Internet DRAFT - draft-ietf-sidr-bgpsec-overview

draft-ietf-sidr-bgpsec-overview



 



SIDR Working Group                                           M. Lepinski
Internet Draft                                                       NCF
Intended status: Informational                                 S. Turner
Expires: December 24, 2016                                         sn3rd
                                                           June 22, 2016


                         An Overview of BGPsec
                   draft-ietf-sidr-bgpsec-overview-08


Abstract

   This document provides an overview of a security extension to the
   Border Gateway Protocol (BGP) referred to as BGPsec.  BGPsec improves
   security for BGP routing.

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 http://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 December 24, 2016.

Copyright Notice

   Copyright (c) 2016 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
   (http://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.
 


Lepinski and Turner    Expires December 24, 2016                [Page 1]

Internet-Draft              BGPsec Overview                June 22, 2016


Table of Contents

   1. Introduction  . . . . . . . . . . . . . . . . . . . . . . . . .  2
   2. Background  . . . . . . . . . . . . . . . . . . . . . . . . . .  3
   3. BGPsec Operation  . . . . . . . . . . . . . . . . . . . . . . .  4
     3.1. Negotiation of BGPsec . . . . . . . . . . . . . . . . . . .  4
     3.2. Update signing and validation . . . . . . . . . . . . . . .  5
   4. Design and Deployment Considerations  . . . . . . . . . . . . .  7
     4.1. Disclosure of topology information  . . . . . . . . . . . .  7
     4.2. BGPsec router assumptions . . . . . . . . . . . . . . . . .  7
     4.3. BGPsec and consistency of externally visible data . . . . .  8
   5. Security Considerations . . . . . . . . . . . . . . . . . . . .  8
   6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . .  9
   7. References  . . . . . . . . . . . . . . . . . . . . . . . . . .  9
     7.1. Normative References  . . . . . . . . . . . . . . . . . . .  9
     7.2. Informative References  . . . . . . . . . . . . . . . . . . 10
   Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 10


1. Introduction

   BGPsec (Border Gateway Protocol Security) is an extension to the
   Border Gateway Protocol (BGP) that provides improved security for BGP
   routing [RFC4271]. This document contains a brief overview of BGPsec
   and its envisioned usage.

   A more detailed discussion of BGPsec is provided in the following set
   of documents:

     *  [RFC7132]:

        A threat model describing the security context in which BGPsec
        is intended to operate.

     *  [RFC7353]:

        A set of requirements for BGP path security, which BGPsec is
        intended to satisfy.

     *  [I-D.sidr-bgpsec-protocol]:

        A standards track document specifying the BGPsec extension to
        BGP.

     *  [I-D.sidr-as-migration]:

        A standards track document describing how to implement an AS
        Number migration while using BGPsec.
 


Lepinski and Turner    Expires December 24, 2016                [Page 2]

Internet-Draft              BGPsec Overview                June 22, 2016


     *  [I-D.sidr-bgpsec-ops]:

        An informational document describing operational considerations.

     * [I-D.sidr-bgpsec-pki-profiles]:

        A standards track document specifying a profile for X.509
        certificates that bind keys used in BGPsec to Autonomous System
        numbers, as well as associated Certificate Revocation Lists
        (CRLs), and certificate requests.

     *  [I-D.sidr-bgpsec-algs]

        A standards track document specifying suites of signature and
        digest algorithms for use in BGPsec.

   In addition to this document set, some readers might be interested in
   [I-D.sriram-bgpsec-design-choices], an informational document
   describing the choices that were made the by the design team prior to
   the publication of the -00 version of draft-ietf-sidr-bgpsec-
   protocol. Discussion of design choices made since the publication of
   the -00 can be found in the archives of the SIDR working group
   mailing list.

2. Background

   The motivation for developing BGPsec is that BGP does not include
   mechanisms that allow an Autonomous System (AS) to verify the
   legitimacy and authenticity of BGP route advertisements (see for
   example, [RFC4272]).

