Adaptable ordering protocols in distributed computation

Communication among group members often needs message ordering to guarantee that every member receives the messages in the same order. There are several ordering algorithms designed with efficiency in mind. However, there might often occur environmental changes in distributed computing and the changes undermine the assumption of the algorithms. Therefore it is difficult to presume a suitable ordering algorithm for the environments and apply a single algorithm throughout the lifetime of a system. We propose to select an ordering algorithm from several algorithms adaptively according to the condition of the environment. We present a protocol for selecting the ordering algorithms based on the cost estimation, agreement among members and switching protocol considering the consistency of message ordering. This protocol helps to reduce the ordering cost by adapting the environment. Therefore the system enables us to provide its service continuously with the proper ordering cost even if the environmental changes occur.