[c-nsp] 7600: eompls on Vlan ?

Ian Cox icox at cisco.com
Tue Feb 14 14:15:23 EST 2006 

At 06:28 PM 2/14/2006 +0000, David Freedman wrote:
>Ian Cox wrote:
> > At 05:18 PM 2/14/2006 +0000, David Freedman wrote:
><snip>
> >
> >
> > You guess incorrectly. The Supervisor L2 switches the packet to the
> > OSM/FlexWAN/SIP module that has joined the specific vlan, and the
> > OSM/FlexWAN/SIP receives the L2 packet and adds the MPLS labels. Similar
> > process happens when bridge packets on ATM interfaces arrive and are
> > switched out onto ethernet interfaces.
>
>Yes, I believe that this is what I described, but in reverse.
>Either way, the OSM does the MPLS and the Sup does the Layer 2.
>
> >
> >> What is the point? is it so hard to do both on the Sup in a two stage
> >> (allbeit slow) process?
> >
> >
> > It is impossible to with the current Sup720 Forwarding ASICs.
>
>So, what you are saying is that this functionality is not in the
>silicon? why not?  this appears to be a design oversight.

Unfortunately it was not a design oversight, it was deliberate. The 
customer EoMPLS was developed for prime requirement was  not to learn 
MAC addresses, they wanted to provide ethernet pipes between two 
locations just like what is done with ATM and Frame Relay, 1q tags 
being the DLCI/vcs. We delivered them a solution with L2 learning, 
and were unhappy with it because they did not want to deal with MAC 
address tables, so that input was taken into the design of the PFC3x 
when it was done.


> >
> >> Are there any plans to add software to support this? and if so when?
> >
> >
> > No.
>
>Thanks, if thats the case, you should at the very least offer an option
>for doing this in software with reduced performance as a result, if not,
>I don't see a use for this architecture moving forward with our current
>EOMPLS plans.

The solution is if you need L2 switching is to utilize 
FlexWAN/OSMs/SIPs for the uplinks (it done in software or microcode 
on the card).


