The MPLS WG Archive

["Chase, Christopher J (Chris), ALNTK" <chase@att.com>]

Indeed.  The VPN definition in terms of sites is ill-defined and has led to
many confusions.  

As for the service we have developed and sell, we define a VPN as a
collection of interfaces.  An interface is always associated with a single
vrf belonging to the VPN.  Although, I must admit that inevitably in
discussions with customers, we often use the term site instead of interface.

Chris Chase


> -----Original Message-----
> From: Ben Black [mailto:ben@layer8.net]
> Sent: Wednesday, December 13, 2000 7:53 PM
> To: Andrew Wu
> Cc: 'James_Huang@Mitel.COM'; mpls@UU.NET; nbvpn@bbo.com
> Subject: Re: How to map a packet to a VRF for route lookup?
> 
> 
> Perhaps James is drawing attention to the ambiguity of the first
> section stating that VRFs are per _site_, but the second section
> stating that VRFs are per _interface_ (allowing for multiple
> connections from the same site to be treated independently).
> 
> 
> ben
> 
> On Wed, Dec 13, 2000 at 04:48:28PM -0800, Andrew Wu wrote:
> > Here is my quick take on this:
> > 
> > Typically, one (sub)interface on a PE is connected 
> > to one CE device and that the (sub)interface at same 
> > time is associated with one VRF on the PE. So the 
> > packets coming from the (sub)interface will 
> > be forwarded upon the result of the lookup in that 
> > VRF(the (sub)interface is associated with ).
> > 
> > On a PE:
> > ===========================
> > (sub)interface1----> VRF1
> > 
> > 
> > (sub)interface2 ----> VRF2
> >            
> >                        ^
> >                        |
> >                        |
> >                    (sub)interface2
> > 
> > 
> > -andrew
> > 
> > -----Original Message-----
> > From: James_Huang@Mitel.COM [mailto:James_Huang@Mitel.COM]
> > Sent: Wednesday, December 13, 2000 4:22 PM
> > To: mpls@UU.NET
> > Subject: How to map a packet to a VRF for route lookup?
> > 
> > 
> > Hi all,
> >      I am very confused by the description in RFC2547bis 
> with regarding to
> > the
> > association of VRFs and CE sites.  In section 1.3 of RFC2547bis, the
> > following
> > description is given:
> >      "Each PE router maintains a number of separate 
> forwarding tables.
> >      Every site to which the PE is attached must be mapped 
> to one of those
> >      forwarding tables."
> > 
> >      Also in section 3, the following text is given:
> >      Each PE router maintains one or more "per-site 
> forwarding tables."
> >      These are known as VRFs, or "VPN Routing and 
> Forwarding" tables.
> >      Every site to which the PE router is attached is 
> associated with one
> >      of these tables.  A particular packet's IP destination 
> address is
> >      looked up in a particular VRF only if that packet has arrived
> >      directly from a site which is associated with that table.
> > 
> >      From these descriptions,  one would conclude that a CE site is
> > associated
> > with exactly one VRF.  But the description in another 
> paragraph of section
> > 1.3
> > seems to indicate otherwise:
> >      A PE router is attached to a site by virtue of being 
> the endpoint of
> >      an interface or "sub-interface" (PVC, VLAN, GRE 
> tunnel, etc.) whose
> >      other endpoint is a CE device.  If there are multiple 
> attachments
> >      between a site and a PE router, all the attachments 
> may be mapped to
> >      the same forwarding table, or different attachments 
> may be mapped to
> >      different forwarding tables.  When a PE router 
> receives a packet from
> >      a CE device, it knows the interface or sub-interface 
> over which the
> >      packet arrived, and this determines the forwarding 
> table used for
> >      processing that packet.  The choice of forwarding table is NOT
> >      determined by the user content of the packet.
> > 
> >      The above description seems to associate an interface 
> or subinterface
> > with
> > a VRF.
> >      Am I missing somethin here?
> > 
> > 
> > -- James Huang
> > 
> 
> -- 
> what great thing would you attempt if you knew you could not fail?
>