Hi @asit1111990
Let me try to contribute with your learning.
What makes Fabricpath interesting and a bit complex is that it scrambles concepts of layer2 technology with layer3 technology. So, to full understand it, it is necessary to know spanning-tree and IS-IS and you'll notice that both technologies has contributed to the fabricpath concept.
Based on your observation:
" I was under impression that when fabric path enters the fabric topology it forwards all traffic as normal L3 protocol would. Does it have to select the best topology to traverse"
Cisco use IS-IS which is a Layer 3 routing protocol as you may know already.
"For broadcast and unicast traffic it identifies multidesination tress. What are multidestination trees and how does traffic route through it."
Here is the explanation:
"
When multi-destination traffic enters a FabricPath domain, the ingress switch uses a hashing algorithm in order to decide which FTag to program in the FabricPath header. Each FabricPath Topology has two multi-destination trees, FTag 1 and FTag 2 trees. Each FTag has a root switch that is calculated similar to a Spanning-tree root. The election is based on FabricPath priority and system ID. The switch with the highest priority, or the system ID when the priority is defaulted, becomes the root for FTag 1 and the runner-up is the root for FTag 2.
Once a FTag is selected by the ingress FabricPath edge switch, the rest of the FabricPath core forwards the multi-destination packet based on that FTag. A multi-destination packet includes any broadcast, multicast, or unknown unicast packet. Each switch forwards the packet based on the lowest cost to the root. Once the root receives the packet, it forwards it to all switches in that FTag except the switch from which it was received."
You can read more about it here:
https://www.cisco.com/c/en/us/support/docs/switches/nexus-5000-series-switches/116303-technote-nexus-00.html
I hope that helps.
-If I helped you somehow, please, rate it as useful.-
