Ethernet is a distributed LAN protocol, which implies that there is no centralized control node, which means that the failure of a single node would never bring the whole network down. However, can you see any benefit in creating a centralized LAN in which a single node controls the whole network and decides who gets messages and who waits? Explain.

Destination IP-based forwarding is used by the per-router control plane (sometimes called the classical network) to connect with other routers. If messages are forwarded depending on their destinations' IP addresses, what are the boundaries of this approach? Include a description of at least three limitations, preferably with examples. To what extent does Software-Defined Networking (SDN) contribute to solving these problems? The examples will serve as the explanation.

Ethernet, for example, is a connectionless protocol that specifies the path taken by a packet between its source and destination addresses. However, a protocol that is connection-oriented, such as ATM, specifies the path taken by a cell using virtual connection IDs (VPIs and VCIs). Could you please elaborate on this statement?

It's a good idea to weigh the advantages and disadvantages of connectionless and connection-based protocols side by side.