Saturday 8 July 2017

BGP/MPLS

PUSH, SWAP, POP functions in MPLS and also explain penultimate hop popping (PHP).
A. Label switching is a discrete mechanism from IP forwarding; here LSR regulates what label function needs to be performed, such as Push, Pop, or Swap, to process a packet throughout the MPLS backbone using label switching.
Push: Add a label
Pop: Remove a label
Swap: Exchange/ replace a label

Penultimate hop popping is a technique to remove the label one hop before its destination so that untagged packet will be received by the destination PE router and only IP lookup will be performed by destination PE router to forward it to the directly connected CE interface.

 How can you differentiate VPNv4 and IPv4 address-family?
Address-family VPNv4 is used to transport VRF traffic, which is quite heavier (96 bits) than normal IPv4 packets, and address-family IPv4 is used to transport traditional 32-bit IPv4 packets. When we deploy L3VPN, we need VPNv4 peering between the PEs to take the VRF routes (labeled with RD, RT) to the corresponding destination PE router.

What are LIB and LFIB? 
The difference between two is almost like RIB and FIB, as RIB uses FIB to forward traffic. Label Information Base (LIB) is a table where prefix to label bindings are built and stored in the control plane and can be seen with “show mpls ldp bindings.” Label Forwarding Information Base (LFIB) is another MPLS table used to forward label packets throughout the MPLS network and build in the data plane. To check LFIB, run “show mpls forwarding-table.


If my BGP neighbor is stuck in idle or active state, what should I do?
If BGP peer is in idle state, then it could be due to physical connectivity failure or the neighbor is not defined properly with respective AS. In connect state, BGP tries to establish a TCP session over port number 179; if it fails to establish the connection, then it goes to active state, where it tries again to establish a TCP connection. By using “debug ip bgp” and “debug ip tcp transactions” commands, we can learn the exact cause of TCP connection failure.

What do you understand by BGP split-horizon rule?
This rule states, “Network prefixes learned from an iBGP peer will not be shared with another iBGP peer” to prevent routing loops. BGP uses the AS_PATH attribute to avoid routing loops, which means a route will not be accepted if the AS_PATH contains the same AS number of the router where it is receiving. Route-reflector and confederation techniques are used to share iBGP learned networks with another iBGP neighbor.

 

No comments:

Post a Comment