Requirements for Resource Public Key Infrastructure (RPKI) Relying Parties

The RPKI Relying Party (RP) software is used by network operators and others to acquire and verify Internet Number Resource (INR) data stored in the RPKI repository system. RPKI data, when verified, allow an RP to verify assertions about which Autonomous Systems (ASes) are authorized to originate routes for IP address prefixes. RPKI data also establishes binding between public keys and BGP routers, and indicates the AS numbers that each router is authorized to represent. Noting that the essential requirements imposed on RPs to support securing Internet routing () are scattered throughout numerous RFC documents that are protocol specific or provide best practices, as follows: RFC 6481 (Repository Structure) RFC 6482 (ROA format) RFC 6486 (Manifests) RFC 6487 (Certificate and CRL profile) RFC 6488 (RPKI Signed Objects) RFC 6489 (Key Rollover) RFC 6810 (RPKI to Router Protocol) RFC 6916 (Algorithm Agility) RFC 7935 (Algorithms) RFC 8209 (Router Certificates) RFC 8210 (RPKI to Router Protocol,Version 1) RFC 8360 (Certificate Validation Procedure) I-D.ietf-sidrops-https-tal (Trust Anchor Locator) This makes it hard for an implementer to be confident that he/she has addressed all of these generalized requirements. Besides, software engineering calls for how to segment the RP system into components with orthogonal functionalities, so that those components could be distributed across the operational timeline of the user. Taxonomy of generalized RP requirements is going to help have ‘the role of the RP’ well framed. To consolidate RP requirements in one document, with pointers to all the relevant RFCs, this document outlines a set of baseline requirements imposed on RPs and provides a single reference point for requirements for RP software for use in the RPKI, as segmented with orthogonal functionalities: • Fetching and Caching RPKI Repository Objects • Processing Certificates and CRLs • Processing RPKI Repository Signed Objects • Distributing Validated Cache of the RPKI Data This document will be update to reflect new or changed requirements as these RFCs are updated, or new RFCs are written.

RP software uses synchronization mechanisms supported by targeted repositories (e.g., , RRDP ) to download all RPKI changed data objects in the repository system and cache them locally. The software validates the RPKI data and uses it to generate authenticated data identifying which ASes are authorized to originate routes for address prefixes, and which routers are authorized to sign BGP updates on behalf of ASes.

In the RPKI, each RP chooses its own set of trust anchors (TAs). Consistent with the extant INR allocation hierarchy, the IANA and/or the five RIRs are obvious candidates to be default TAs for the RP. An RP does not retrieve TAs directly. A set of Trust Anchor Locators (TALs) is used by each RP to retrieve and verify the authenticity of each TA. TAL acquisition and processing are specified in Section 3 of .

The RPKI repository system is a distributed one, consisting of multiple repository instances. Each repository instance contains one or more repository publication points. An RP discovers publication points using the Subject Information Access (SIA) and the Authority Information Access (AIA) extensions from (validated) certificates. Section 5 of specifies how an RP locates all RPKI objects by using the SIA and AIA extensions. Detailed specifications of SIA and AIA extensions in a resource certificate are described in Section 4 of .

An RP takes the key rollover period into account with regard to its frequency of synchronization with RPKI repository system. RP requirements in dealing with key rollover are described in Section 3 of and Section 3 of .

The set of cryptographic algorithms used with the RPKI is expected to change over time. Each RP is expected to be aware of the milestones established for the algorithm transition and what actions are required at every juncture. RP requirements for dealing with algorithm transition are specified in Section 4 of .

Each RP is expected to maintain a local cache of RPKI objects. The cache needs to be as up to date and consistent with repository publication point data as the RP’s frequency of checking permits. The last paragraph of Section 5 of provides guidance for maintenance of a local cache.

The RPKI make use of X.509 certificates and CRLs, but it profiles these standard formats . The major change to the profile established in is the mandatory use of a new extension to X.509 certificate .

Certificates in the RPKI are called resource certificates, and they are required to conform to the profile . An RP is required to verify that a resource certificate adheres to the profile established by Section 4 of . This means that all extensions mandated by Section 4.8 of must be present and value of each extension must be within the range specified by this RFC. Moreover, any extension excluded by Section 4.8 of must be omitted. Section 7.1 of gives the procedure that the RP should follow to verify resource certificate and syntax.

The INRs in issuer’s certificate are required to encompass the INRs in the subject’s certificate. This is one of necessary principles of certificate path validation in addition to cryptographic verification i.e., verification of the signature on each certificate using the public key of the parent certificate). Section 7.2 of gives the procedure that the RP should follow to perform certificate path validation. Certificate Authorities that want to reduce aspects of operational fragility will migrate to the new OIDs , informing the RP of using an alternative RPKI validation algorithm. An RP is expected to support the amended procedure to handle with accidental over-claim.

