|
Maintaining data consistency in mobile database broadcasts
 A. J. Seetha seetha_aj@cs.annauniv.edu A. Kannan School of Computer Science and Engineering Anna University, Chennai kannan@annauniv.edu
Abstract The broadcast-based data dissemination in wireless environments poses new challenging issues on data consistency of transaction processing. Due to concurrent execution of update transactions and broadcast of data items, mobile transactions generated by mobile clients may observe inconsistency of data values. Among the concurrency control techniques providing data consistency in broadcast environments, Update First with Order (UFO) algorithm has been shown to be more efficient and has minimal impact on server, client and the broadcast schedule. In this paper, we propose a mobile database system that uses a modified UFO algorithm utilizing rule-based approach for concurrency control among mobile and update transactions in the broadcast environment. We use Event-Condition-Action (ECA) rules to make decisions on broadcasting without conflicting. Moreover, this system has an inference engine that makes deduction using forward chaining control flow to provide effective consistency checking during updates. The system discussed in this paper, provides site autonomy between the mobile clients and the server with minimum upstream communication and data consistency that are desirable features to the scalability of applications, which are running in broadcast environments. Introduction The increasing ability to interconnect computers through internetworking, wireless networks, high bandwidth satellite, and cable networks has spawned a new class of information centered applications based on data dissemination. In contrast to traditional, where data are delivered from servers to clients on demand, a wide range of emerging database applications benefit from a broadcast mode for data dissemination. In such applications, the server repetitively broadcasts data to a client population without a specific request. Clients monitor the broadcast channel and retrieve the data items they need as they arrive on the broadcast channel. Such applications typically involve a small number of servers and a much larger number of clients with similar interests. Examples include weather information systems, electronic commerce applications, such as auction and electronic tendering, and traffic control information systems. With data broadcast, mobile transactions do not need to inform the broadcast server before accessing a data item and can get it from the “air” while it is being broadcast. As data dissemination systems continue to evolve, more and more sophisticated client applications will require reading current and consistent data despite updates at the server. Since many data items in mobile computing systems are used to record the real-time information such as current traffic conditions of the roads, current weather conditions of the cities, and news updates, their values will be highly dynamic and sensitive. The updates capture the most current information for the system and refresh the values of the data items in the database [9][10]. Accessing out-dated data items is undesirable and will significantly affect the usefulness of the information to mobile clients [2]. Thus allowing the execution of updates while it is being broadcast is important in maintaining the validity and freshness of the data items. However, if concurrent execution of updates and data broadcast is allowed, the problem of concurrency control must be addressed. Given the limited amount of bandwidth available for clients to communicate with the broadcast server, achieving data consistency efficiently is a challenging research issue. Unfortunately, conventional concurrency protocols, such as two phase locking, optimistic method and timestamp ordering [1] are not suitable for mobile computing systems as the overhead for setting locks and detecting data conflicts in a mobile environment can be very heavy [8]. In this paper, we study the problem of disseminating consistent data items to mobile transactions while allowing updates to be executed concurrently at the database server. An efficient and pioneering method is by broadcasting multiple versions of data items [18]. Consistent data items are provided to mobile transactions by requiring the mobile transactions to read data items committed at the same point of time. The basic multi-version broadcast method is extended in [16,17] for systems with client caches where multiple versions of data items are maintained. By reading cached data items, which are committed at the same time, the data access delay can be much reduced and at the same time data consistency is ensured. Another efficient method for concurrency control between mobile transactions and update transactions is a data re-broadcast scheme called Update First with Ordering (UFO) [3]. Although UFO can provide the most updated values of data items to mobile transactions and at the same time maintain the serializability of the execution between update and mobile transactions, it is designed for read-only mobile transactions in which operations are unordered. In this paper, we extend the UFO protocol for read-write mobile transactions, which uses rule-based approach.  Figure 1: Architecture of Mobile Database System The following are the contributions of the paper
Page 1 of 5
| Next | |