wdiff draft-ietf-idr-bgp-flowspec-oid-08.txt draft-ietf-idr-bgp-flowspec-oid-09.txt

Network Working Group J. Uttaro
Internet-Draft AT&T
Updates: 5575bis (if approved) J. Alcaide
Intended status: Standards Track C. Filsfils
Expires: November 10, 2019 January 9, 2020 D. Smith
Cisco
P. Mohapatra
Sproute Networks
May 9,
July 8, 2019

Revised Validation Procedure for BGP Flow Specifications
draft-ietf-idr-bgp-flowspec-oid-08
draft-ietf-idr-bgp-flowspec-oid-09

Abstract

This document describes a modification to the validation procedure
defined in RFC 5575bis for the dissemination of BGP flow
specifications. Flow
Specifications. RFC 5575bis requires that the originator of the flow
specification Flow
Specification matches the originator of the best-match unicast route
for the destination prefix embedded in the flow specification. Flow Specification. This
allows only BGP speakers within the data forwarding path (such as
autonomous system border routers) to originate BGP flow
specifications. Flow
Specifications. Though it is possible to disseminate such flow
specifications Flow
Specifications directly from border routers, it may be operationally
cumbersome in an autonomous system with a large number of border
routers having complex BGP policies. The modification proposed
herein enables flow specifications Flow Specifications to be originated from a
centralized BGP route controller.

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 November 10, 2019. January 9, 2020.

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. Requirements Language . . . . . . . . . . . . . . . . . . . . 2
2. Motivation . Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2
3. Introduction Motivation . . . . . . . . . . . . . . . . . . . . . . . . 4 . 3
4. Revised Validation Procedure . . . . . . . . . . . . . . . . 5
4.1. Revision of Route Feasibility . . . . . . . . . . . . . . 5
4.2. Revision of AS_PATH Validation . . . . . . . . . . . . . 6
5. Other RFC5575bis Considerations . . . . . . . . . . . . . . . 7
6. Topology Considerations . . . . . . . . . . . . . . . . . . . 8
7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 7
6. 9
8. Security Considerations . . . . . . . . . . . . . . . . . . . 7
7. 9
9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 7
8. 10
10. Normative References . . . . . . . . . . . . . . . . . . . . 7 10
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 8 10

1. Requirements Language

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC 2119 [RFC2119].

2. Motivation

Step (a) of the validation procedure in [RFC5575bis], section 6 is Introduction

[RFC5575bis] defined with the underlying assumption that the flow specification
NLRI traverses the same path, in the inter-domain and intra-domain
route distribution graph, as a new BGP [RFC4271] capability that of the longest-match unicast route
for the destination prefix embedded can be used
to distribute traffic Flow Specifications amongst BGP speakers in the flow specification.

In the case
support of inter-domain traffic filtering, for example, the flow
specification originator at the egress border routers filtering. The primary intention of ASN1 (RTR-D
and RTR-E in figure 1) matches the EBGP neighbor that advertised [RFC5575bis]
is to enable downstream autonomous systems to signal traffic
filtering policies to upstream autonomous systems. In this way,
traffic is filtered closer to the
longest match destination prefix (RTR-F source and RTR-G respectively).
Similarly, at the ingress border routers of ASN1 (RTR-A and RTR-B in
figure 1), upstream autonomous
system(s) avoid carrying the flow specification originator matches traffic to the egress IBGP
border routers downstream autonomous
system only to be discarded. [RFC5575bis] also enables more granular
traffic filtering based upon upper layer protocol information (e.g.,
protocol port numbers) as opposed to coarse IP destination prefix-
based filtering. Flow specification NLRIs received from a BGP peer
are subject to validity checks before being considered feasible and
subsequently installed within the respective Adj-RIB-In.

