ORDER: A Dynamic Replication Algorithm for Periodic Transactions in Distributed Real-Time Databases

Many real-time applications need data services in distributed environments. However, providing such data services is a challenging task due to long remote data accessing delays and stringent time requirements of real-time transactions. Replication can help distributed real-time database systems meet the stringent time requirements of application transactions. In this paper, we present two dynamic replication control algorithms designed for medium and large scale distributed realtime database systems. With the data needs information from incoming transactions, our algorithms dynamically determine where and how often the replicas are replicated. In our algorithms, the data replicas are dynamically created upon the requests by the incoming transactions and their update frequencies are determined by the data freshness requirements of these transactions. A detailed simulation study shows that our algorithms can greatly improve the system performance compared to the systems without replication or systems with simple replication strategies such as full replication.