Status of This Memo

This is an Internet Standards Track document.¶

This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community. It has received public review and has been approved for publication by the Internet Engineering Steering Group (IESG). Further information on Internet Standards is available in Section 2 of RFC 7841.¶

Information about the current status of this document, any errata, and how to provide feedback on it may be obtained at https://www.rfc-editor.org/info/rfc0000.¶

Copyright Notice

Copyright (c) 2020 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 Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶

Table of Contents

1. Introduction

This document obsoletes "Dissemination of Flow Specification Rules" [RFC5575] (see Appendix B for the differences). This document also obsoletes "Clarification of the Flowspec Redirect Extended Community" [RFC7674], since it incorporates the encoding of the BGP Flow Specification Redirect Extended Community in Section 7.4.¶

Modern IP routers have the capability to forward traffic and to classify, shape, rate limit, filter, or redirect packets based on administratively defined policies. These traffic policy mechanisms allow the operator to define match rules that operate on multiple fields of the packet header. Actions, such as the ones described above, can be associated with each rule.¶

The n-tuple consisting of the matching criteria defines an aggregate traffic Flow Specification. The matching criteria can include elements such as source and destination address prefixes, IP protocol, and transport protocol port numbers.¶

Section 4 of this document defines a general procedure to encode Flow Specifications for aggregated traffic flows so that they can be distributed as a BGP [RFC4271] NLRI. Additionally, Section 7 of this document defines the required Traffic Filtering Actions BGP Extended Communities and mechanisms to use BGP for intra- and inter-provider distribution of traffic filtering rules to filter (distributed) denial-of-service (DoS) attacks.¶

By expanding routing information with Flow Specifications, the routing system can take advantage of the ACL (Access Control List) or firewall capabilities in the router's forwarding path. Flow Specifications can be seen as more specific routing entries to a unicast prefix and are expected to depend upon the existing unicast data information.¶

A Flow Specification received from an external autonomous system will need to be validated against unicast routing before being accepted (Section 6). The Flow Specification received from an internal BGP peer within the same autonomous system [RFC4271] is assumed to have been validated prior to transmission within the internal BGP (iBGP) mesh of an autonomous system. If the aggregate traffic flow defined by the unicast destination prefix is forwarded to a given BGP peer, then the local system can install more specific Flow Specifications that may result in different forwarding behavior, as requested by this system.¶

From an operational perspective, the utilization of BGP as the carrier for this information allows a network service provider to reuse both internal route distribution infrastructure (e.g., route reflector or confederation design) and existing external relationships (e.g., inter-domain BGP sessions to a customer network).¶

While it is certainly possible to address this problem using other mechanisms, this solution has been utilized in deployments because of the substantial advantage of being an incremental addition to already deployed mechanisms.¶

In current deployments, the information distributed by this extension is originated both manually as well as automatically, the latter by systems that are able to detect malicious traffic flows. When automated systems are used, care should be taken to ensure the correctness of the automated system. The limitations of the receiving systems that need to process these automated Flow Specifications need to be taken in consideration as well (see also Section 12).¶

This specification defines required protocol extensions to address most common applications of IPv4 unicast and VPNv4 unicast filtering. The same mechanism can be reused and new match criteria added to address similar filtering needs for other BGP address families, such as IPv6 families [IDR-FLOW-SPEC].¶

4. Dissemination of IPv4 Flow Specification Information

This document defines a Flow Specification NLRI type (Figure 1) that may include several components, such as destination prefix, source prefix, protocol, ports, and others (see Section 4.2 below).¶

This NLRI information is encoded using MP_REACH_NLRI and MP_UNREACH_NLRI attributes, as defined in [RFC4760]. When advertising Flow Specifications, the Length of the Next-Hop Network Address MUST be set to 0. The Network Address of the Next-Hop field MUST be ignored.¶