The validation procedure defined within [RFC5575bis] requires that had advertised
the originator of the Flow Specification NLRI matches the originator
of the best-match unicast route for the best-
match destination prefix (RTR-D and RTR-E respectively). embedded
in the Flow Specification. This allows only BGP speakers within the
data forwarding path (such as autonomous system border routers) to
originate BGP Flow Specification NLRIs. Though it is
true even when ingress possible to
disseminate such Flow Specification NLRIs directly from border
routers, it may be operationally cumbersome in an autonomous system
with a large number of border routers select paths having complex BGP policies.

This document describes a modification to the [RFC5575bis] validation
procedure allowing Flow Specification NLRIs to be originated from different
egress a
centralized BGP route controller within the local autonomous system
that is not in the data forwarding path. While the proposed
modification cannot be used for inter-domain coordination of traffic
filtering, it greatly simplifies distribution of intra-domain traffic
filtering policies within an autonomous system which has a large
number of border routers as best path based upon IGP distance (as an
example, RTR-A chooses RTR-D's path as best; RTR-B chooses RTR-E as having complex BGP policies. By relaxing
the best path).

/ - - - - - - - - - - - - - -
| ASN1 |
+-------+ +-------+
| validation procedure for iBGP, the proposed modification allows
Flow Specifications to be distributed in a standard and scalable
manner throughout an autonomous system.

3. Motivation

Step (a) of the validation procedure in [RFC5575bis], section 6 is
defined with the underlying assumption that the Flow Specification
NLRI traverses the same path, in the inter-domain and intra-domain
route distribution graph, as that of the longest-match unicast route
for the destination prefix embedded in the Flow Specification.

In the case of inter-domain traffic filtering, the Flow Specification
originator at the egress border routers of an AS (e.g. RTR-D and
RTR-E of ASN1 in figure 1) matches the eBGP neighbor that advertised
the longest match destination prefix (see RTR-F and RTR-G
respectively in figure 1). Similarly, at the ingress border routers
of ASN (see RTR-A and RTR-B of ASN1 in figure 1), the Flow
Specification originator matches the egress iBGP border routers that
had advertised the unicast route for the best-match destination
prefix (see RTR-D and RTR-E respectively in figure 1). This is true
even when ingress border routers select paths from different egress
border routers as best path based upon IGP distance. For example, in
figure 1:

RTR-A chooses RTR-D's path as best

RTR-B chooses RTR-E as the best path

/ - - - - - - - - - - - - - -
| ASN1 |
+-------+ +-------+
| | | | | |
| RTR-A | | RTR-B |
| | | | | |
+-------+ +-------+
| \ / |
IBGP
iBGP \ / IBGP iBGP
| \ / |
+-------+
| | | |
| RTR-C |
| | RC | |
+-------+
| / \ |
/ \
| IBGP iBGP / \ IBGP iBGP |
+-------+ +-------+
| | RTR-D | | RTR-E | |
| | | |
| | | | | |
+-------+ +-------+
| | | |
- - -|- - - - - - - - -|- - -/
| EBGP EBGP eBGP eBGP |
- - -|- - - - - - - - -|- - -/
| | | |
+-------+ +-------+
| | | | | |
| RTR-F | | RTR-G |
| | | | | |
+-------+ +-------+
| ASN2 |
/ - - - - - - - - - - - - - -

Figure 1

It is highly desirable that each ASN is able to protect itself
independently from network security attacks using the BGP flow
specification NLRI for intra-domain purposes only. Network operators
often deploy a dedicated Security Operations Center (SOC) within
their ASN to monitor and detect such security attacks. To mitigate
attacks in a scalable intra-domain manner, operators require the
ability to originate intra-domain flow specification NLRIs from a
central BGP route controller that is not within the data forwarding
plane. In this way, operators can direct border routers within their
ASN with specific attack mitigation actions (drop the traffic,
forward to a clean-pipe center, etc.). To originate a flow
specification NLRI, a central BGP route controller must set itself as
the originator in the flowspec NLRI. This is necessary given the
route controller is originating the flow specification not reflecting
it, and to avoid the complexity of having to determine the egress
border router whose path was chosen as the best in each of the
ingress border routers. It thus becomes necessary to modify step (a)
of the [RFC5575bis] validation procedure such that an IBGP peer that
is not within the data forwarding plane may originate flow
specification NLRIs.

