardized procedures and protocols. Attestation evidence is complex. Operating systems used in larger embedded devices are often multi-threaded, so the order of completion for individual processes is non-deterministic. While the hash of a specific component is stable, once extended into a PCR, the resulting values are dependent on the (non-deterministic) ordering of events, so there will never be a single known-good value for some PCRs. Careful analysis of event logs can provide proof that the expected modules loaded, but it’s much more complicated than simply comparing reported and expected digests, as collected in TPM Platform Configuration Registers (PCRs). Software configurations can have seemingly infinite variability. This problem is nearly intractable on PC and Server equipment, where end users have unending needs for customization and new applications. However, embedded systems, like networking equipment, are often simpler, in that there are fewer variations and releases, with vendors typically offering fewer options for mixing and matching. Standards-based mechanisms to encode and distribute Reference Integrity Measurements, (i.e., expected values) are still in development. Software updates can be complex. Even the most organized network operator may have many different releases in their network at any given time, with the result that there’s never a single digest or fingerprint that indicates the software is “correct”; digests formed by hashing software modules on a device can only show the correct combination of versions for a specific device at a specific time. None of these issues are insurmountable, but together, they’ve made deployment of attestation a major challenge. The intent of this document is to outline an attestation profile that’s simple enough to deploy, while yielding enough security to be useful.

Even in embedded systems, adding Attestation at the OS level (e.g. Linux IMA, Integrity Measurement Architecture ) increases the number of objects to be attested by one or two orders of magnitude, involves software that’s updated and changed frequently, and introduces processes that begin in unpredictable order. TCG and others (including the Linux community) are working on methods and procedures for attesting the operating system and application software, but standardization is still in process.

Table 1 summarizes many of the actions needed to complete an Attestation system, with links to relevant documents. While documents are controlled by several standards organizations, the implied actions required for implementation are all the responsibility of the manufacturer of the device, unless otherwise noted.

Some components of an Attestation system have been implemented for end-user machines such as PCs and laptops. Figure 7 shows the corresponding protocol stacks.

This memo includes no request to IANA.