Wednesday, January 23, 2019

Help me understand MPLS label switching

So I'm trying to understand MPLS. I know what it's typically used for, and I understand the benefits over IP routing under those circumstances. What I don't understand is the label switching and why the protocol is designed this way. From what I'm reading, it seems that every router along the path replaces the label with a new one. I don't see why the label couldn't just be static.

For instance: You assign a label (100) to a packet/frame as it enters the PE, which then forwards it to the next router. At the next router, and the next, and the next (...so on...), you have a label forwarding table. If the label remained 100 throughout its entire journey, each table just has to know what to do with a packet having label 100 (just like IP routing tables, only faster because the table is smaller, the number is shorter, and there's no masking involved). Simple enough, and you can build your pre-determined paths, just as you could build pre-determined unidirectional paths with static IP routes.

I don't understand what you gain by switching the label at every stop... But I'm sure there's something I'm missing. Can someone please explain this?



No comments:

Post a Comment