The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2002-Apr> msg00184



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

PHP

  • From: Eric Rosen <erosen@cisco.com>
  • Date: Mon, 22 Apr 2002 10:28:31 -0400
  • cc: Markus Jork <mjork@avici.com>, Vach Kompella <vkompella@timetra.com>, David Charlap <David.Charlap@marconi.com>, mpls@UU.NET
  • User-Agent: EMH/1.14.1 SEMI/1.14.3 (Ushinoya) FLIM/1.14.3(Unebigoryōmae) APEL/10.3 Emacs/21.2(sparc-sun-solaris2.6) MULE/5.0 (SAKAKI)


Kireeti> One could use  IP LSPs for IP, IP VPNs, etc.,  where the payload is
Kireeti> IP; and establish a parallel set of LSPs for non-IP traffic. 

I think the impact on the core makes this one a non-starter. 

Markus> Maybe it should  be redefined to only apply to  packets with a label
Markus> stack of 1 while  all other packets on the LSP with  label stack > 1
Markus> are simply of unknown protocol type. 

Isn't this really the de facto standard interpretation?

Eric> So in practice,  PHP always results in either an IP  packet or an MPLS
Eric> packet, and the appropriate ethertype (IP or MPLS) can then be applied
Eric> before the packet is transmitted to its next hop. 

Kireeti> How does one distinguish IPv4 from IPv6? 

Kireeti> One could use the 'check first  nibble' hack :-( It would be neater
Kireeti> to use the L3PID. 

Well, I wouldn't want to have to  set up a separate TE tunnel just to handle
the IPv6 traffic, so I'd recommend the "check first nibble" hack.

  • Follow-Ups:
    • PHP
      • From: Eric Rosen <erosen@cisco.com>
  • References:
    • PHP
      • From: Kireeti Kompella <kireeti@juniper.net>