3. Introduction

[RFC5575bis] defined a new BGP capability that can be used to
distribute traffic flow specifications amongst BGP speakers in
support of traffic filtering. The primary intention of [RFC5575bis]
is to enable downstream autonomous systems to signal traffic
filtering policies to upstream autonomous systems. In this way,
traffic is filtered closer to the source and the upstream autonomous
system(s) avoid carrying the traffic to the downstream autonomous
system only to be discarded. [RFC5575bis] also enables more granular
traffic filtering based upon upper layer protocol information (e.g.,
protocol port numbers) as opposed to coarse IP destination prefix-
based filtering. Flow specification NLRIs received from a BGP peer
are subject to validity checks before being considered feasible and
subsequently installed within the respective Adj-RIB-In. The
validation procedure defined within [RFC5575bis] requires that the
originator of the flow specification NLRI matches the originator of
the best-match unicast route for the destination prefix embedded in
the flow specification. This allows only BGP speakers [RFC4271]
within the data forwarding path (such as autonomous system border
routers) to originate BGP flow specification NLRIs. Though it -

Figure 1

It is
possible highly desirable that the mechanisms exist to disseminate such flow specification NLRIs directly protect each ASN
independently from
border routers, it may be operationally cumbersome in an autonomous
system with a large number of border routers having complex network security attacks using the BGP
policies. This document describes Flow
Specification NLRI for intra-domain purposes only. Network operators
often deploy a modification dedicated Security Operations Center (SOC) within
their ASN to monitor and detect such security attacks. To mitigate
attacks within a domain (AS or group of ASes), operators require the [RFC5575bis]
validation procedure allowing flow specification NLRIs
ability to be
originated originate intra-domain Flow Specification NLRIs from a centralized
central BGP route controller within the local
autonomous system that is not in within the data forwarding path. While
plane. In this way, operators can direct border routers within their
ASN with specific attack mitigation actions (drop the
proposed modification cannot be used for inter-domain coordination of
traffic filtering, it greatly simplifies distribution traffic,
forward to a clean-pipe center, etc.).

To originate a Flow Specification NLRI, a central BGP route
controller must set itself as the originator in the Flow
Specification NLRI. This is necessary given the route controller is
originating the Flow Specification rather than reflecting it, and to
avoid the complexity of intra-domain
traffic filtering policies having to determine the egress border router
whose path was chosen as the best in an autonomous system with a large
number each of the ingress border routers having complex BGP policies. By relaxing
routers. Thus, it is necessary to modify step (a) of the
[RFC5575bis] validation procedure for IBGP, the proposed modification allows
flow specifications to be distributed in a standard and scalable
manner throughout such that an autonomous system. iBGP peer that is not
within the data forwarding plane may originate Flow Specification
NLRIs.

4. Revised Validation Procedure

4.1. Revision of Route Feasibility

Step (a) of the validation procedure specified in [RFC5575bis],
section 6 is redefined as follows:

a. One of the following conditions MUST hold true.

1. The originator of the flow specification Flow Specification matches the
originator of the best-match unicast route for the
destination prefix embedded in the flow specification.

* Flow Specification.

2. The AS_PATH attribute of the flow specification Flow Specification does not
contain AS_SET and/or AS_SEQUENCE segments.

1. This condition SHOULD be enabled by default. This
default behavior should validate an empty AS_PATH.

2. This condition MAY be disabled by configuration on a BGP
speaker.

3. As an exception to this rule, a given AS_PATH without with AS_SET
and/or AS_SEQUENCE segments indicates MAY be validated by policy.

Explanation:

In this context, an empty AS_PATH means that
the flow specification was originated inside it does not have
AS_SET and/or AS_SEQUENCE segments, and local domain means the
local AS [RFC4271] or inside the local confederation of ASes (in the case
that the local AS belongs to a confederation of ASes) [RFC5065]. ASes [RFC5065]).

