The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2002-Feb> msg00021



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

[Fwd: Penultimate Hop layer-2 EtherType setting for MPLS]

  • From: Tricci So <tso@caspiannetworks.com>
  • Date: Mon, 04 Feb 2002 13:21:27 -0800
  • Organization: Caspian Networks

Oops... I should reply to all......


  • From: Tricci So <tso@caspiannetworks.com>
  • Date: Mon, 04 Feb 2002 12:33:35 -0800
  • Organization: Caspian Networks
Hi Jason,

Where is 0x300 come from??  Do you mean 0x0003 which is the MPLS reserved label represents the "implicit null" and is used to indicate the
PHOP label popping?   As you saw in my previous email title, my question is regarding the Ethertype encoding for the link layer (i.e.
Ethernet or POS).   From the protocol trace, it implies that Cisco sets the Ethertype to "0x8847" (as MPLS) even though it spilts out IP
packet to the egress LER.

Comments?????  Thanks....
Tricci

Jason Strayer wrote:

> Traci,
>         If you are asking what the code is for PHOP, it is 0x300.  That tells the MPLS device to pop (N-1) labels.  Then, if we are not
> talking MPLS VPN, you would have L3 operation from here.  Is that the answer you were looking for here ?
>
> /Jason
>
> -----Original Message-----
> From: owner-mpls@UU.NET [mailto:owner-mpls@UU.NET]On Behalf Of Tricci So
> Sent: Monday, February 04, 2002 11:00 AM
> Cc: mpls@UU.NET
> Subject: Penultimate Hop layer-2 EtherType setting for MPLS
>
> Hi all,
>
> Could somebody please comment on the following email regarding if this is a
> correct interpretation of the spec for PHOP to forward IP packet?  Obviously,
> the same question applies for POS as well.
>
> Many thanks in advance........
> Tricci
>
> =====> Forwarded Message.....
>
> mpls over ethernet
>
>      From: "Zhao, Xingguo" <xzhao@celoxnetworks.com>
>      Date: Wed, 14 Nov 2001 17:03:40 -0600
>
> Hi,
>
> Can Ethernet frame (EtherType = "0x8847", MPLS) carry an unlabelled IP packet?
> This looks like illegal.
>
> But we observed Cisco 12000 actually sent out such a frame.
>
> Thanks!