The NLRI field of the MP_REACH_NLRI and MP_UNREACH_NLRI is encoded as one or more 2-tuples of the form <length, NLRI value>. It consists of a 1- or 2-octet length field followed by a variable-length NLRI value. The length is expressed in octets.¶

                  +-------------------------------+
                  |    length (0xnn or 0xfnnn)    |
                  +-------------------------------+
                  |    NLRI value   (variable)    |
                  +-------------------------------+

Implementations wishing to exchange Flow Specification MUST use BGP's Capability Advertisement facility to exchange the Multiprotocol Extension Capability Code (Code 1), as defined in [RFC4760]. The (AFI, SAFI) pair carried in the Multiprotocol Extension Capability MUST be (AFI=1, SAFI=133) for IPv4 Flow Specification and (AFI=1, SAFI=134) for VPNv4 Flow Specification.¶

                    0   1   2   3   4   5   6   7
                  +---+---+---+---+---+---+---+---+
                  | e | a |  len  | 0 |lt |gt |eq |
                  +---+---+---+---+---+---+---+---+

                    0   1   2   3   4   5   6   7
                  +---+---+---+---+---+---+---+---+
                  | e | a |  len  | 0 | 0 |not| m |
                  +---+---+---+---+---+---+---+---+

                   0   1   2   3   4   5   6   7
                 +---+---+---+---+---+---+---+---+
                 | 0 | 0 | 0 | 0 |LF |FF |IsF|DF |
                 +---+---+---+---+---+---+---+---+

5. Traffic Filtering

Traffic filtering policies have been traditionally considered to be relatively static. Limitations of these static mechanisms caused this new dynamic mechanism to be designed for the three new applications of traffic filtering:¶

• Prevention of traffic-based, denial-of-service (DOS) attacks¶
• Traffic filtering in the context of BGP/MPLS VPN service¶
• Centralized traffic control for SDN/NFV networks¶

These applications require coordination among service providers and/or coordination among the AS within a service provider.¶

The Flow Specification NLRI defined in Section 4 conveys information about traffic filtering rules for traffic that should be discarded or handled in a manner specified by a set of predefined actions (which are defined in BGP Extended Communities). This mechanism is primarily designed to allow an upstream autonomous system to perform inbound filtering in their ingress routers of traffic that a given downstream AS wishes to drop.¶

In order to achieve this goal, this document specifies two application-specific NLRI identifiers that provide traffic filters and a set of actions encoding in BGP Extended Communities. The two application-specific NLRI identifiers are:¶

• IPv4 Flow Specification identifier (AFI=1, SAFI=133) along with specific semantic rules for IPv4 routes and¶
• VPNv4 Flow Specification identifier (AFI=1, SAFI=134) value, which can be used to propagate traffic filtering information in a BGP/MPLS VPN environment.¶

Encoding of the NLRI is described in Section 4 for IPv4 Flow Specification and in Section 8 for VPNv4 Flow Specification. The filtering actions are described in Section 7.¶

6. Validation Procedure

Flow Specifications received from a BGP peer that are accepted in the respective Adj-RIB-In are used as input to the route selection process. Although the forwarding attributes of two routes for the same Flow Specification prefix may be the same, BGP is still required to perform its path selection algorithm in order to select the correct set of attributes to advertise.¶

The first step of the BGP Route Selection procedure (Section 9.1.2 of [RFC4271]) is to exclude from the selection procedure routes that are considered non-feasible. In the context of IP routing information, this step is used to validate that the NEXT_HOP attribute of a given route is resolvable.¶

The concept can be extended, in the case of the Flow Specification NLRI, to allow other validation procedures.¶

The validation process described below validates Flow Specifications against unicast routes received over the same AFI but the associated unicast routing information SAFI:¶

• Flow Specification received over SAFI=133 will be validated against routes received over SAFI=1.¶
• Flow Specification received over SAFI=134 will be validated against routes received over SAFI=128.¶

In the absence of explicit configuration, a Flow Specification NLRI MUST be validated such that it is considered feasible if and only if all of the conditions below are true:¶

