Abstract: Concurrency control is the problem that database management system (DBMS) meets with difficulties, especially distributed DBMS. There are two main methods of concurrency control such as locking-based and timestamp-based. Each method gets involved in its own disadvantages, but locking-based approach is often realized in most distributed DBMS because its feasibility and strictness lessen danger in distributed environment. Otherwise, timestamp ordering algorithm is merely implemented in central DBMS due to the issue of local counter synchronization among sites in distributed environment. The common solution is broadcasting the message about the change of local counter (at one site) over distributed network so that all remaining sites “know” to update their own counters. However, this solution raises some disadvantages of low performance. So, an advanced approach is proposed to overcome such disadvantages by giving out another measure so-called active number that is responsible for harmonizing local counters among distributed sites. Moreover, another method is proposed to apply minimum spanning tree into reduce cost of broadcasting messages over distributed network.
Cite this paper:
Nguyen, L. (2015) An Advanced Approach of Local Counter Synchronization to Timestamp Ordering Algorithm in Distributed Concurrency Control. Open Access Library Journal, 2, 1-5. doi: 10.4236/oalib.1100982.
 Ozsu, M.T. and Valduriez, P. (2011) Principles of Distributed Database Systems. 3rd Edition, Springer, Berlin.