Internet-Draft | Selective Synchronization for RTR | March 2024 |
Geng, et al. | Expires 4 September 2024 | [Page] |
The RPKI-to-Router (RTR) protocol synchronizes all the verified RPKI data to routers. This document proposes to extend the existing RTR protocol to support selective data synchronization. Selective synchronization can avoid some unnecessary synchronizations. The router can obtain only the data that it really needs, and it does not need to save the data that are not needed.¶
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 4 September 2024.¶
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.¶
The RPKI-to-Router (RTR) protocol is a simple but reliable approach, which help synchronize the validated RPKI data from a trusted cache to routers. There are already several versions of the protocol [RFC6810][RFC8210][I-D.ietf-sidrops-8210bis]. The supported types of data that can be transferred increase, which is shown in Table 1.¶
Version 0 | Version 1 | Version 2 |
---|---|---|
IPv4 Prefix | IPv4 Prefix | IPv4 Prefix |
IPv6 Prefix | IPv6 Prefix | IPv6 Prefix |
Router Key | Router Key | |
ASPA |
The RTR protocol keeps the synchronization of all types of data, and selective synchronization is not supported. However, routers may be interested in a part of data types, instead of all. In such cases, storing unused data on the router is unreasonable, and synchronizing all types of data will induce some unnecessary transmission and storage overhead. Since multiple types of data are transmitted together, the router cannot use any type of these data unless it waits for all data to complete transmission. Furthermore, there may be more types of data in the cache, , which makes the above issue more significant and worse. The followings are example types, and some of them may be possibly supported in the RTR protocol in the future:¶
Secured Routing Policy Specification Language (RPSL) [RFC7909]¶
Signed Prefix Lists [I-D.ietf-sidrops-rpki-prefixlist]¶
Autonomous Systems Cones [I-D.ietf-grow-rpki-as-cones]¶
Mapping Origin Authorizations (MOAs) [I-D.xie-sidrops-moa-profile]¶
Signed SAVNET-Peering Information (SiSPI) [I-D.chen-sidrops-sispi]¶
Path validation with RPKI [I-D.van-beijnum-sidrops-pathrpki]¶
Signed Groupings of Autonomous System Numbers [I-D.spaghetti-sidrops-rpki-asgroup]¶
This document describes the synchronization problem of the RTR protocol and provides some possible solutions.¶
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.¶
The RTR protocol does not distinguish data types in the cache. Different types of data share one serial number and one End of Data PDU. When the Relay Party (RP) synchronizes the cache to the router, various PDUs, such as IPv4 Prefix, IPv6 Prefix, Router Key, and ASPA, are mixed. The router cannot select one or more really required PDUs or deny receiving a certain kind of PDU. For example, if the router supports RTR v2 but does not support or enable ASPA, the ASPA PDU messages will still be transmitted. Another example is the router in an IPv6-only network unreasonably has to receive IPv4 RPKI data. Overall, the transimitted Data PDU type cannot be flexibly selected by the router.¶
The negative effects of the above problem are as follows:¶
Storing unused data on the router, which is unreasonable.¶
Unnecessary transmission and storage overhead.¶
Inefficient end-of-transmission acknowledgment. Multiple types of data are transmitted together. The router cannot use any type of these data unless it waits for all data to complete transmission.¶
The above negative effects will become worse when there are more kinds of RPKI data available [I-D.van-beijnum-sidrops-pathrpki][I-D.ietf-grow-rpki-as-cones][I-D.spaghetti-sidrops-rpki-asgroup]. The main problem of the RTR protocol is the lack of selective synchronization capability.¶
How about using different RTR versions for controlling the synchronized data, e.g., using RTR v0 if ASPA data are unwanted? This is not a good solution. First, the data selection is restricted to RTR versions and thus is not flexible either. Second, upgrading the version of RTR for future new RPKI data is not a proper choice, which is also a problem of existing RTR design. Specifically, existing RTR protocol has low extension capability. When there are new PDUs defined for transmission, a new RTR version needs to be issued. The new version protocol is not well compatible with the older ones, which induces some challenges on version negotiation, protocol implementation, and deployment. This document will primarily focus on the solving the inflexible synchronization problem. How to define an extensible protocol needs to be further discussed.¶
This section preliminarily proposes some independent solutions for achieving selective synchronization in the RTR protocol, while trying to keep the protocol's simplicity. A new protocol version may not necessarily be required.¶
Define a new type of PDU called Subscribing Data PDU. The new PDU will indicate the data types that the router is interested in. An example format of the PDU is shown in Figure 1. The field of PDU type is TBD. The Data Type fields indicate the interested data types (i.e., 4: IPv4 Prefix, 6: IPv6 Prefix, 9: Router Key, 11: ASPA).¶
The router can send the Subscribing Data PDU to the cache. After finishing the subscribing, the following PDUs, including Serial Notify, Serial Query, Reset Query, Cache Response, and Cache Reset, are only for the subscribed data. If the router wants to modify the subscription, a new Subscribing Data PDU can be sent for overwriting the previous subscription.¶
The existing PDUs, including Serial Notify, Serial Query, Reset Query, Cache Response, and Cache Reset, can be extended to carry the Data Type field. The values of the Data Type field can be 4 for IPv4 Prefix, for IPv6 Prefix, 9 for Router Key, and 11 for ASPA. An example format of the extended Serial Query PDU is shown in Figure 2. A router can send the extended Serial Query PDU for requesting a specific type of data.¶
End of Data PDU tells the router that all the requested data are synchronized. The End of Specific Data PDU can be defined for indicating a specific type of data has been synchronized. An example format of End of Specific Data PDU is shown in Figure 3. The field of PDU type is TBD. The Data Type field indicate the interested data types (i.e., 4: IPv4 Prefix, 6: IPv6 Prefix, 9: Router Key, 11: ASPA).¶
The type of data specified in End of Specific Data PDU will become ready for use. The router does not need to wait for all the data to complete transmission before it can use the specified data.¶