1. A destination prefix component is embedded in the Flow Specification.¶
2. The originator of the Flow Specification matches the originator of the best-match unicast route for the destination prefix embedded in the Flow Specification (this is the unicast route with the longest possible prefix length covering the destination prefix embedded in the Flow Specification).¶
3. There are no "more-specific" unicast routes, when compared with the flow destination prefix, that have been received from a different neighboring AS than the best-match unicast route, which has been determined in rule b.¶

However, rule a MAY be relaxed by explicit configuration, permitting Flow Specifications that include no destination prefix component. If such is the case, rules b and c are moot and MUST be disregarded.¶

By "originator" of a BGP route, we mean either the address of the originator in the ORIGINATOR_ID Attribute [RFC4456] or the source IP address of the BGP peer, if this path attribute is not present.¶

BGP implementations MUST also enforce that the AS_PATH attribute of a route received via the External Border Gateway Protocol (eBGP) contains the neighboring AS in the left-most position of the AS_PATH attribute. While this rule is optional in the BGP specification, it becomes necessary to enforce it here for security reasons.¶

The best-match unicast route may change over the time independently of the Flow Specification NLRI. Therefore, a revalidation of the Flow Specification NLRI MUST be performed whenever unicast routes change. Revalidation is defined as retesting rules a to c as described above.¶

Explanation:¶

The underlying concept is that the neighboring AS that advertises the best unicast route for a destination is allowed to advertise Flow Specification information that conveys a destination prefix that is more or equally specific. Thus, as long as there are no "more-specific" unicast routes received from a different neighboring AS, which would be affected by that Flow Specification, the Flow Specification is validated successfully.¶

The neighboring AS is the immediate destination of the traffic described by the Flow Specification. If it requests these flows to be dropped, that request can be honored without concern that it represents a denial of service in itself. The reasoning is that this is as if the traffic is being dropped by the downstream autonomous system, and there is no added value in carrying the traffic to it.¶

7. Traffic Filtering Actions

This document defines a minimum set of Traffic Filtering Actions that it standardizes as BGP extended communities [RFC4360]. This is not meant to be an inclusive list of all the possible actions but only a subset that can be interpreted consistently across the network. Additional actions can be defined as either requiring standards or as vendor specific.¶

The default action for a matching Flow Specification is to accept the packet (treat the packet according to the normal forwarding behavior of the system).¶

This document defines the following extended communities values shown in Table 8 in the form 0xttss, where tt indicates the type and ss indicates the sub-type of the extended community. Encodings for these extended communities are described below.¶

Multiple Traffic Filtering Actions defined in this document may be present for a single Flow Specification and SHOULD be applied to the traffic flow (for example, traffic-rate-bytes and rt-redirect can be applied to packets at the same time). If not all of the Traffic Filtering Actions can be applied to a traffic flow, they should be treated as interfering Traffic Filtering Actions (see below).¶

Some Traffic Filtering Actions may interfere with each other or even contradict. Section 7.7 of this document provides general considerations on such Traffic Filtering Action interference. Any additional definition of Traffic Filtering Actions SHOULD specify the action to take if those Traffic Filtering Actions interfere (also with existing Traffic Filtering Actions).¶

All Traffic Filtering Actions are specified as transitive BGP Extended Communities.¶

   0                   1                   2                   3
   0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
  | Traffic Action Field                                          |
  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
  | Tr. Action Field (cont.)  |S|T|
  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   0                   1                   2                   3
   0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
  |   reserved    |   reserved    |   reserved    |   reserved    |
  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
  |   reserved    | r.|    DSCP   |
  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

8. Dissemination of Traffic Filtering in BGP/MPLS VPN Networks

Provider-based Layer 3 VPN networks, such as the ones using a BGP/ MPLS IP VPN [RFC4364] control plane, may have different traffic filtering requirements than Internet service providers. But also Internet service providers may use those VPNs for scenarios like having the Internet routing table in a VRF, resulting in the same traffic filtering requirements as defined for the global routing table environment within this document. This document defines an additional BGP NLRI type (AFI=1, SAFI=134) value, which can be used to propagate Flow Specification in a BGP/MPLS VPN environment.¶