Thus, receiving a Flow Specification with an empty AS_PATH
indicates that the Flow Specification was originated inside the
local domain.

With this the above modification to the [RFC5575bis] validation
procedure, it is now
possible for an IBGP a BGP peer within the local domain that is not within
the data forwarding path
to can originate flow specification NLRIs. This applies whether the AS
belongs or not to a confederation of ASes. Checking the (newly
introduced) second condition above MAY be disabled by configuration
on a BGP speaker. However, it SHOULD be enabled by default. Flow Specification.

Disabling the new condition above (a.2.2) may be a good practice
when the administrator operator knows with certainty that there are is not flow specification NLRI a Flow
Specification originated inside the local AS (or local confederation). The default
behavior is thus domain.

Also, policy may be useful to validate an empty AS_PATH. In this context, an
empty AS_PATH means that it does not have AS_SET and/or AS_SEQUENCE
segments. Optionally, an implementation MAY also validate a specific set of non-empty AS_PATH.
AS_PATHs (a.2.3). For instance, example, it could validate a flowspec NLRI Flow
Specification whose AS_PATH contains only an AS_SEQUENCE of with ASes
that are all known (via
configuration) to belong to the same administrative domain.

Further,

4.2. Revision of AS_PATH Validation

[RFC5575bis] states that "BGP (flow specification) states:

o BGP implementations MUST also enforce that the AS_PATH attribute
of a route received via the External Border Gateway Protocol (EBGP)
(eBGP) contains the neighboring AS in the left-most position of
the AS_PATH attribute". attribute.

This rule is not valid for all topologies. For example, it prevents the exchange of BGP flow specification Flow Specification NLRIs at
Internet exchanges with BGP route servers. Therefore, this document
also redefines the [RFC5575bis] AS_PATH validation procedure
referenced above as follows:

o BGP flow specification Flow Specification implementations MUST enforce that the last AS
added within in
the left-most position of the AS_PATH attribute of a EBGP learned flow
specification NLRI MUST match Flow
Specification route received via the External Border Gateway
Protocol (eBGP) matches the last AS added within in the left-most position of the
AS_PATH attribute of the best-match unicast route for the
destination prefix embedded in the flow specification. Flow Specification NLRI.

Explanation:

For clarity, the AS in the left-most position of the AS_PATH means
the AS that was last added to the AS_SEQUENCE.

This proposed modification enables the exchange of BGP flow specification Flow
Specification NLRIs at Internet exchanges with BGP route servers
while at the same time, for security reasons, prevents an EBGP eBGP
peer from advertising an inter-domain flow
specification Flow Specification for a
destination prefix that it does not provide reachability
information for. Note, comparing

Comparing only the last ASes added is sufficient for EBGP eBGP learned flow specification
Flow Specification NLRIs. Requiring a full AS_PATH match would
limit origination of inter-
domain flow specifications inter-domain Flow Specifications to the
origin (or first) AS of the best-
match best-match unicast route for the destination
prefix embedded in the Flow Specification only. As such, a full
AS_PATH validity check may prevent transit ASes from originating
inter-domain Flow Specifications, which is not desirable.

Redefinition of this AS_PATH validation rule for a Flow
Specification does not mean that the original rule in [RFC5575bis]
cannot be enforced as well. Its enforcement remains optional per
[RFC4271]. That is, we can enforce the first AS in the AS_PATH to
be the same as the neighbor AS for any address-family route
(including a Flow Specification).

Using the new rule to validate a Flow Specification received from
an Internal Border Gateway Protocol (iBGP) peer is out of the
scope of this document. Note that in most scenarios such
validation would be redundant.

Using the new rule to validate a Flow Specification route received
from an External Border Gateway Protocol (eBGP) peer belonging to
the same local domain (in the case that the local AS belongs to a
confederation of ASes) is out of the scope of this document. Note
that although it's possible, its utility is dubious.

