# Switching Techniques ## What are Switching Techniques Switching techniques refer to the different methods that switches use to forward frames from one port to another. The choice of switching technique affects the latency introduced by the switch, the error checking performed, and the overall performance of the network. There are three main switching techniques used in Ethernet switches: store-and-forward switching, cut-through switching, and fragment-free switching. ## Store-and-Forward Switching In store-and-forward switching the switch receives the entire frame before it begins forwarding it. After receiving the complete frame the switch performs a CRC check to verify the frame has no errors. If the frame is error-free it is forwarded to the appropriate port. If the frame has errors it is discarded. Store-and-forward switching provides the highest level of error checking but introduces the most latency because the entire frame must be received before forwarding begins. The latency depends on the frame size. Larger frames take longer to receive and therefore introduce more latency. Store-and-forward switching is required when the switch needs to translate between different port speeds, for example from a 10 Gbps uplink to a 1 Gbps access port. ## Cut-Through Switching In cut-through switching the switch begins forwarding the frame before it has completely received it. The switch reads only the destination MAC address, which is in the first 6 bytes of the Ethernet frame, and immediately begins forwarding the frame to the appropriate port while the rest of the frame is still being received. This dramatically reduces latency compared to store-and-forward switching because forwarding begins almost immediately. However cut-through switching does not perform CRC error checking, so corrupted frames are forwarded rather than discarded. This can propagate errors through the network. Cut-through switching is suitable for low-latency applications where error rates on the physical links are very low. ## Fragment-Free Switching Fragment-free switching is a compromise between store-and-forward and cut-through. The switch receives the first 64 bytes of the frame before beginning to forward it. The first 64 bytes are significant because most collision fragments, which are malformed frames resulting from collisions, are shorter than 64 bytes. By checking the first 64 bytes the switch filters out the most common type of corrupted frames while still starting to forward before receiving the entire frame. Fragment-free switching provides lower latency than store-and-forward while filtering collision fragments that cut-through switching would pass. ## Adaptive Switching Some advanced switches use adaptive switching, which dynamically switches between cut-through and store-and-forward modes based on the error rate observed on a port. If the error rate is low the switch uses cut-through mode for low latency. If the error rate rises above a threshold the switch switches to store-and-forward mode to filter errors.