The NLRI format for this address family consists of a fixed-length Route Distinguisher field (8 octets) followed by the Flow Specification NLRI value (Section 4.2). The NLRI length field shall include both the 8 octets of the Route Distinguisher as well as the subsequent Flow Specification NLRI value. The resulting encoding is shown in Figure 7.¶

                 +--------------------------------+
                 | length (0xnn or 0xfn nn)       |
                 +--------------------------------+
                 | Route Distinguisher (8 octets) |
                 +--------------------------------+
                 |    NLRI value  (variable)      |
                 +--------------------------------+

Propagation of this NLRI is controlled by matching Route Target extended communities associated with the BGP path advertisement with the VRF import policy, using the same mechanism as described in BGP/ MPLS IP VPNs [RFC4364].¶

Flow Specifications received via this NLRI apply only to traffic that belongs to the VRF(s) in which it is imported. By default, traffic received from a remote PE is switched via an MPLS forwarding decision and is not subject to filtering.¶

Contrary to the behavior specified for the non-VPN NLRI, Flow Specifications are accepted by default, when received from remote PE routers.¶

The validation procedure (Section 6) and Traffic Filtering Actions (Section 7) are the same as for IPv4.¶

10. Error Handling

Error handling according to [RFC7606] and [RFC4760] applies to this specification.¶

This document introduces Traffic Filtering Action Extended Communities. Malformed Traffic Filtering Action Extended Communities in the sense of Section 7.14 of [RFC7606] are Extended Community values that cannot be decoded according to Section 7 of this document.¶

11. IANA Considerations

This section complies with [RFC7153].¶

<CODE BEGINS>
   enum flow-spec-safi {
          value 133;
          description
            "Dissemination of Flow Specification rules SAFI.";
        }
   enum l3vpn-flow-spec-safi {
          value 134;
          description
            "L3VPN Dissemination of Flow Specification rules SAFI.";
        }

<CODE ENDS>

<CODE BEGINS>
   revision [revision date] {
     description "Non-backwards-compatible change of SAFI names
                  (SAFI values 133, 134).";
     reference
       "RFCXXXX: Dissemination of Flow Specification Rules.";
  }

<CODE ENDS>

12. Security Considerations

As long as Flow Specifications are restricted to match the corresponding unicast routing paths for the relevant prefixes (Section 6), the security characteristics of this proposal are equivalent to the existing security properties of BGP unicast routing. Any relaxation of the validation procedure described in Section 6 may allow unwanted Flow Specifications to be propagated, and thus unwanted Traffic Filtering Actions may be applied to flows.¶

Where the above mechanisms are not in place, this could open the door to further denial-of-service attacks, such as unwanted traffic filtering, remarking, or redirection.¶

Deployment of specific relaxations of the validation within an administrative boundary of a network are useful in some networks for quickly distributing filters to prevent denial-of-service attacks. For a network to utilize this relaxation, the BGP policies must support additional filtering, since the origin AS field is empty. Specifications relaxing the validation restrictions MUST contain security considerations that provide details on the required additional filtering. For example, the use of origin validation can provide enhanced filtering within an AS confederation.¶

Inter-provider routing is based on a web of trust. Neighboring autonomous systems are trusted to advertise valid reachability information. If this trust model is violated, a neighboring autonomous system may cause a denial-of-service attack by advertising reachability information for a given prefix for which it does not provide service (unfiltered address space hijack). Since validation of the Flow Specification is tied to the announcement of the best unicast route, the failure in the validation of best path route may prevent the Flow Specification from being used by a local router. Possible mitigations are [RFC6811] and [RFC8205].¶

