Scalable Remote Attestation for Systems, Containers, and Applications

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶

This Internet-Draft will expire on 29 September 2024.¶

Copyright Notice

Copyright (c) 2024 IETF Trust and the persons identified as the document authors. All rights reserved.¶

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶

Table of Contents

1. Introduction

Posture assessment has long been desired, but has been difficult to achieve due to complexities of customization requirements at each organization. By using policy and measurement sets that may be offered at various assurance levels defined in an Entity Attestation Token (EAT) profile [I-D.ietf-rats-eat], automating posture assessment through attestation becomes achievable for organizations of all sizes. The measurement and policy groupings in an EAT profile may be provided by the vendor or by a neutral third party to enable ease of use and consistent implementations. This provides simpler options to enable posture assessment at selected levels by organizations without the need to have in-house expertise. The measurement and policy sets may also be customized, but not necessary to achieve posture assessment to predefined options. This document describes a method to use existing attestation formats and protocols while allowing for defined profiles of policies, benchmarks, and measurements for specific assurance levels that scale to provide transparency to posture assessment results summarized with remote attestation.¶

By way of example, the Center for Internet Security (CIS) hosts recommended configuration settings to secure operating systems, applications, and devices in CIS Benchmarks developed with industry experts. Attestations aligned to the CIS Benchmarks or other configuration guide such as a DISA STIG could be used to assert the configuration meets expectations. This has already been done for multiple platforms to demonstrate assurance for firmware according to NIST SP 800-193, Firmware Resiliency Guidelines [FIRMWARE]. In order to scale remote attestation, a single attestation for a set of benchmarks or policies being met may be sent to the remote attestation management system. On traditional servers, assurance to NIST SP 800-193 is provable through attestation from a root of trust (RoT), using the Trusted Computing Group (TCG) Trusted Platform Module (TPM) chip and attestation formats.¶

At boot, policy and measurement expectations are verified against a set of "golden policies" from collected and attested evidence. Device identity and measurements can also be attested at runtime. The attestations on evidence (e.g. hash of boot element) and verification of attestations are typically contained within a system and are limited to the control plane for management. The policy and measurement sets for comparison are protected to assure the result in the attestation verification process for boot element. Event logs and PCR values may be exposed to provide transparency into the verified attestations. Remote attestation provides a summary of a local assessment of posture for managed systems and across various layers (operating system, application, containers) in each of these systems in a managed environment.¶

There is a balance of exposure and evidence needed to assess posture when providing assurance of controls and system state. Currently, logs and TPM PCR values may be passed to provide assurance of verification of attestation evidence meeting set requirements. Providing the assurance can be accomplished with a remote attestation format such as the Entity Attestation Token (EAT) [I-D.ietf-rats-eat] and a RESTful interface such as ROLIE [RFC8832] or RedFish [REDFISH]. Policy definition blocks may be scoped to control measurement sets, where the EAT profile asserts compliance to the policy or measurement block specified and may include claims with the log and PCR value evidence. Measurement and Policy sets, referenced in an EAT profile may be published and maintained by separate entities (e.g. CIS Benchmarks, DISA STIGs). The policy and measurement sets should be maintained separately even if associated with the same benchmark or control set. This avoids the need to transition the verifying entity to a remote system for individual policy and measurements which are performed locally for more immediate remediation as well as other functions.¶

Examples of measurement and policy sets that could be defined in EAT profiles include, but are not limited to:¶

• Hardware attribute certificates, TCG¶

• Hardware Attribute Certificate Comparison Results, TCG¶

• Reference Integrity Measurements for firmware, TCG¶

• Operating system benchmarks at Specified Assurance Levels, CIS¶

• Application hardening Benchmarks at Specified Assurance Levels, CIS, DISA STIG¶

• Container security benchmarks at Specified Assurance Levels, CIS¶

Scale, ease of use, full automation, and consistency for customer consumption of a remote attestation function or service are essential toward the goal of consistently securing systems against known threats and vulnerabilities. Mitigations may be baked into policy. Claim sets of measurements and policy verified to meet or not meet Endorsed values [I-D.ietf-rats-eat] are conveyed in an Entity Attestation Token made available to a RESTful interface in aggregate for the systems managed. The Measurement or Policy Set may be registered in the IANA registry created in this document (Section 8), detailing the specific configuration policies and measurements required to adhere or prove compliance to the associated document. Levels (e.g. high, medium, low, 1, 2, 3) or vendor specific instances of the policy defined in code required to verify the policy and measurements would be registered using a name for the policy set, that would also be used in the reporting EAT that includes the MPS along with other artifacts to prove compliance.¶