   The Resource Public Key Infrastructure (RPKI), described in
   [RFC6480], provides a first step towards addressing the validation of
   BGP routing data. RPKI resource certificates are issued to the
   holders of AS number and IP address resources, providing a binding
   between these resources and cryptographic keys that can be used to
   verify digital signatures. Additionally, the RPKI architecture
   specifies a digitally signed object, a Route Origination
   Authorization (ROA), that allows holders of IP address resources to
   authorize specific ASes to originate routes (in BGP) to these
   resources. Data extracted from a valid ROA can be used by a BGP
   speaker to determine whether the origin AS asserted in a received
   route has been authorized (by the Internet Number Resource holder) to
   originate that route (see [RFC6483] and [RFC7115]).

   By instituting a local policy that prefers routes with origins
   validated using RPKI data (versus routes to the same prefix that
   cannot be so validated) an AS can protect itself from configuration
 


Lepinski and Turner    Expires December 24, 2016                [Page 3]

Internet-Draft              BGPsec Overview                June 22, 2016


   errors by network operators and from certain mis-origination attacks.
   However, use of RPKI data alone provides little or no protection
   against a sophisticated attacker. Such an attacker could, for
   example, conduct a route hijacking attack by appending an authorized
   origin AS to an otherwise illegitimate AS path. (See [RFC7132] for a
   detailed discussion of the BGPsec threat model.)

   BGPsec extends the RPKI by adding an additional type of certificate,
   referred to as a BGPsec Router Certificate, that binds an AS number
   to a public signature verification key.  The corresponding private
   key is held by one or more BGP speakers within this AS. Private keys
   corresponding to public keys in such certificates are used within
   BGPsec to enable a BGP speaker to sign on behalf of its AS. The
   certificates thus allow a relying party to verify that a BGPsec
   signature was produced by a BGP speaker belonging to a given AS. The
   goal of BGPsec is to use such signatures to protect the AS path data
   in BGP update messages, so that each BGP speaker can assess the
   validity of this data in update messages that it receives.

3. BGPsec Operation

   The core of BGPsec is a new optional (non-transitive) attribute,
   called BGPsec_Path. This attribute includes both AS Path data as well
   as a sequence of digital signatures, one for each AS in the path.
   (The use of this new attribute is formally specified in [I-D.sidr-
   bgpsec-protocol].) A new signature is added to this sequence each
   time an update message leaves an AS. The signature is constructed so
   that any tampering with the AS path data or Network Layer
   Reachability Information (NLRI) in the BGPsec update message can be
   detected by the recipient of the message. 

3.1. Negotiation of BGPsec

   The use of BGPsec is negotiated using BGP capability advertisements
   [RFC5492]. Upon opening a BGP session with a peer, BGP speakers who
   support (and wish to use) BGPsec include a newly-defined capability
   in the OPEN message [I-D.sidr-bgpsec-protocol].

   The use of BGPsec is negotiated separately for each address family.
   This means that a BGP speaker could, for example, elect to use BGPsec
   for IPv6, but not for IPv4 (or vice versa) routes. Additionally, the
   use of BGPsec is negotiated separately in the send and receive
   directions. This means that a BGP speaker could, for example,
   indicate support for sending BGPsec update messages but require that
   messages it receives be traditional (non-BGPsec) update message. (To
   see why such a feature is useful, see Section 4.2.)

   If the use of BGPsec is negotiated in a BGP session (in a given
 


Lepinski and Turner    Expires December 24, 2016                [Page 4]

Internet-Draft              BGPsec Overview                June 22, 2016


   direction, for a given address family) then both BGPsec update
   messages (ones that contain the BGPsec_Path_Signature attribute) and
   traditional BGP update messages (that do not contain this attribute)
   can be sent within the session.

   If a BGPsec-capable BGP speaker finds that its peer does not support
   receiving BGPsec update messages, then the BGP speaker must remove
   the BGPsec_Path attribute from any update messages it sends to this
   peer.