5. Other RFC5575bis Considerations

This section clarifies some of the terminology and rules referenced
in [RFC5575bis]. Namely:

o [RFC5575bis] references "the best-match unicast route for the
destination prefix embedded in the Flow Specification". For
clarity, this route is defined hereby as the best path of the
unicast network with the longer prefix-length that covers the
destination prefix embedded in the Flow Specification.

Explanation:

We consider only the best-match network and we do not consider
unicast route for non-best paths (even if it is received from the destination prefix embedded in same peer
as the flow
specification only. As such, Flow Specification).

o If a full AS_PATH validity check may
prevent transit ASes from originating inter-domain flow
specifications, which is not desirable.

This document also clarifies proper handling when the BGP flow
specification Flow Specification does not embed a destination prefix component. The
component, then the default behavior SHOULD be not to perform any the
validation procedure.
Further, support procedures.

Explanation:

The reason for two-octet AS number space this is out of because the scope of best-match unicast route is not
defined in this document. case.

o In the context of this context, document and [RFC5575bis], AS_PATH
attribute is defined as the reconstructed AS
Path path information (by
combining AS_PATH and AS4_PATH attributes, if the BGP speaker is a
NEW speaker and receives the route from an OLD speaker), according
to section 4.2.3 of [RFC6793].

[RFC5575bis] references "the best-match unicast route

o Support for the
destination prefix embedded in the flow specification". For clarity,
this route two-octet AS only implementations is defined hereby as the best path out of the unicast network scope
of this document (i.e. it's assumed that covers destination prefix embedded in the flow specification
with the longer prefix-length. In other words, we consider only the
best-match network and we do not consider unicast non-best paths
(even if it is received from the same peer than BGP speaker supports
[RFC6793]).

6. Topology Considerations

[RFC5575bis] indicates that the flowspec route).

Note that, per [RFC5575bis], originator may refer to the BGP
ORIGINATOR_ID
originator path attribute (ORIGINATOR_ID) or (if the attribute is not
present) the transport address of the peer from which we received the
update. If the later, latter applies, a network must should be designed so it
has a congruent topology. Otherwise, using

With the additional second condition (a.2) in the validation
procedure, non-congruent topologies are supported within the local
domain if the Flow Specification is originated within the local
domain.

Explanation:

Consider the following scenarios without the second condition
(a.2) being added to the validation procedure:

1. Consider a topology with two BGP speakers with two peering
sessions between the same pair of BGP speakers, them, one for unicast and one for flowspec, will cause the flowspec validation
procedure to fail. Consider, for example, the case where a BGP route
reflector receives the NLRIs from Flow
Specification. This is a route reflector client, thus not
receiving non-congruent topology. Let's
assume that the ORIGINATOR_ID attribute. If the speaker belongs to a
confederation [RFC5065] and we are receiving attribute was not received (e.g.
a flowspec route reflector receiving routes from
different peers than its best match unicast route, clients). In this
case, the flowspec Flow Specification validation procedure will fail as well.
because of the first condition (a.1).

2. Consider also a
misconfiguration where flowspec address-family is not configured for topology with a particular peering between different member-AS (but it BGP speaker within a confederation
of ASes, inside local AS X. ORIGINATOR_ID attribute is
configured for unicast). Even if we receive not
advertised within the flowspec route via a
redundant peer, we may receive local domain. Let's assume the unicast Flow
Specification route is received from peer A and the flowspec best-match
unicast route is received from different peers, peer B. Both peers belong in
local AS Y. Both AS X and thus flowspec AS Y belong to the same local
domain. The Flow Specification validation procedure will fail. Thus,
with also
fail because of the first condition (a.1).

In the examples above, if Flow Specifications are originated in
the same local domain, AS_PATH will not contain AS_SET and/or
AS_SEQUENCE segments. When the (newly introduced) second condition above applied,
incongruent (a.2) in the
validation procedure is used, the validation procedure will pass.
Thus, non-congruent topologies are supported.

Note that supported if the flowspec NLRI Flow
Specification is learned from another AS (and thus originated in the AS_PATH same local domain.