2. Conventions and Definitions

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶

3. Policy and Measurement Set Definitions

This document defines EAT claims in the JWT [RFC7519] and CWT [RFC8392] registries to provide attestation to a set of verified claims within a defined grouping. The trustworthiness will be conveyed on original verified evidence as well as the attestation on the grouping. The claims provide the additional information needed for an EAT to convey compliance to a defined policy or measurement set to a system or application collecting evidence on policy and measurement assurance, for instance a governance, risk, and compliance (GRC) system.¶

4. Supportability and Re-Attestation

The remote attestation framework shall include provisions within the system and attestation authority to allow for Product modification.¶

Over its lifecycle, the Product may experience modification due to: maintenance, failures, upgrades, expansion, moves, etc..¶

The customer can chose to:¶

• Run remote attestation after product modification, or¶

• Not take action and remain un-protected¶

In the case of Re-Attestation:¶

• framework needs to invalidate previous TPM PCR values and tokens,¶

• framework needs to collect new measurements,¶

• framework needs to maintain history or allow for history to be logged to enable change traceability attestation, and¶

• framework needs to notify that the previous attestation has been invalidated¶

5. Configuration Sets

In some cases, it may be difficult to attest to configuration settings for the initial or subsequent attestation and verification processes. The use of an expected hash value for configuration settings can be used to compare the attested configuration set. In this case, the creator of the attestation verification measurements would define a set of values for which a message digest would be created and then signed by the attestor. The expected measurements would include the expected hash value for comparison. The configuration set could be the full attestation set to a Benchmark or a defined subset. These configuration sets can be registered for general use to reduce the need to replicate the policy and measurement assessments by others aiming to assure at the same level for a benchmark or hardening guide. This document creates an IANA registry for this purpose, creating consistency between automated policy and measurement set levels and the systems used to collect and report aggregate views for an organization across systems and applications, such as a GRC platform.¶

6. Remediation

If policy and configuration settings or measurements attested do not meet expected values, remediation is desireable. Automated remediation performed with alignment to zero trust architecture principles would require that the remediation be performed prior to any relying component executing. The relying component would verify before continuing in a zero trust architecture.¶

Ideally, remediation would occur on system as part of the process to attest to a set of attestations, similar to how attestation is performed for firmware in the boot process. If automated remediation is not possible, an alert should be generated to allow for notification of the variance from expected values.¶

7. Security Considerations

This document establishes a pattern to list sets of benchmarks and controls within CWT and JWT formats. The contents of the benchmarks and controls are out of scope for this document. This establishes an architectural pattern whereby a remote attestation could be issued for a complete set of benchmarks or controls as defined and grouped by external entities, preventing the need to send over individual attestations for each item within a benchmark or control framework. This document does not add security consideration over what has been described in the EAT, JWT, or CWT specifications.¶

8. IANA Considerations

Draft section - authors know more work is needed to properly define the registry and claims. This section is here now to assist in understanding the concepts.¶

This document requests the creation of a Measurement and Policy Set (MPS) registry. The MPS registry will contain the names of the Benchmarks, Policy sets, DISA STIGS, controls, or other groupings as a policy and measurement set that MAY correlate to standards documents containing assurance guidelines, compliance requirements, or other defined claim sets for verification of posture assessment to that MPS. The MPS registry will include the policy definition for specific levels of MPS assurance to enable interoperability between assertions of compliance (or lack thereof) and reporting systems.¶

The MPS name includes versions or level information, allowing for distinct policy or measurement sets and definitions of those sets (including the supporting formats used to write the definitions).¶

Contributors

Thank you to reviewers and contributors who helped to improve this document. Thank you to Nick Grobelney, Dell Technologies, for your review and contribution to separate out the policy and measurement sets. Thank you, Samant Kakarla and Huijun Xie from Dell Technologies, for your detailed review and corrections on boot process details. Section 3 has been contributed by Rudy Bauer from Dell as well and an author will be added on the next revision. IANA section added in version 7 by Kathleen Moriarty, expanding the claims registered and adding a proposed registry to define policy and measurement sets.¶

Authors' Addresses