A broad overview is provided in this chapter of core concepts at the network and transport layers. The coverage extends from architectural principles (such as end-to-end argument) to transport and network layer issues, and includes a short overview of popular protocols. The purpose of this chapter is to provide a broad context into which optical networking fits. In particular, the control plane of optical networks is an extension of the multiprotocol label switching (MPLS) architecture, known as generalized MPLS (G-MPLS).

Optical networks deal with large aggregates of packet traffic, but it is important to ensure that such aggregates preserve the integrity of end-to-end microflows (for example, transmission control protocol (TCP) flows are affected by reordered packets). The discussion of routing algorithms provides a counterpoint to the style of control-plane algorithms used in optical networks. Optical networks also change the trade-off, with respect to problems such as quality of service (QoS). Complex QoS mechanisms need to be placed in the network only when network capacity is scarce. With optical networks, it is likely that, for several years, the core network may have more capacity than the sum of all its feeder or metro-area networks. Moreover, the fast provisioning and reconfigurability of such networks provides new traffic engineering tools for a network operator to meet the dynamic QoS needs of its customers.

The discussion of the end-to-end principle shows that functions (for example, intelligent optical networking) can be placed inside the Internet, provided that this offers a favorable economic trade-off in achieving the basic goal of the Internet: connectivity.