Lable allocation mode per interface and per platform mode



I am a begineer in MPLS and have been preparing for my CCIE lab, my friend happened to raise a questions which raised my curiosity and wish somebody could clarify.

Out of this label distribution mode per platform and per interface mode, the per platfor mode where one single label is assigned to a destination network and announced to all neighbors and valid on all local interfaces. Whereas Per-interface label allocation mode where local labels are assigned to IP destination prefixes on a per-interface basis. These labels must be unique on a per-interface basis.

The Advantage of using per platform is quicker label exchange and lesser label usage, and DisAdvantage as label spoofing where adjacent router can send a labeled packet with a label that has not been previously advertised to this router. 


So is that why per interface used to avoid spoofing or is there any other usage of per interface allocation in frame mode


  • Per platform does bring the added security risk with it but there are ways of mitigating against intrusion that a network should be set up for to prevent unauthorised use. I guess the main benefit is the lower lib size that comes with platform based, depending on your networks size you could end up with a monstrous table. That really should be a thing of the past though with scalability of modern routers, able to handle hundreds of thousands or even millions of labels. Plus you have filtering mechanisms on a cisco to lower the lfib size, alu and possibly juniper do this by default, way more efficient.

    Per interface just smacks of RPF logic, must match interface, and everyone loves RPF right?!?  Don't forget the human readability element, hard for people to sift through X more labels when troubleshooting.

    I wonder if there is an element of 'we run per platform on prob every vendors kit, lets leave it that way.'

Sign In or Register to comment.