Configure P & PE on a single router

[Alex Manu]

Hi,

just wanted to know if you can configure a single router to be a P and PE router? If yes, what are the disadvantages of a single router doing both? what is the main reason for separating the two routers on networks? 

thank you in advance

[Philip D'Ath]

Yes you can.  Here is a basic config.

http://www.cisco.com/c/en/us/support/docs/multiprotocol-label-switching-mpls/mpls/13733-mpls-vpn-basic.html

All the P router does is provide connectivity to PE routers.  So you can connect two PE routers directly together if you like.

I recently linked a pair of 890 series routers in exactly this way.

[driss JABBAR]

Hello,

An MPLS network consists of edge devices known as PE (Provider Edge) and core router known as P (Provider).

The basic roles of PE router are:

•    At the ingress side: encapsulate(push) a label and forward the packet toward the egress PE
•    At the egress side: decapsulate (pop) the label and forward the packet to final destination (if PHP is not activated)

The P router reads the incoming label and identify to which egress interface must be sent (swap label).

Back to your questions: you can configure a single router to be a P and PE router? the answer is Yes.

what are the disadvantages of a single router doing both? it depends. How large is your MPLS network, what are the performances of your routers ..etc

what is the main reason for separating the two routers on networks? There are many reasons for that:

• Scalability reason by offloading the role of label swapping to the P router and let the PEs focuses only for the Push and the PoP operations
• Simplifying the design of your MPLS networks 

The answer for why i need a core layer in enterprise network design? could be applied as an answer for why i need a P router in MPLS design 

Hope this helps 

[Alex Manu]

Thank you Philip and driss 

the information provided are very helpful

much appreciated

[Adam Vitkovsky]

There are several reason for separating P and PE functionality

1)            More stable version of code can be run on P routers, while PE routers might need newer code release to support new features

2)            Little to no configuration changes on core routers

3)            Less core links,

- usual POP design is based on two P routers that provide connectivity to other POPs in the backbone

and then each of the PE routers connects to both P routers in a local POP,

since there can be many PE routers with many different roles/functions in a local POP you can see how it would be very expensive to connect them all together and to other PE routers in different POPs

adam

netconsultings.com

::carrier-class solutions for the telecommunications industry::