ABSTRACT

Replication [5] is the technique of using multiple copies of a server or a resource for better availability and performance.Each copy is called a replica. The main goal of replication is to improve availability, since a service is available even if some of its replicas are not. This helps mission critical services, such as many financial systems or reservation systems, where even a short outage can be very disruptive and expensive.A prerequisite for realizing the banefits of replication, however, is the devlopement of high performance concurrency machenism. Current applications, such as Web-based services, electronic commerce, mobile telecommunication system, etc., are distributed in nature and manipulate time-critical databases. In order to enhance the performance and the availability of such applications, one of the main techniques is to replicate data on multiple sites of the network. Therefore, the major issue is to develop efficient replica concurrency control protocols that are able to tolerate the overload of the distributed system. In fact, if the system is not designed to handle overloads, the effects can be catastrophic and some primordial transactions of the application can miss their deadlines. In this paper we present CIRS (Concurrency control In Replicated realtime Systems) a state conscious concurrency control protocol in replicated distributed environment which is specially for firm realtime database system. CIRS mechanism uses S2PL (Static Two Phase Locking) for deadlock free environment. It also includes veto power given to a cohort after receiving PREPARE message from its coordinator. Also with some more assumptions like sending an extra message in execution phase but after completion of execution at local copy which is described later in this paper the proposed mechanism has a significant increased performance over O2PL and MIRROR in decreasing execution time of the current transaction and it also decreases the waiting time of transactions in wait queue..