Even when the second condition (a.2) is not empty), used in the original validation procedures defined
procedure, a Flow Specification originated in [RFC5575bis] still apply a different local
domain needs a congruent topology. AS_SEQUENCE is not empty and incongruent topologies may cause
the first condition (a.1) in the validation rules procedure needs to be
evaluated. Because transport addresses for Flow Specification and
unicast routes are different, the validation procedure will fail.

5.
This is true both across domains and within domains. Consider
both cases:

* Consider the first example. If the Flow Specification route is
originated in another AS, the validation procedure will fail
because the topology is non-congruent within the domain.

* Consider the second example and modify it so AS X and AS Y
belong to different local domains (no confederation of ASes
exists). The validation procedure will fail because the
topology is non-congruent across domains.

7. IANA Considerations

This memo includes no request to IANA.

8. Security Considerations

No new security issues are introduced by relaxing the validation
procedure for IBGP iBGP learned flow specifications. Flow Specifications. With this proposal,
the security characteristics of BGP flow specifications Flow Specifications remain
equivalent to the existing security properties of BGP unicast
routing. Traffic flow specifications Flow Specifications learned from IBGP iBGP peers are
trusted, hence, it is not required to validate that the originator of
an intra-domain traffic flow specification Flow Specification matches the originator of
the best-match unicast route for the flow destination prefix.
Conversely, this proposal continues to enforce the validation
procedure for EBGP eBGP learned traffic flow specifications. Flow Specifications. In this way,
the security properties of [RFC5575bis] are maintained such that an
EBGP
eBGP peer cannot cause a denial-of-service attack by advertising an
inter-domain flow specification Flow Specification for a destination prefix that it does
not provide reachability information for.

9. Acknowledgements

The authors would like to thank Han Nguyen for his direction on this
work as well as Waqas Alam, Keyur Patel, Robert Raszuk, Eric Rosen
and Shyam Sethuram for their review comments.

10. Normative References

[I-D.ietf-idr-rfc5575bis]
Hares, S., Loibl, C., Raszuk, R., McPherson, D., and M.
Bacher, "Dissemination of Flow Specification Rules",
draft-ietf-idr-rfc5575bis-14 (work in progress), April
2019.

[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119,
DOI 10.17487/RFC2119, March 1997,
<https://www.rfc-editor.org/info/rfc2119>.

[RFC4271] Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A
Border Gateway Protocol 4 (BGP-4)", RFC 4271,
DOI 10.17487/RFC4271, January 2006,
<https://www.rfc-editor.org/info/rfc4271>.

[RFC4456] Bates, T., Chen, E., and R. Chandra, "BGP Route
Reflection: An Alternative to Full Mesh Internal BGP
(IBGP)", RFC 4456, DOI 10.17487/RFC4456, April 2006,
<https://www.rfc-editor.org/info/rfc4456>.

[RFC5065] Traina, P., McPherson, D., and J. Scudder, "Autonomous
System Confederations for BGP", RFC 5065,
DOI 10.17487/RFC5065, August 2007,
<https://www.rfc-editor.org/info/rfc5065>.

[RFC6793] Vohra, Q. and E. Chen, "BGP Support for Four-Octet
Autonomous System (AS) Number Space", RFC 6793,
DOI 10.17487/RFC6793, December 2012,
<https://www.rfc-editor.org/info/rfc6793>.

Authors' Addresses
James Uttaro
AT&T
200 S. Laurel Ave
Middletown, NJ 07748
USA

Email: ju1738@att.com

Juan Alcaide
Cisco
7100 Kit Creek Road
Research Triangle Park, NC 27709
USA

Email: jalcaide@cisco.com

Clarence Filsfils
Cisco

Email: cf@cisco.com

David Smith
Cisco
111 Wood Ave South
Iselin, NJ 08830
USA

Email: djsmith@cisco.com

Pradosh Mohapatra
Sproute Networks

Email: mpradosh@yahoo.com