3.2. Update signing and validation

   When a BGP speaker originates a BGPsec update message, it creates a
   BGPsec_Path attribute containing a single signature. The signature
   protects the Network Layer Reachability Information (NLRI), the AS
   number of the originating AS, and the AS number of the peer AS to
   which the update message is being sent. Note that the NLRI in a
   BGPsec update message is restricted to contain only a single prefix.

   When a BGP speaker receives a BGPsec update message and wishes to
   propagate the route advertisement contained in the update to an
   external peer, it adds a new signature to the BGPsec_Path attribute.
   This signature protects everything protected by the previous
   signature, plus the AS number of the new peer to which the update
   message is being sent.

   Each BGP speaker also includes a reference, called a Subject Key
   Identifier (SKI).  The SKI identifies the BGPsec Router Certificate
   of the BGP speaker signing the BGPsec_Path attribute.  The SKI is
   used by a recipient to select the public key (and associated router
   certificate data) needed to validate the signature.

   As an example, consider the following case in which an advertisement
   for 192.0.2/24 is originated by AS 1, which sends the route to AS 2,
   which sends it to AS 3, which sends it to AS 4. When AS 4 receives a
   BGPsec update message for this route, it will contain the following
   data:

     *  NLRI: 192.0.2/24
     *  AS path data: 3 2 1
     *  BGPsec_Path contains 3 signatures :
          o  Signature from AS 1 protecting
             192.0.2/24, AS 1 and AS 2
          o  Signature from AS 2 protecting
             Everything AS 1's signature protected, and AS 3
          o  Signature from AS 3 protecting
             Everything AS 2's signature protected, and AS 4

 


Lepinski and Turner    Expires December 24, 2016                [Page 5]

Internet-Draft              BGPsec Overview                June 22, 2016


   When a BGPsec update message is received by a BGPsec speaker, the
   BGPsec speaker can validate the message as follows. For each
   signature, the BGP speaker first determines if there is a valid RPKI
   Router certificate matching the SKI and containing the appropriate AS
   number. (This would typically be done by looking up the SKI in a
   cache of data extracted from valid RPKI objects. A cache allows
   certificate validation to be handled via an asynchronous process,
   which might execute on another device.)

   The BGPsec speaker then verifies the signature using the public key
   from this BGPsec router certificate. If each of the signatures can be
   verified in this fashion, the BGPsec speaker is assured that the
   update message it received was propagated via the AS path specified
   in the update message. 

   In the above example, upon receiving the BGPsec update message, a BGP
   speaker for AS 4 would do the following. First, it would look at the
   SKI for the first signature and see if this corresponds to a valid
   BGPsec Router certificate for AS 1. Next, it would verify the first
   signature using the key found in this valid certificate. Finally, it
   would repeat this process for the second and third signatures,
   checking to see that there are valid BGPsec router certificates for
   AS 2 and AS 3 (respectively) and that the signatures can be verified
   with the keys found in these certificates. Note that the BGPsec
   speaker for AS 4 should additionally perform origin validation as per
   RFC 6483 [RFC6483]. However, such origin validation is independent of
   BGPsec.

   The deployment model for BGPsec requires that all ASs in a BGPsec
   protected path must be BGPsec speakers.  It does not permit BGPsec
   protection of an update that propagates through ASs that do not
   support BGPsec.  In particular, it does not permit what is called
   "partial path signing", in which a BGPsec AS attaches a BGPsec_Path
   attribute to an unprotected update that was received from a
   downstream neighbor.

   Partial path signing might be viewed as supplying information about a
   portion of a path that could be used in making better routing
   decisions, preferring a partially protected route.  However, partial
   path signing implies that the entire AS path is not rigorously
   protected.  Rigorous AS path protection is a key requirement of
   BGPsec [RFC7353].  Partial path signing also introduces the following
   attack vulnerability: If a BGPsec speaker can attach a BGPsec_Path
   attribute to an unprotected update, and if BGPsec protected updates
   would be preferred to unprotected updates, then a BGPsec speaker can
   manufacture any unprotected update it wants and attach a BGPsec_Path
   attribute to it, and thereby increase the chance that its
   manufactured update will be preferred.  Partial path signing then
 


Lepinski and Turner    Expires December 24, 2016                [Page 6]

Internet-Draft              BGPsec Overview                June 22, 2016


   becomes a privilege elevation attack vector, that could be employed
   by any BGPsec AS at any point.