On Internet Exchange Points (IXPs), routes are often exchanged via route servers that do not extend the AS_PATH. In such cases, it is not possible to enforce the left-most AS in the AS_PATH to be the neighbor AS (the AS of the route server). Since the validation of Flow Specification (Section 6) depends on this, additional care must be taken. It is advised to use a strict inbound route policy in such scenarios.¶

Enabling firewall-like capabilities in routers without centralized management could make certain failures harder to diagnose. For example, it is possible to allow TCP packets to pass between a pair of addresses but not ICMP packets. It is also possible to permit packets smaller than 900 or greater than 1000 octets to pass between a pair of addresses but not packets whose length is in the range 900-1000. Such behavior may be confusing, and these capabilities should be used with care whether manually configured or coordinated through the protocol extensions described in this document.¶

Flow Specification BGP speakers (e.g., automated DDoS controllers) not properly programmed, algorithms that are not performing as expected, or simply rogue systems may announce unintended Flow Specifications, send updates at a high rate, or generate a high number of Flow Specifications. This may stress the receiving systems, exceed their capacity, or lead to unwanted Traffic Filtering Actions being applied to flows.¶

While the general verification of the Flow Specification NLRI is specified in this document (Section 6), the Traffic Filtering Actions received by a third party may need custom verification or filtering. In particular, all non-traffic-rate actions may allow a third party to modify packet forwarding properties and potentially gain access to other routing-tables/VPNs or undesired queues. This can be avoided by proper filtering/screening of the Traffic Filtering Action communities at network borders and only exposing a predefined subset of Traffic Filtering Actions (see Section 7) to third parties. One way to achieve this is by mapping user-defined communities, which can be set by the third party, to Traffic Filtering Actions and not accepting Traffic Filtering Action extended communities from third parties.¶

This extension adds additional information to Internet routers. These are limited in terms of the maximum number of data elements they can hold as well as the number of events they are able to process in a given unit of time. Service providers need to consider the maximum capacity of their devices and may need to limit the number of Flow Specifications accepted and processed.¶

Appendix A. Example Python code: flow_rule_cmp

<CODE BEGINS>
"""
Copyright (c) 2020 IETF Trust and the persons identified as
authors of the code.  All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, is permitted pursuant to, and subject to the license
terms contained in, the Simplified BSD License set forth in Section
4.c of the IETF Trust's Legal Provisions Relating to IETF Documents
(https://trustee.ietf.org/license-info).
"""

import itertools
import collections
import ipaddress


EQUAL = 0
A_HAS_PRECEDENCE = 1
B_HAS_PRECEDENCE = 2
IP_DESTINATION = 1
IP_SOURCE = 2

FS_component = collections.namedtuple('FS_component',
                                      'component_type op_value')


class FS_nlri(object):
    """
    FS_nlri class implementation that allows sorting.

    By calling .sort() on an array of FS_nlri objects these will be
    sorted according to the flow_rule_cmp algorithm.

    Example:
    nlri = [ FS_nlri(components=[
             FS_component(component_type=IP_DESTINATION,
                    op_value=ipaddress.ip_network('10.1.0.0/16') ),
             FS_component(component_type=4,
                    op_value=bytearray([0,1,2,3,4,5,6])),
             ]),
             FS_nlri(components=[
             FS_component(component_type=5,
                    op_value=bytearray([0,1,2,3,4,5,6])),
             FS_component(component_type=6,
                    op_value=bytearray([0,1,2,3,4,5,6])),
             ]),
           ]
    nlri.sort() # sorts the array according to the algorithm
    """
    def __init__(self, components = None):
        """
        components: list of type FS_component
        """
        self.components = components

    def __lt__(self, other):
        # use the below algorithm for sorting
        result = flow_rule_cmp(self, other)
        if result == B_HAS_PRECEDENCE:
            return True
        else:
            return False


