Collaborative computing/Reliable/unreliable service

From Wikiversity
Jump to navigation Jump to search

Whether a connection service is reliable or unreliable comes down to the quality of service the connection service offers.

A reliable service is usually implemented by asking the receiver to acknowledge the receipt of each datapackage/frame so the sender is confirmed that the data has been received. Thereby the service will never lose any data. This process do of cause have disadvantages in the sense that it introduces overhead and delays, which sometimes are undesirable.

Unreliabe services on the other hand, never ask the receiver of a message to acknowledge the receipt, this ultimately means that a unreliable service sometimes will lose data due to damage in transit. The advantage of this type of service is that it introduces less overhead and fewer delays.

Examples of reliable and unreliable services:

Tanenbaum, A. & Wetherall, D. (2011) “Computer Networks”, 5th edition, Prentice hall, page 36-37