The CRL processing requirements imposed on CAs and RP are described in Section 5 of . CRLs in the RPKI are tightly constrained; only the AuthorityKeyIndetifier and CRLNumber extensions are allowed, and they are required to be present. No other CRL extensions are allowed, and no CRLEntry extensions are permitted. RPs are required to verify that these constraints have been met. Each CRL in the RPKI must be verified using the public key from the certificate of the CA that issued the CRL. In the RPKI, RPs are expected to pay extra attention when dealing with a CRL that is not consistent with the Manifest associated with the publication point associated with the CRL. Processing of a CRL that is not consistent with a manifest is a matter of local policy, as described in the fourth paragraph of Section 6.6 of .

Before an RP can use a signed object from the RPKI repository, the RP is required to check the signed object syntax. Section 3 of lists all the steps that the RP is required to execute in order to validate the top level syntax of a repository signed object. Note that these checks are necessary, but not sufficient. Additional validation checks must be performed based on the specific type of signed object.

To determine whether a manifest is valid, the RP is required to perform manifest-specific checks in addition to those specified in . Specific checks for a Manifest are described in Section 4 of . If any of these checks fails, indicating that the manifest is invalid, then the manifest will be discarded and treated as though no manifest were present.

To validate a ROA, the RP is required perform all the checks specified in as well as the additional ROA-specific validation steps. The IP address delegation extension present in the end-entity (EE) certificate (contained within the ROA), must encompass each of the IP address prefix(es) in the ROA. More details for ROA validation are specified in Section 4 of .

The Ghostbusters Record is optional; a publication point in the RPKI can have zero or more associated Ghostbuster Records. If a CA has at least one Ghostbuster Record, RP is required to verify that this Ghostbusters Record conforms to the syntax of signed object defined in . The payload of this signed object is a (severely) profiled vCard. An RP is required to verify that the payload of Ghostbusters conforms to format as profiled in .

A BGPsec Router Certificate is a resource certificate, so it is required to comply with . Additionally, the certificate must contain an AS Identifier Delegation extension, and must not contain an IP Address Delegation extension. The validation procedure used for BGPsec Router Certificates is identical to the validation procedure described in Section 7 of , but using the constraints applied come from specification of Section 7 of . Note that the cryptographic algorithms used by BGPsec routers are found in . Currently, the algorithms specified in and are different. BGPsec RPs will need to support algorithms that are used to validate BGPsec signatures as well as the algorithms that are needed to validate signatures on BGPsec certificates, RPKI CA certificates, and RPKI CRLs.

For a given publication point, the RP ought to perform tests, as specified in Section 6.1 of , to determine the state of the Manifest at the publication point. A Manifest can be classified as either valid or invalid, and a valid Manifest is either current and stale. An RP decides how to make use of a Manifest based on its state, according to local (RP) policy. If there are valid objects in a publication point that are not present on a Manifest, does not mandate specific RP behavior with respect to such objects. However, most RP software ignores such objects and the authors of this document suggest this behavior be adopted uniformly. In the absence of a Manifest, an RP is expected to accept all valid signed objects present in the publication point. If a Manifest is stale or invalid (see ) and an RP has no way to acquire a more recently valid Manifest, the RP is expected to contact the repository manager via Ghostbusters record and thereafter make decision according to local (RP) policy.

An RP may encounter a stale Manifest or CRL, or an expired CA certificate or ROA at a publication point. An RP is expected to use the information from the Ghostbusters record to contact the maintainer of the publication point where any stale/expired objects were encountered. The intent here is to encourage the relevant CA and/or repository manager to update the slate or expired objects.

On a periodic basis, BGP speakers within an AS request updated validated origin AS data and router/ASN data from the validated cache of RPKI data. The RP may either transfer the validated data to the BGP speakers directly, or it may transfer the validated data to a cache server that is responsible for provisioning such data to BGP speakers. The specification of the protocol designed to deliver validated cache data to a BGP Speaker is provided in and .

The RP links the RPKI provisioning side and the routing system, establishing the local view of global RPKI data to BGP speakers. The security of the RP is critical to BGP messages exchanging. The RP implementation is expected to offer cache backup management to facilitate recovery from outage state. The RP implementation also should support application of secure transport (e.g., IPsec ) that is able to protect validated cache delivery in a unsafe environment.

This document has no actions for IANA.

The authors thank David Mandelberg, Wei Wang, Tim Bruijnzeels, George Michaelson and Oleg Muravskiy for their review, feedback and editorial assistance in preparing this document.