i understand how mpls label switching works, i understand each router has a cached database of all the labels it would use for each network (for each vpnv4 network), and i understand the header also contains rd information for processing. but why? why not just use cef? transit routers dont need to be aware of vrf or rd's, they just forward packets from one end to the other. how is label switching any different than cef then? just look at the adjacency table, slap on a new ethernet header, and pass it along without thinking. wouldnt adding more headers be a waste of bandwith?
No comments:
Post a Comment