def flow_rule_cmp(a, b):
    """
    Example of the flowspec comparison algorithm.
    """
    for comp_a, comp_b in itertools.zip_longest(a.components,
                                           b.components):
        # If a component type does not exist in one rule
        # this rule has lower precedence
        if not comp_a:
            return B_HAS_PRECEDENCE
        if not comp_b:
            return A_HAS_PRECEDENCE
        # Higher precedence for lower component type
        if comp_a.component_type < comp_b.component_type:
            return A_HAS_PRECEDENCE
        if comp_a.component_type > comp_b.component_type:
            return B_HAS_PRECEDENCE
        # component types are equal -> type specific comparison
        if comp_a.component_type in (IP_DESTINATION, IP_SOURCE):
            # assuming comp_a.op_value, comp_b.op_value of
            # type ipaddress.IPv4Network
            if comp_a.op_value.overlaps(comp_b.op_value):
                # longest prefixlen has precedence
                if comp_a.op_value.prefixlen > \
                        comp_b.op_value.prefixlen:
                    return A_HAS_PRECEDENCE
                if comp_a.op_value.prefixlen < \
                        comp_b.op_value.prefixlen:
                    return B_HAS_PRECEDENCE
                # components equal -> continue with next component
            elif comp_a.op_value > comp_b.op_value:
                return B_HAS_PRECEDENCE
            elif comp_a.op_value < comp_b.op_value:
                return A_HAS_PRECEDENCE
        else:
            # assuming comp_a.op_value, comp_b.op_value of type
            # bytearray
            if len(comp_a.op_value) == len(comp_b.op_value):
                if comp_a.op_value > comp_b.op_value:
                    return B_HAS_PRECEDENCE
                if comp_a.op_value < comp_b.op_value:
                    return A_HAS_PRECEDENCE
                # components equal -> continue with next component
            else:
                common = min(len(comp_a.op_value), len(comp_b.op_value))
                if comp_a.op_value[:common] > comp_b.op_value[:common]:
                    return B_HAS_PRECEDENCE
                elif comp_a.op_value[:common] < \
                        comp_b.op_value[:common]:
                    return A_HAS_PRECEDENCE
                # the first common bytes match
                elif len(comp_a.op_value) > len(comp_b.op_value):
                    return A_HAS_PRECEDENCE
                else:
                    return B_HAS_PRECEDENCE
    return EQUAL

<CODE ENDS>

Appendix B. Comparison with RFC 5575

This document includes numerous editorial changes to [RFC5575]. It also completely incorporates the redirect action clarification document [RFC7674]. It is recommended to read the entire document. The authors, however, want to point out the following technical changes to [RFC5575]:¶

• Section 1 introduces the Flow Specification NLRI. In [RFC5575], this NLRI was defined as an opaque key in BGPs database. This specification has removed all references to an opaque key property. BGP implementations are able to understand the NLRI encoding.¶
• Section 4.2.1.1 defines a numeric operator and comparison bit combinations. In [RFC5575], the meaning of those bit combination was not explicitly defined and left open to the reader.¶
• Sections 4.2.2.3-4.2.2.8, 4.2.2.10, and 4.2.2.11 make use of the above numeric operator. The allowed length of the comparison value was not consistently defined in [RFC5575].¶
• Section 7 defines all Traffic Filtering Action Extended communities as transitive extended communities. [RFC5575] defined the traffic-rate action to be non-transitive and did not define the transitivity of the other Traffic Filtering Action communities at all.¶
• Section 7.2 introduces a new Traffic Filtering Action (traffic-rate-packets). This action did not exist in [RFC5575].¶
• Section 7.4 contains the same redirect actions already defined in [RFC5575], however, these actions have been renamed to "rt-redirect" to make it clearer that the redirection is based on route-target. This section also completely incorporates the [RFC7674] clarifications of the Flowspec Redirect Extended Community.¶
• Section 7.7 contains general considerations on interfering traffic actions. Section 7.3 also cross-references Section 7.7. [RFC5575] did not mention this.¶
• Section 10 contains new error handling.¶

Authors' Addresses