   The need to avoid introducing that vulnerability forced the stringent
   deployment model.

4. Design and Deployment Considerations

   In this section we provide a brief overview of several additional
   topics that commonly arise in the discussion of BGPsec.

4.1. Disclosure of topology information

   A key requirement in the design of BGPsec was that it not disclose
   any new information about BGP peering topology.  Since many ISPs feel
   peering topology data is proprietary, further disclosure of it would
   inhibit BGPsec adoption.

   In particular, the topology information that can be inferred from
   BGPsec update messages is exactly the same as that which can be
   inferred from equivalent (non-BGPsec) BGP update messages.

4.2. BGPsec router assumptions

   In order to achieve its security goals, BGPsec assumes additional
   capabilities in routers. In particular, BGPsec requires adding
   digital signatures to BGP update messages, which will significantly
   increase the size of these messages. Therefore, an AS that wishes to
   receive BGPsec update messages will require additional memory in its
   routers to store (e.g., in ADJ RIBs) the data conveyed in these
   larger update messages. Additionally, the design of BGPsec assumes
   that an AS that elects to receive BGPsec update messages will do some
   cryptographic signature verification at its edge router. This
   verification may require additional capability in these edge routers.

   Additionally, BGPsec requires that all BGPsec speakers support 4-byte
   AS Numbers [RFC6793]. This is because the co-existence strategy for
   4-byte AS numbers and legacy 2-byte AS speakers that gives special
   meaning to AS 23456 is incompatible with the security properties that
   BGPsec seeks to provide.

   For this initial version of BGPsec, optimizations to minimize the
   size of BGPsec updates or the processing required in edge routers
   have not been considered. Such optimizations may be considered in the
   future.

   Note also that the design of BGPsec allows an AS to send BGPsec
   update messages (thus obtaining protection for routes it originates)
 


Lepinski and Turner    Expires December 24, 2016                [Page 7]

Internet-Draft              BGPsec Overview                June 22, 2016


   without receiving BGPsec update messages. An AS that sends, but does
   not receive, BGPsec update messages, will require much less
   capability in its edge routers to deploy BGPsec. In particular, a
   router that only sends BGPsec update messages does not need
   additional memory to store larger updates and requires only minimal
   cryptographic capability (as generating one signature per outgoing
   update requires less computation than verifying multiple signatures
   on each incoming update message). See [I-D.sidr-bgpsec-ops] for
   further discussion related to Edge ASes that do not provide transit.

4.3. BGPsec and consistency of externally visible data

   Finally note that, by design, BGPsec prevents parties that propagate
   route advertisements from including inconsistent or erroneous
   information within the AS-Path (without detection).  In particular,
   this means that any scenarios in which a BGP speaker constructs such
   an inconsistent or erroneous AS Path attribute will break when BGPsec
   is used.

   For example, when BGPsec is not used, it is possible for a single
   autonomous system to have one peering session where it identifies
   itself as AS 111 and a second peering session where it identifies
   itself as AS 222.  In such a case, it might receive route
   advertisements from the first peering session (as AS 111) and then
   add AS 222 (but not AS 111) to the AS-Path and propagate them within
   the second peering session.

   Such behavior may very well be innocent and performed with the
   consent of the legitimate holder of both AS 111 and 222.  However, it
   is indistinguishable from the following man-in-the-middle attack
   performed by a malicious AS 222. First, the malicious AS 222
   impersonates AS 111 in the first peering session (essentially
   stealing a route advertisement intended for AS 111).  The malicious
   AS 222 then inserts itself into the AS path and propagates the update
   to its peers.