Ian


>Kind Regards,
>
>David Freedman
>
> >
> >
> > Ian
> >
> >> Thanks,
> >>
> >> Dave.
> >>
> >>
> >>
> >> Kristofer Sigurdsson wrote:
> >> > David,
> >> >
> >> > The Sup2 can't do MPLS on the Sup - it has to use the line cards, ie.
> >> > OSM/FlexWAN etc.
> >> >
> >> > The Sup720 (vanila) is the same.  The Sup720-3B and -3BXL can do MPLS,
> >> > but they have this restriction.
> >> >
> >> > -Kristo
> >> >
> >> > 2006/2/14, David Freedman <david.freedman at uk.clara.net>:
> >> >> I'm right in thinking that this restriction only applies to Sup2
> >> >> architecture and not Sup720, surely the 720 should be able to do SVI
> >> >> EOMPLS without the need for the PE->P link to be on a WAN module?
> >> >>
> >> >> Dave.
> >> >>
> >> >> Bill Wade (wwade) wrote:
> >> >> > SVI based EoMPLS --
> >> >> >
> >> >> >   interface Vlan501
> >> >> >    xconnect 192.1.1.1 501 encapsulation mpls
> >> >> >
> >> >> > requires an OSM+, FlexWAN, enhanced FlexWAN or SIP-600 as MPLS core
> >> >> > facing interface.
> >> >> >
> >> >> > Bill
> >> >> >
> >> >> >> -----Original Message-----
> >> >> >> From: cisco-nsp-bounces at puck.nether.net
> >> >> >> [mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of
> >> >> >> Alexandre Snarskii
> >> >> >> Sent: Tuesday, February 14, 2006 9:43 AM
> >> >> >> To: Cisco-NSP Mailing List
> >> >> >> Subject: [c-nsp] 7600: eompls on Vlan ?
> >> >> >>
> >> >> >>
> >> >> >> Hi!
> >> >> >>
> >> >> >> As my tests show me, EoMPLS VC configured on interface Vlan
> >> >> >> does not come up (and debug shows "Imposition setup
> >> >> >> incomplete because unable to find a tunnel label" error), but
> >> >> >> the same xconnect config applied to subinterface works well,
> >> >> >> vc comes up and passes traffic.
> >> >> >>
> >> >> >> IOS version: 12.2(18)SXD7, also tested with 12.2(18)SXF2, ip
> >> >> >> services.
> >> >> >> Testbed network topology: c2950 <-> Gi4/4 Sup720 Gi4/12 <->
> >> >> >> GE3/1 Sup2/OSM
> >> >> >>
> >> >> >> Is this a bug or some kind of feature ? If a bug: is there an
> >> >> >> IOS version unaffected by this bug, or when it planned to be
> >> >> >> fixed ? If a feature - are there any hints to get eompls
> >> >> >> working together with switching without hardware upgrade ?
> >> >> >> (box should stay as simple as chassis+sup+ WS-X6516-GBIC combo).
> >> >> >>
> >> >> >> Full log of "failed" connection:
> >> >> >>
> >> >> >> Feb 13 13:45:54.451 MSK: AToM SMGR [RE.MO.TE.IP, 900]:
> >> >> >> Adjacency creation succeeded for Vlan900.
> >> >> >> Feb 13 13:45:54.515 MSK: DISP XDR cktidx: 196 Feb 13
> >> >> >> 13:45:54.515 MSK: AToM SMGR [RE.MO.TE.IP, 900]: Processing
> >> >> >> imposition update, vc_handle 50D68200, update_action 1,
> >> >> >> remote_vc_label 731 Feb 13 13:45:54.515 MSK: CWAN_ATOM:
> >> >> >> AtomSmgr Set parentRew, vcID: 900 Feb 13 13:45:54.515 MSK:
> >> >> >> CWAN_ATOM: GetParentRew failed, vcId: 900, rewCnt: 0 Feb 13
> >> >> >> 13:45:54.515 MSK: CWAN_ATOM: lb-mode: default, vcId: 900,
> >> >> >> rew: n/a Feb 13 13:45:54.515 MSK: CWAN_ATOM: find parentRew
> >> >> >> for vcId: 900 failed Feb 13 13:45:54.515 MSK: AToM SMGR
> >> >> >> [RE.MO.TE.IP, 900]: no parent rewrite: LFIB entry not present
> >> >> >> or unresolved Feb 13 13:45:54.515 MSK: AToM SMGR
> >> >> >> [RE.MO.TE.IP, 900]: Imposition setup incomplete because
> >> >> >> unable to find a tunnel label Feb 13 13:45:54.515 MSK: AToM
> >> >> >> SMGR: Processing TFIB event for RE.MO.TE.IP Feb 13
> >> >> >> 13:45:54.515 MSK: CWAN_ATOM: AtomSmgr Set parentRew, vcID:
> >> >> >> 900 Feb 13 13:45:54.515 MSK: CWAN_ATOM: GetParentRew failed,
> >> >> >> vcId: 900, rewCnt: 0 Feb 13 13:45:54.515 MSK: CWAN_ATOM:
> >> >> >> lb-mode: default, vcId: 900, rew: n/a Feb 13 13:45:54.515
> >> >> >> MSK: CWAN_ATOM: find parentRew for vcId: 900 failed Feb 13
> >> >> >> 13:45:54.515 MSK: AToM SMGR [RE.MO.TE.IP, 900]: no parent
> >> >> >> rewrite: LFIB entry not present or unresolved Feb 13
> >> >> >> 13:45:54.515 MSK: AToM SMGR [RE.MO.TE.IP, 900]: Imposition
> >> >> >> setup incomplete because unable to find a tunnel label
> >> >> >>
> >> >> >> The same log, using subinterface with the same vlan configured:
> >> >> >>
> >> >> >> Feb 13 14:25:48.745 MSK: AToM SMGR [RE.MO.TE.IP, 900]:
> >> >> >> Processing disposition update, vc_handle 50D68200,
> >> >> >> update_action 1, local_vc_label 489 Feb 13 14:25:48.745 MSK:
> >> >> >> AToM SMGR [RE.MO.TE.IP, 900]: Adjacency creation succeeded
> >> >> >> for GigabitEthernet4/4.900.
> >> >> >> Feb 13 14:25:48.805 MSK: AToM SMGR [RE.MO.TE.IP, 900]:
> >> >> >> Processing imposition update, vc_handle 50D68200,
> >> >> >> update_action 1, remote_vc_label 731 Feb 13 14:25:48.805 MSK:
> >> >> >> CWAN_ATOM_IMP: XDR HW l2ckt: 900, ACtype: 4, local_vc_label:
> >> >> >> 489 Feb 13 14:25:48.805 MSK: AToM SMGR [RE.MO.TE.IP, 900]:
> >> >> >> Imposition Programmed, Output Interface: Gi4/12, flags 0x0011
> >> >> >> Feb 13 14:25:48.805 MSK: AToM SMGR: Processing TFIB event for
> >> >> >> RE.MO.TE.IP Feb 13 14:25:48.805 MSK: CWAN_ATOM_IMP: XDR HW
> >> >> >> l2ckt: 900, ACtype: 4, local_vc_label: 489 Feb 13
> >> >> >> 14:25:48.805 MSK: AToM SMGR [RE.MO.TE.IP, 900]: Imposition
> >> >> >> Programmed, Output Interface: Gi4/12, flags 0x0011
> >> >> >>
> >> >> >>
> >> >> >> _______________________________________________
> >> >> >> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> >> >> >> https://puck.nether.net/mailman/listinfo/cisco-nsp
> >> >> >> archive at http://puck.nether.net/pipermail/cisco-nsp/
> >> >> >>
> >> >> >
> >> >> > _______________________________________________
> >> >> > cisco-nsp mailing list  cisco-nsp at puck.nether.net
> >> >> > https://puck.nether.net/mailman/listinfo/cisco-nsp
> >> >> > archive at http://puck.nether.net/pipermail/cisco-nsp/
> >> >> >
> >> >>
> >> >> _______________________________________________
> >> >> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> >> >> https://puck.nether.net/mailman/listinfo/cisco-nsp
> >> >> archive at http://puck.nether.net/pipermail/cisco-nsp/
> >> >>
> >> >
> >> > _______________________________________________
> >> > cisco-nsp mailing list  cisco-nsp at puck.nether.net
> >> > https://puck.nether.net/mailman/listinfo/cisco-nsp
> >> > archive at http://puck.nether.net/pipermail/cisco-nsp/
> >> >
> >>
> >> _______________________________________________
> >> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> >> https://puck.nether.net/mailman/listinfo/cisco-nsp
> >> archive at http://puck.nether.net/pipermail/cisco-nsp/
>
>_______________________________________________
>cisco-nsp mailing list  cisco-nsp at puck.nether.net
>https://puck.nether.net/mailman/listinfo/cisco-nsp
>archive at http://puck.nether.net/pipermail/cisco-nsp/

More information about the cisco-nsp mailing list