Comparative Study of Concurrency Control Techniques in Distributed Databases Anand Mhatre Department of Computer Engineering Ramrao Adik Institute of Technology, Nerul Navi Mumbai, India anand9304@gmail.com Rajashree Shedge Department of Computer Engineering Ramrao Adik Institute of Technology, Nerul Navi Mumbai, India rajashree.shedge@rait.ac.in Abstract— In today’s world many of researches have been done on distributed databases. The main issue in distributed databases is to maintain consistency in databases. To maintain consistency in database, correctness criteria must be met. Many of the concurrency control methods are presented earlier, but they have problems about delay, performance, waiting time and number of message exchanges while maintaining correctness. Our paper presents comparison of the recent concurrency control methods considering the above mentioned parameters. Keywords- Distributed Database; Concurrency control; I. INTRODUCTION In distributed databases, there are multiple sites involved on which the data is stored. Hence concurrency control is necessary to ensure reliability and consistency of transactions on these databases [1]. Database is inconsistent when transactions are in deadlock. Therefore concurrency control is needed to maintain database in consistent state. For concurrency control serializability is the most important criterion. The serializability ensures the correctness of the database by converting conflict equivalent schedule to a serial schedule [1]. The basic concurrency control methods in distributed system are: two phase locking (2PL), where transaction obtain lock on data item when they read and convert this lock to write when they need to update it. In wound wait (WW) rather than waiting for the information from all sites deadlocks are prevented by use of timestamps. The third method is Basic Timestamp Ordering. Like WW it employ transaction startup timestamp but use it differently. Distributed certification is operated by exchanging certification information. In all above methods there is some point of deadlock situation formed between the executions of operations. Hence some advance concurrency control methods are proposed like Speculative locking [2], Validation Queue [3], Stamp based [4]. In validation queue transactions are validated on client side and also on server side. In Stamp based, validation is done by matching stamp values. In speculative locking, validation is based on status of preceding transaction. But these methods vary in terms of performance, delay, waiting time and number of message exchanges. Rest of the paper is organized as follows section II describes related work; Section III shows promises of distributed databases; section IV describes distributed concurrency control algorithms; section V gives comparative analysis; finally conclusion is done in section VI. II. RELATED WORK There are many techniques proposed for controlling the concurrent transactions in distributed databases apart from the basic techniques. In [2] Mohit Goyal, T. Ragunathan and P. Krishna Reddy proposed the Speculative locking protocols for distributed environment. Fahren Bukhari and Santosh Shrivastava [3] proposed ROCC (Read Commit Order Concurrency Control) scheme. There are two validation queues, one CVQ (Cache validation queue) which is located at the client side and maintained by local cache manager and another is SVQ (Server validation queue) located at server and maintained by scheduler component. In [4], priority protocol is explained by implementing a timestamp where another value of flag is added and this value doesn’t change unless transaction update has been successfully committed. Atul Adya, Robert Gruber Barbara, Liskov Umesh Maheshwari [5] describes an efficient optimistic concurrency control scheme for use in distributed database systems in which objects are cached and manipulated at client machines while persistent storage and transactional support are provided by servers. The scheme provides both serializability and external consistency for committed transactions; it uses loosely synchronized clocks to achieve global serialization. In [6] T. Ragunathan, P. Krishna Reddy have given semantics-based high performance asynchronous speculation based protocol to improve parallelism among Update transactions and read only transactions. In [7] the approach reduces waiting time for read only transactions and improves its performance. Speculation based locking along with synchronous approach is suggested in [8]. In [9] T. Ragunathan, P. Krishna Reddy, and Mohit Goyal propose semantics- based high performance asynchronous speculation based protocol to improve parallelism among Update transactions and read only transactions. In [10] general concurrency control algorithms in the distributed environment is proposed. These includes the locking algorithms, time stamp algorithm and optimistic algorithm. Arun Kumar Yadav and Ajay Agarwal also proposed the transaction processing in distributed environment. Kamal Solaiman, Graham Morgan [11] suggested optimistic algorithm for transactions resides on resource constraint system. 2014 Fourth International Conference on Communication Systems and Network Technologies 978-1-4799-3070-8/14 $31.00 © 2014 IEEE DOI 10.1109/CSNT.2014.81 378 2014 Fourth International Conference on Communication Systems and Network Technologies 978-1-4799-3070-8/14 $31.00 © 2014 IEEE DOI 10.1109/CSNT.2014.81 378