   Therefore, when BGPsec is used, such an autonomous system would
   either need to assert a consistent AS number in all external peering
   sessions, or else it would need to add both AS 111 and AS 222 to the
   AS-Path (along with appropriate signatures) for route advertisements
   that it receives from the first peering session and propagates within
   the second peering session. See [I-D.sidr-as-migration] for a
   detailed discussion of how to reasonably manage AS number migrations
   while using BGPsec.

5. Security Considerations

   This document provides an overview of BGPsec; it does not define the
 


Lepinski and Turner    Expires December 24, 2016                [Page 8]

Internet-Draft              BGPsec Overview                June 22, 2016


   BGPsec extension to BGP.  The BGPsec extension is defined in [I-
   D.sidr-bgpsec-protocol].  The threat model for the BGPsec is
   described in [RFC7132].

6. IANA Considerations

   None.

7. References

7.1. Normative References

   [RFC4271] Rekhter, Y., Li, T., and S. Hares, Eds., "A Border Gateway
   Protocol 4 (BGP-4)", RFC 4271, January 2006.

   [RFC6793] Vohra, Q. and E. Chen, "BGP Support for Four-octet AS
   Numbers", RFC 6793, December 2012.

   [RFC5492] Scudder, J. and R. Chandra, "Capabilities Advertisement
   with BGP-4", RFC 5492, February 2009.

   [RFC6480] Lepinski, M. and S. Kent, "An Infrastructure to Support
   Secure Internet Routing", February 2012.

   [RFC6483] Huston, G., and G. Michaelson, "Validation of Route
   Origination using the Resource Certificate PKI and ROAs", February
   2012.

   [RFC7132] Kent, S., and A. Chi, "Threat Model for BGP Path Security",
   RFC 7132, February 2014.

   [RFC7115] Bush, R., "RPKI-Based Origin Validation Operation", RFC
   7115, January 2014.

   [I-D.sidr-bgpsec-protocol] Lepinski, M., Ed., "BPSEC Protocol
   Specification", draft-ietf-sidr-bgpsec-protocol, work-in-progress.

   [I-D.sidr-bgpsec-ops] Bush, R., "BGPsec Operational Considerations",
   draft-ietf-sidr-bgpsec-ops, work-in-progress.

   [I-D.sidr-bgpsec-algs] Turner, S., "BGPsec Algorithms, Key Formats, &
   Signature Formats", draft-ietf-sidr-bgpsec-algs, work-in-progress.

   [I-D.sidr-bgpsec-pki-profiles] Reynolds, M., Turner, S., and S. Kent,
   "A Profile for BGPsec Router Certificates, Certificate Revocation
   Lists, and Certification Requests",
   draft-ietf-sidr-bgpsec-pki-profiles, work-in-progress.

 


Lepinski and Turner    Expires December 24, 2016                [Page 9]

Internet-Draft              BGPsec Overview                June 22, 2016


   [I-D.sidr-as-migration] George, W. and S. Murphy, "BGPsec
   Considerations for AS Migration", draft-ietf-sidr-as-migration,
   work-in-progress.

7.2. Informative References

   [RFC4272] Murphy, S., "BGP Security Vulnerabilities Analysis", RFC
   4272, January 2006

   [I-D.sriram-bgpsec-design-choices] Sriram, K., "BGPsec Design Choices
   and Summary of Supporting Discussions",
   draft-sriram-bgpsec-design-choices, work-in-progress.

   [RFC7353] Bellovin, S., R. Bush, and D. Ward, "Security Requirements
   for BGP Path Validation", RFC 7353, August 2014.

Authors' Addresses

   Matt Lepinski
   New College of Florida
   5800 Bay Shore Road
   Sarasota, FL 34243
   USA

   Email: mlepinski@ncf.edu

   Sean Turner
   sn3rd

   Email: sean@sn3rd.com





















Lepinski and Turner    Expires December 24, 2016               [Page 10]