Network Working Group
Internet Draft
Document: draft-squire-ppvpn-vpn-discovery-reqts-00.txt

Matt Squire
Hatteras Networks

November 2001                                       Expires May 2002


                 VPN Discovery Discussions and Options


Status of this Memo

   This document is an Internet-Draft and is in full conformance with
   all provisions of Section 10 of RFC2026 [1].

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF), its areas, and its working groups. Note that
   other groups may also distribute working documents as Internet
   Drafts. 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."
   The list of current Internet-Drafts can be accessed at
   http://www.ietf.org/ietf/1id-abstracts.txt
   The list of Internet-Draft Shadow Directories can be accessed at
   http://www.ietf.org/shadow.html.

 Abstract

   VPNs are a common service being offered by providers.  The PPVPN WG
   is tasked with defining a limited number of solutions to support the
   interoperable deployment of VPN services.  As part of this effort, a
   design team was tasked with defining the requirements of PPVPN
   discovery proposals, to analyze if additional discovery schemes are
   needed, and to characterize potential solutions to the problem.
   This draft is the output of that design work.  Consensus was not
   reached on the solution characteristics.  However, this draft
   attempts to capture the discussions and positions of the design team
   exchanges.

1  Introduction

   VPNs come in many shapes and sizes.  In [MARTINISIG], as an example,
   an emulated VC consists of two LSPs (Label Switched Paths), one in
   each direction.  Each endpoint initiates the setup of the LSP that
   carries packets in the "incoming" direction.  In order for the
   signaling to proceed, each endpoint has apriori knowledge of
        (a)  the address of the other endpoint, and
        (b)  a VC id.

                                                              [Page 1]


                draft-squire-ppvpn-vpn-discovery-reqts-00.txt  September 2001


   On a given emulated VC, the same VC id must be used for both LSPs.

   In this context, "apriori knowledge" simply means information that
   must be known prior to the initiation of signaling.  The draft
   [MARTINISIG] provides one example of how to use signaling in
   establishing a VPN.  The information required as apriori knowledge
   may differ depending on the signaling protocol and assumptions.

   In the context of PPVPN, discovery is the process by which a PE
   learns the required apriori knowledge.

1.1 Specification of Requirements

   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.

2  Necessity of a Solution

   There are currently two proposals for performing VPN endpoint
   discovery.  The two methods can be characterized as BGP and
   multicast.  RFC 2547 (and related work) define mechanisms to use BGP
   multi-protocol extensions to include VPN information in the BGP
   routing information.  Discovering which PE equipment is in which
   VPNs can then be determined by querying the routing tables.

   RFC 2917 defines how multicast can be used to discover VPN
   membership.  Each VPN is assigned a specific multicast address.
   This address implicitly defines a communication channel to all VPN
   members over which members can determine the unicast address of
   other PE equipment in that VPN.

   Current PPVPN discovery methods have been developed as part of
   particular deployment solutions.  The BGP and multicast approaches
   introduced above were specifically developed for discovery of L3VPNs
   in BGP or multicast enabled networks.  Most members of the design
   team believe that there are contexts where additional discovery
   mechanisms are needed.  Some members are firmly opposed to this
   belief.

   PPVPN discovery is only one part of the PPVPN solution.  As there
   are currently multiple models and architectures for PPVPN solutions,
   including virtual routers, overlay L3 networks, Layer2 VPNS, etc.,
   one must consider particular discovery solutions in the context of
   the PPVPN architectures for which they are intended.

3  Requirements

   There was consensus from the design team on many requirements for a
   VPN discovery protocol.  However, there was contention over the
   exact interpretation of the requirements.  This following list
   summarizes the requirements while later subsections discuss where
   these requirements are in dispute.

                                                              [Page 2]


                draft-squire-ppvpn-vpn-discovery-reqts-00.txt  September 2001



   Any VPN discovery process or protocol must satisfy the following
   requirements.

