The MPLS WG Archive

[Eric Rosen <erosen@cisco.com>]

> If it turns out the WG consensus is to remove the FEC, then Ina will have 
> to restore the two changes she removed, including changing the semantics of 
> the Prefix Address FEC to provide equivalent functionality. 

We  really  need  to get  clear  on  the  technical  issue here.   The  ONLY
difference  between the Host  Address FEC  and an  Address Prefix  FEC which
specifies a /32  is the following.  If these two FECs  are both present, and
each is bound to a different LSP, then a packet whose IP destination address
matches both of these FECs will be assigned to the LSP with the Host Address
FEC.  Further,  a packet whose IP  destination address does  not match these
FECS cannot be assigned to the LSP  with the Host Address FEC, though it can
be assigned to the  LSP with the Address Prefix FEC (i.e.,  the LSP with the
Host Address FEC can only be used  for packets which are destined to the LSP
egress, not for packets which are to be further forwarded by the LSP egress).

Hence it simply does not make any  sense to say that if the Host Address FEC
is  eliminated,  the Address  Prefix  FEC needs  to  be  changed to  provide
equivalent functionality!   The reason for eliminating the  Host Address FEC
is to eliminate the associated functionality, because that functionality has
never been used.

It's  already been  stated that  the  MPLS Forum  work could  have used  the
Address Prefix FEC  with a /32 prefix.   If that's the case, then  it is not
using the Host Address FEC correctly. 





_______________________________________________
mpls mailing list
mpls@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/mpls