Switching can be performed on a per-hop basis during the actual transfer, or prior to transfer by the sender of the data. The first method, hop-by-hop, is the Internet's most common switching method. Each data packet is labeled only with the final destination address, and a new switching decision is made at each step along the path to select the next hop. The second method, commonly referred to as source routing, gives the sender complete control over the path taken by the data, but requires this path information to be precomputed and transmitted with the data. It also is more diffcult to recovery from network faults which occur between path selection and the actual data transfer. Internet supports source routing with several options in the IP Protocol, particuarly Loose Source and Record Route and Strict Source and Record Route.
For the modern network engineer, the two most important distinctions are between bridging and routing. Bridging requires no correspondance between network addresses and physical locations, but is less efficient and is used primarily at the Data Link Layer. Routing is more efficient, but requires network addresses to be assigned in predictable patterns and is used primarily at the Network Layer. Switching nodes that perform bridging are called bridges; those that perform routing are called routers; those that can perform both are sometimes called brouters.