In networks with data packet switching, the transmission delay plays an essential role, since certain real-time applications can no longer be realized at high latency times. To keep this latency as low as possible in switching networks of switching equipment and switches, the cut-through method initiates the switching process even before the complete data packet has arrived.
Packet evaluation in the cut-through process
Unlike the store-and-forward method, the cut-through method does not evaluate the complete data packet, but only the first bytes that contain the MAC address or destination address. Already after evaluation of the destination address with the internal address table, the data packet is forwarded to the corresponding output port via the switching network without error correction. As a result, an extremely low latency is achieved with the cut-through method. Since no error correction is performed, damaged data fragments can also be forwarded, which are later sorted out by layer 2 switches. However, there are also process variants that support fragment-free transmission. In this variant for Ethernet, the data packet ischecked up to the 64th byte, because collisions are usually detected in these bytes.
A variant of the cut-through method is the error-free cut-through method, Error Free Cut-Through. In this method, the switch copies the forwarded data packet and calculates its checksum. Error checksums are counted and if a certain limit is exceeded, the switch switches to the error-free store-and-forward procedure. Only when the error rate drops again does it switch back to the faster cut-through.