[nsp] Delay in MPLS PE lookup

[Eric Osborne]

On 8 Jun 2004 12:03:48 -0000, khiz <khiznms at rediffmail.com> wrote:

>   
> Hi Gurus,
>   I have implemented L3 MPLS VPN  in the foll  test scenario
>
> CE-1 ---[PE-1]--|FR Switch|------WAN--------- |FR  
> Switch|----[PE-2]--[CE-2]
>
> Everything is working absolutely fine as far as  the MPLS L3 VPN  
> functionalty is concerned.
> My Concern is of the delay invloved in pinging from CE-1 to Ce-2
> PE1 and Pe-2 are talking  MP BGP . There is no intermediate P routers  
> for my network
>
> Ping response from
> 1> CE 1 to CE2 (with CE -1 directly connecting to CE-2 over the L2 Frame  
> relay backbone using FR encapsulation) =  20 ms
> 2>Ce-1 to CE-2 (terminating  both the CE routers on the corresponding PE  
> router) =  50 ms via the MPLS VPN
>
> Is this delay expected when moving from a layer 2 connectivty to an  
> MPLS  connectivity given the fact that the PEs have to do L3 lookup and  
> further label association


It's really just because you have an extra routed hop in there.  With FR  
direct, it's

CE1---CE2

with the PEs, it's

CE1---PE1---PE2---CE2

where PE1 has to do a routing lookup (and impose a label), and PE2 has to  
pop a label and do a routing lookup.  Label operations are pretty much the  
same speed as routing lookups, so I'd expect to see the same delay whether  
it was MPLS-VPN or just regular IP between all 4 routers.



eric

>
> ( PE 1 and PE2 still connect to each other via a FR PVC)
>
> PE 1  and PE2 are Cisco 7513
>
> Cheers
> Knms
>
>
>
>
>
>
> _______________________________________________
> 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/