ARQ protocol

from Wikipedia, the free encyclopedia

ARQ protocols ( English Automatic Repeat reQuest , dt. Automatic repeat request ) are used in computer networks to ensure reliable data transmission by means of retransmissions.

With the possibility of error detection , a receiver can determine transmission errors that have occurred. This can communicate the result of the error detection to the sender of the message via a return channel . This is usually done by transmitting what are known as ACK / NAK signals (acknowledgment or negative acknowledgment, i.e. correct receipt confirmed or repeat request). If necessary, a disturbed message is retransmitted until it has reached the recipient without errors.

The following protocols , which mainly differ in the achievable throughput , are considered to be fundamental:

Stop-and-wait

Stop-and-wait (. English for holding on and waiting , even as Mail and-wait referred to) is the simplest method is: Having sent a piece of data, the sender must wait for acknowledgment before sending the next packet can. If it does not receive an acknowledgment within a certain time frame ( timeout ), it must send the data unit again.

with ;
: maximum throughput
: Number of information bits
: Number of bits while waiting for acknowledge
: Average number of transmission attempts
: Probability of retransmission in the event of an error

Example: TFTP, the Trivial File Transfer Protocol .

Go-Back-N

Go-Back-N (Engl. For Jump N back ) is a procedure that allows a significantly higher throughput as opposed to stop-and-wait.

The sender can send several data units without having to wait for an acknowledgment. How many there are depends on the so-called window size. If this is n, the sender can send n-1 further data units before the recipient has to confirm the first unit. The receiver can also confirm several data units at once (cumulatively), so an acknowledgment for n + i indicates that all units from n to n + i have been correctly received. If a timeout occurs while waiting for the confirmations, the sender retransmits all data units in the window. So he goes back to the last unacknowledged sequence number N . Since it can be the case that only one data unit was not properly transmitted and all of the data that were sent afterwards are nevertheless retransmitted, transmission capacity is wasted at this point.

The number of data units that can be sent without confirmation (i.e. the window size) is determined by a flow control procedure .

Example: TCP, the Transmission Control Protocol .

Selective repeat

(Engl. for Selective Repeat ) The other common error handling strategy for ARQ protocols is selective repetition. Here an incorrect frame is discarded, but the frames received afterwards are stored in a buffer in the receiver and confirmed. When the time has expired at the sender, only the oldest unacknowledged frame is retransmitted. If this frame arrives correctly, the receiver can subsequently transfer all frames stored in the buffer to the network layer. The selective repetition is often combined with the sending of a negative acknowledgment (NAK, Negative Acknowledgment) by the recipient if the recipient detects an error such as a checksum error or a frame out of sequence. NAKs trigger the retransmission before the corresponding timer expires and therefore improve performance. If the NAK is lost, the sender waits for the timeout for the frame and sends it (and only it) again on its own initiative. The selective repetition corresponds to a reception window greater than 1.

If these protocols are combined with the methods of channel coding , one speaks of the hybrid ARQ protocols.

Example: TCP SACK

Web links