Attestation of File Content using an X.509 Certificate

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 5 November 2022.¶

Copyright Notice

Copyright (c) 2022 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

Linear hashing is a common technique for protecting the integrity of data. A fixed-size hash, or digest, is computed over the bytes in a data set using a deterministic and collision-resistant algorithm. An example of such an algorithm is [FIPS.180-4].¶

Filesystem designers often employ this technique to protect the integrity of both individual files and filesystem metadata. For instance, to protect an individual file's integrity, the filesystem computes a digest from the beginning of its content to its end. The filesystem then stores that digest along with the file content. The integrity of that digest can be further protected by cryptographically signing it. The filesystem recomputes the digest when the file is retrieved and compares the locally-computed digest with the saved digest to verify the file content.¶

Over time, linear hashing has proven to be an inadequate fit with the way filesystems manage file content. A content verifier must read the entire file to validate its digest. Reading whole files is not onerous for small files, but reading a large file every time its digest needs verification quickly becomes costly.¶

Filesystems read files from persistent storage in small pieces (blocks) on demand to manage large files efficiently. When memory is short, the system evicts these data blocks and then reads them again when needed later. There is no physical guarantee that a subsequent read of a particular block will give the same result as an earlier one. Thus the initial verification of a file's becomes stale, sometimes quickly.¶

To address this shortcoming, some have turned to hash trees [Merkle88]. A hash tree leaf node contains the linear hash of a portion of the protected content. Interior nodes in a hash tree contain hashes of the nodes below them, up to the root node which stores a hash of everything in the tree. Validating a leaf node means validating only the portion of the file content protected by that node and its parents in the hash tree.¶

Hash trees present a new challenge, however. Even when signed, a single linear hash is the same size no matter how much content it protects. The size of a hash tree, however, increases logarithmically with the size of the content it protects.¶

Transporting and storing a hash tree can therefore be unwieldy. It is particularly a problem for legacy storage formats that do not have mechanisms to handle extensive amounts of variably-sized metadata. Software distribution and packaging formats might not be flexible enough to transport this possibly large amount of integrity data. Backup mechanisms such as tar or rsync might be unable to handle variably-sized metadata per file.¶

Moreover, we can readily extend network file storage protocols to exchange a hash tree associated with every file. However, to support such extensions, file servers and the ecosystems where they run must be updated to manage and store this metadata. Thus it is not merely an issue of enriching a file storage protocol to handle a new metadata type.¶

2. Requirements Language

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. Hash Tree Metadata

Reconstituting a hash tree (as opposed to building a more generic directed graph of hashes) requires the protected content, a basic set of metadata, and an understanding of how to use the metadata to reconstitute the hash tree:¶

• The algorithm used to compute the tree's digests¶
• The divergence factor (defined as one for a hash list and two for binary hash trees)¶
• The tree height (from root to the lowest leaf node)¶
• The block size covered by each leaf node in the tree¶
• An optional salt value¶

More research might be needed to cover recent innovations in hash tree construction; in particular, the use of prefixes to prevent second pre-image attacks.¶

The digest algorithm used to construct the hash tree MUST match the digest algorithm used to sign the certificate. Thus if SHA-2 is used to construct the hash tree, the certificate signature is created with SHA-2. The verifier then uses SHA-2 when validating the certificate signature and reconstituting the hash tree. The object identifiers for the supported algorithms and the methods for encoding public key materials (public key and parameters) are specified in [RFC3279], [RFC4055], and [RFC4491].¶

The block size value of the tree is specified in octets. For example, if the block size is 4096, then each leaf node of the hash tree digests 4096 octets of the protected file (aligned on 4096-octet boundaries).¶

The internal nodes are digests constructed from the hashes of two adjacent child nodes up to the root node (further detail needed here). The tree's height is the distance, counted in nodes, from the root to the lowest leaf node.¶

The leaf nodes are ordered (left to right) by the file offset of the block they protect. Thus, the left-most leaf node represents the first block in the file, and the right-most leaf node represents the final block in the file.¶

An explanation of the salt value goes here.¶

Further, when computing each digest, an extra byte might be prefixed to the pre-digested content to reduce the possibility of a second-preimage attack.¶

4. File Provenance Certificates

X.509 certificates are specified in [X.509]. The current document extends the Internet X.509 certificate profile specified in [RFC5280] to represent file content protected by hash tree metadata.¶

File provenance certificates are end-entity certificates. The certificate's signature identifies the attestor and cryptographically protects the hash tree metadata.¶

The Subject field MUST be an empty sequence. The SubjectAltName list carries a filename and the root hash, encoded in a new otherName type-ID, shown below. The current document requests allocation of this new type-ID on the id-on arc, defined in [RFC7299]. The following subsections describe how the fields in this new type-ID are used.¶

  id-pkix OBJECT IDENTIFIER ::=
             { iso(1) identified-organization(3) dod(6) internet(1)
                        security(5) mechanisms(5) pkix(7) }
  id-on OBJECT IDENTIFIER ::= { id-pkix 8 }
  id-on-fileContentAttestation OBJECT IDENTIFIER ::= { id-on II }
  FileContentAttestation ::= SEQUENCE {
     treeRootDigest OCTET STRING,
     treeDivergenceFactor INTEGER  (1..2),
     treeHeight INTEGER,
     treeBlockSize INTEGER,
     treeSaltValue OCTET STRING
  }

5. Implementation Status

This section records the status of known implementations of the protocol defined by this specification at the time of posting of this Internet-Draft, and is based on a proposal described in [RFC7942]. The description of implementations in this section is intended to assist the IETF in its decision processes in progressing drafts to RFCs.¶

Please note that the listing of any individual implementation here does not imply endorsement by the IETF. Furthermore, no effort has been spent to verify the information presented here that was supplied by IETF contributors. This is not intended as, and must not be construed to be, a catalog of available implementations or their features. Readers are advised to note that other implementations may exist.¶

There are no known implementations of the X.509 certificate extensions described in the current document.¶

6. Security Considerations

It is important to note the narrow meaning of the digital signature in X.509 certificates as defined in this document. That signature connotes that the data content of the certificate has not changed since the certificate was signed, and it identifies the signer cryptographically. The signature does not confer any meaning or guarantees other than the integrity of the certificate's data content.¶

7. IANA Considerations

  id-pkix OBJECT IDENTIFIER ::=
             { iso(1) identified-organization(3) dod(6) internet(1)
                        security(5) mechanisms(5) pkix(7) }
  id-on OBJECT IDENTIFIER ::= { id-pkix 8 }
  id-on-fileContentAttestation OBJECT IDENTIFIER ::= { id-on II }

Acknowledgments

The editor is grateful to Bill Baker, Eric Biggers, James Bottomley, Russ Housley, Benjamin Kaduk, Rick Macklem, Greg Marsden, Paul Moore, Martin Thomson, and Mimi Zohar for their input and support.¶

Finally, special thanks to Transport Area Directors Martin Duke and Zaheduzzaman Sarker, NFSV4 Working Group Chairs David Noveck and Brian Pawlowski, and NFSV4 Working Group Secretary Thomas Haynes for their guidance and oversight.¶

Author's Address