ROWA procedure

This article or the following section is not adequately provided with supporting documents ( e.g. individual evidence ). Information without sufficient evidence could be removed soon. Please help Wikipedia by researching the information and including good evidence.

The Read-One-Write-All method ( English for read one, describe all ) or ROWA method is a simple method for synchronizing replicated data . The principle is based on the fact that changes to a data object are always carried out synchronously on all replicas; a change can only be successful if it has been made on all replicas.

The advantage of the process is that all replicas are always up to date and none are out of date. The availability of a replicated data object is very high for read operations because the current value of a data object can be retrieved from any replica without delay. Read availability increases with the degree of replication (number of replicas).

In order to be able to carry out a write operation successfully, all replicas must be reachable. The higher the degree of replication, however, the lower the probability that all replicas can also be reached to carry out a change. Because of this, the method has only a low write availability. In addition, as the degree of replication increases, so does the effort for change operations (for example, the number of messages required).

The ROWAA procedure provides a remedy for the problem of poor write availability. With ROWAA ( R ead- O ne- W rite- A ll- A vailable) only the actually available replicas are changed so that a change can be carried out successfully even if not all replicas are accessible. For replicas that are saved on computers that have failed in the meantime, the changes are logged and followed up on restart. However, the procedure requires additional work for validation. In addition, the mutual consistency of the replicas can only be guaranteed in the event of computer failures, but not in the case of network partitioning, because changes to the same data object can then take place in different partitions which can no longer be synchronized.

Another possibility to increase the write availability is to combine the ROWA procedure with so-called voting procedures (e.g. weighted voting ). Examples of such hybrid methods are " missing writes " and " virtual partition ".