M.S.Khatib* et al. / (IJITR) INTERNATIONAL JOURNAL OF INNOVATIVE TECHNOLOGY AND RESEARCH Volume No.2, Issue No. 3, April – May 2014, 985 – 990. ISSN 2320 –5547 @ 2013 http://www.ijitr.com All rights Reserved. Page | 985 An Analysis of Transaction Management in Distributed Real Time Databases: An Overview M.S.KHATIB Professor Department Of Computer Science & Engineering Anjuman College of Engineering & Technology Nagpur, India. Dr. MOHAMMAD ATIQUE P.G. Department Of Computer Science S.G.B. Amravati University Amravati, India Abstract — A real time distributed computing has heterogeneously networked computers to solve a single problem. So coordination of activities among computers is a complex task and deadlines make more complex. The performance of the system depends on many factors such as traffic workloads data base system architecture, underlying processor, disk speeds, concurrency control, transaction management etc. A simulation study have to be performed to analyze the performance under different transaction scheduling, different workloads, arrival rate priority policies, altering slack factors and preemptive policies. The performance of the distributed system under various conditions is to be monitored and parameters such as arrival rate, transaction size, transaction distribution policies, and execution time are to be analyzed. Keywords- Database systems, real time systems, transaction management, two phase commit, earliest deadline first, rate monotonic analysis I. INTRODUCTION In safety-critical environments[1][2] such as nuclear power plants, automated landing systems for bad weather conditions, distributed collaboration of elementary level, have not only to meet deadlines but most of them are critical in the sense that subsequent instances are to be executed prior to their deadlines. Rigid dependability[3][4] of requirements in this type of distributed systems call for replication of information across sites, for availability even under site crashes and even for faster, i.e. local accessibility are conflicting goals there is not static trade-off between these restrictions, due to the unpredictability of the environment. Thus a high amount of adaptability of system could tolerate deadline failures only to a task-specific degree. After that the next instance will be essentially critical, i.e., its failure would possibly have hazardous consequences for the system, to handle all essentially critical task instances successfully has been denoted by survivability. Into the same vein, task deadlines in such environments cannot be guaranteed, due analysis and synthesis is required. In order to take care of highly critical situations task periods would have to be very short thus potentially causing traffic jams on the network even in situations of moderate criticality. Thus it is imperative for distributed safety-critical systems [5][6] to model the relevant tasks as a periodic in nature. While the principles and constraints mentioned above have strong implications for task management they are even more important for distributed transaction handling. II. LITERATURE REVIEW A. Critically, Sensitivity, Survivability: Roll-back, or restart after preemption are costly and have an unpredictable effect on meeting hard deadlines. From the MELODY system we borrow various adaptability concepts and techniques about task execution. These are used for stepwise generating an adaptive transaction model, which explicitly utilizes the nested transaction structure. The key idea is to save as many tasks or subtask from executing after a transaction invocation. This is based on a novel format concept of transaction similarity [7] [8][9] In MELODY object information is maintained through replication in two different types of copies public and private copies, depending on the decency of their information. This is done for both faster access and higher object availability. Also, the closer the task instance is to essential criticality the more will be the need for most recent information to be neglected while private copies are available locally for the essentially critical stage. This is formally handled through a parameter called sensitivity. Sensitivity is adjusted in combination with criticality. Finally concepts of task and object similarity had been defined and interrelated as measures for deciding about task invocation . Soft real-time systems or predictable applications, typical transaction procedures like roll-back are infeasible in safety-critical systems. During the roll- back phase the resources will remain locked causing harmful effects for conflicting transactions, and even worse, this would be uncontrollable from the sites that manage the latter. Thus roll back should be avoided by all means. A similar problem arises when using restart after a transaction has been preempted by a high priority transaction. While on brought to you by CORE View metadata, citation and similar papers at core.ac.uk provided by International Journal of Innovative Technology and Research (IJITR)