[c-nsp] EoMPLS between 7600 & 7200 config clarification

Justin Shore justin at justinshore.com
Tue Feb 5 16:45:38 EST 2008


Gert Doering wrote:
> Hi,
> 
> On Tue, Feb 05, 2008 at 02:15:40PM -0600, Justin Shore wrote:
>> I'm in a similar boat as Jose.  What options for EoMPLS do we people 
>> with 6700s have?  I'm trying physical to physical with no luck. 
> 
> physical to physical should work, according to the documentation.  I
> haven't tried it yet, though.
> 
> What IOS version did you try?

I'm running SRB on 3BXLs.  Here's what I get when I try it on the physicals:

7613-1.clr#sh xconnect all
Legend: XC ST=Xconnect State, S1=Segment1 State, S2=Segment2 State
UP=Up, DN=Down, AD=Admin Down, IA=Inactive, NH=No Hardware
XC ST  Segment 1                         S1 Segment 2 
       S2
------+---------------------------------+--+---------------------------------+--
DN     ac   Gi9/28(Ethernet)             UP mpls 10.64.0.20:1234 
       DN


7613-2.clr#sh xcon
7613-2.clr#sh xconnect all
Legend: XC ST=Xconnect State, S1=Segment1 State, S2=Segment2 State
UP=Up, DN=Down, AD=Admin Down, IA=Inactive, NH=No Hardware
XC ST  Segment 1                         S1 Segment 2 
       S2
------+---------------------------------+--+---------------------------------+--
DN     ac   Gi9/28(Ethernet)             UP mpls 10.64.0.10:1234 
       DN

I got the same thing when trying SVIs.

>> Sub-interface isn't an option for a particular design that I'm working 
>> on either.
> 
> Sub-if is what we use today (because the other endpoint is a 7200 that can 
> *only* do sub-if in 12.2S).  Works great :)

I can arrange for a sub-int test later this week.

I would love to own ES20s if the price was more reasonable.  I can't 
justify buying the ES20 at 9x the cost ($60k + $40k +16k / 20 vs $15k + 
$15k / 48).  That doesn't even include the SFPs.

I'd like to see Cisco apply the "leather seat" theory to the ES blades. 
  The theory states that if all car manufacturers switched to offering 
leather-trimmed seats by default (which the majority want) and charging 
extra for cloth seats (which the majority don't want) that the higher 
price of leather would shortly be offset by volume needed and would 
ultimately bring the price down to be on par with what the cloth seats 
previously cost.  Applying this theory to the ES linecards, Cisco should 
make all Ethernet-based linecards with the ES capabilities, thus driving 
down the overall cost for linecards with the ES-capabilities and ending 
countless amounts of frustration for us engineers. :-)

Justin



More information about the cisco-nsp mailing list