      - It MUST support inter-provider VPNs.
      - It MUST  be possible  to deploy  the auto-discovery scheme in a
        manner which prevents unauthorized access and allows
        authentication of the source
      - It MUST respond to VPN membership changes in a timely fashion
        (see 3.1)
      - It SHOULD limit VPN information to only those PEs involved in
        that VPN.
      - It MUST provide VPN endpoint information consisting of at last
        the IP address of associated PE equipment, and MAY be
        extendible to provide additional information (see 3.2).


3.1 Timely Fasion

   The responsiveness of VPN discovery to membership changes is hotly
   contested.  Although faster is always better, the main question is
   how fast is fast enough?  When a PE is added to a VPN (or deleted
   for that matter), how quickly must the other PE equipment in that
   VPN notice the change?

   The two positions put forward by the design team and captured by
   this document are rougly:
   a) "Routing" time frame.
   b) "Provisioning" time frame.

   This document does not attempt to exactly quantify the two
   possibilities.  However, it should be clear that (a) is a more
   stringent requirement than (b).  As a rough quantification, (a) is
   usually measured in seconds, while (b) is measured in minutes.

3.2 Extended Discovery or Signaling

   There are two different ways of viewing the VPN discovery process.
   It could be viewed as a limited process where each member learns
   enough about other members of the VPN to complete VPN signaling.
   Alternatively it could be viewed as this limited process, plus the
   VPN signaling.

   The VPN signaling is required to exchange parameters of a more or
   less dynamic nature, e.g. information used to dynamically distribute
   MPLS labels set up LSP tunnels and VC LSP's.

   The point of contention centers on the boundary between discovery
   and signaling.  It is clear that, at a minimum, each PE device must
   know the IP address of other PE devices that serve a VPN.  The
   question then becomes whether the IP address enough information.
   The two conflicting positions are

                                                              [Page 3]


                draft-squire-ppvpn-vpn-discovery-reqts-00.txt  September 2001


     a) Yes.  All additional information should be exchanged via the
        signaling protocol.
     b) No.  It should be required to support additional information
        that   may be a prequesite for signaling.
   It is clear that the initialization process must be extensible to
   new parameters and features, the question lies in where those
   parameters are added.

4  Conclusion

   Being an IETF design team, we have realized our responsibility to
   not reach consensus.

   The design team was able to reach consensus on some of the VPN
   discovery requirements.  These are described in Section 3.  However,
   there was intense contention on several key issues as described
   outlined in Sections 3.1 thru 3.2.  As a result, we can conclude
   that any one PPVPN discovery solution is unlikely to satisfy all
   providers, developers, and scenarios.

5  Referenc

   [RFC2026] S. Bradner, "The Internet Standards Process -- Revision
      3", BCP 9, RFC 2026, October 1996.

   [RFC2547] E. Rosen & Y. Rekhter, "BGP/MPLS VPNs," RFC 2547, March
      1999.

   [RFC2917] K. Muthukrishnan & A. Malis, "A Core IP MPLS VPN
      Architecture," RFC 2917, September 2000.

   [MARTINISIG] L. Martini, et al., "Transport of Layer 2 Frames over
      MPLS," draft-martini-l2circuit-trans-mpls-08.txt, Work in
      Progress.


6  Acknowledgments

   This draft is the result of collaborative effort of the PPVPN
   discovery design team.  The members of that design team are:

        Loa Andersson (Utfors)
        Ron Bonica (MCI)
        Juha Heinanen (Song Networks)
        Jim Luciani (Crescent Networks)
        Dave McDysan (WorldCom)
        Dave Meyer (Sprint)
        Hamid Ould-Brahim (Nortel Networks)
        Yakov Rekhter (Juniper Networks)
        Eric Rosen (Cisco)
        Tissa Senevirathne (Force 10 Networks)
        Matt Squire (Hatteras Networks)


                                                              [Page 4]


             draft-squire-ppvpn-vpn-discovery-reqts-00.txt September 2001


   All members made valuable contributions to this effort.

7  Author's Addresses

   Matt Squire
   Hatteras Networks
   639 Davis Drive
   Research Triangle Park, NC 27709
   Email: msquire@hatterasnetworks.com













































                                                              [Page 5]