Under connectionless communication between two network endpoints, a message can be sent from one endpoint to another without prior arrangement. The device at one end of the communication transmits data addressed to the other, without first ensuring that the recipient is available and ready to receive the data. Some protocols allow for error correction by requesting retransmission.
Connectionless protocols are stateless protocols because the endpoints have no protocol-defined way to remember where they are in a "conversation" of message exchanges. It has lower overhead than connection-oriented communication because, in connection-oriented communication, the communicating peers must first establish a logical or physical data channel or connection in a dialog preceding the exchange of user data. It allows for multicast and broadcast operations in which the same data are transmitted to several recipients in a single transmission.
In connectionless transmissions the service provider usually cannot guarantee that there will be no loss, error insertion, misdelivery, duplication, or out-of-sequence delivery of the packet. However, the effect of errors may be reduced by implementing error correction within an application protocol.
In connectionless mode, there is less opportunity for optimization when sending several data units between the same two peers. By establishing a connection at the beginning of such a data exchange the components (routers, bridges) along the network path would be able to pre-compute (and hence cache) routing-related information, avoiding re-computation for every packet. In connection-oriented communication, network components can also reserve capacity for the transfer of the subsequent data units of a video download, for example.