i am trying to learn mpls TE  , all goes greate exept of 1 command :

conf t->if tunnel1->tunnel mpls traffic-eng autoroute

 i read alot about it , most of post explain this as "add the tunnel to IGP calculation" etc .... things that not fully explain the why!


so my question is why would traffic will go over the tunnel when i issue this command ?




  • in a way what you have done with that command is created a 1 hop IGP route from PE to [P|PE].  that route is then passed down to your FIB -> LIB and is seen as the best path.

  • peetypeety ✭✭✭

    MPLS TE tunnels are "special". Because they're unidirectional, IGP adjacencies aren't possible (and that's probably a good thing). Because the IGP doesn't form adjacencies, you are responsible for putting traffic onto the tunnels some other way. My "Traffic Engineering with MPLS" book (by Eric Osborne, awesome book for TE and also for MPLS in general) is tied up in storage at the moment, but I vaguely recall three ways to put traffic over tunnels:

    1) Static routes

    2) PBR

    3) autoroute

    Static and PBR should be obvious in how they work. Autoroute is a little different, in that it hooks into the IGP to not disturb the adjacencies/metrics/whatever, but to substitute the tunnel in place of the IGP-calculated next-hop where the IGP agrees that the tunnel is equal/better to the IGP's best route. In doing so, remember that it's a substitution, not an addition, so the route(s) that it replaces are no longer presented in the RIB, but the TE tunnel shows up in replacement.

    Remember that TE will not load-share between a tunnel and an alternate physical path TO a destination, but it will load-share between a tunnel and an alternate physical path BEYOND a destination.

  • petty = i did understand ur answer

    here is my topology  in the link .....!xJFlDCjQ!MGklOeS9s01mkUNcHczqlDSXuvXfeWT-uHW6gKn8plY

     my TE is the green line ......
    static explicit path ....
    and the cost in this path is higher than the regular apth ....
    TE path : R1->r3->r5->r4->r2 ....
     regular path : R1->R3->R2

    why ? what is the spesific reason that traffic beyon R2 will start going via the tunnel .... what wil link this toutes to the tunnel
    what is the spesific thing that does the magic ?


    BTW .... appriciate ur affort to help - )


  • During SPF each new node found is moved from a TENTative list to PATHS list. Now the first-hop is being determined via:
        A.  Check if there is any TE tunnel terminating at  this node from the current router and if so do  the metric check
        B.  If there is no TE tunnel and the node is directly    connected use the first-hop from adj database
        C.  In non of the above applies the first-hop is   copied from the parent of this new node.

    Tunnel metric:
     A.  Relative +/- X
     B.  Absolute Y
       The default is relative metric of 0.

    Metric of native IP path to the found node =  50
    1. Tunnel with relative metric of  -10 =>   40
    2. Tunnel with relative metric of  +10 =>   60
    3. Tunnel with absolute metric of 10 =>  10

    If the metric of the found TE tunnel at this node is higher then the metric for other tunnels or native IGP path this tunnel is not installed as next hop
    If the metric of the found TE tunnel is equal to other TE tunnels the tunnel is added to the existing next-hops
    If the metric of the found TE tunnel is lower then the metric of other TE tunnels or native IGP the tunnel replaces them as the only next-hop.

  • peetypeety ✭✭✭

    I think we're getting caught up a bit in semantics.

    'tunnel mpls traffic-eng autoroute' causes the TE tunnel to be made available to the IGP for routing calculations. That doesn't mean that the tunnel will absolutely be used as the route of choice...if the metric is higher than an existing IGP route, the existing IGP route remains selected. However, I think the tunnel inherits (by default) the metric of the best route, not the actual LSP, so by default it would be the best path and would be selected because of that reason.

  • i agree, that was just the best one source version of the process i could find via a nanog ppt file.  :)

Sign In or Register to comment.