A New Method for Optimization of Deadlock resolution of Distributed Database with Formal Model

Masoomeh Ghodrati; Ali Harounabadi

A distributed system is a network of sites that send messages to each other, they exchange information. One of the most important applications of distributed systems is enabling resource sharing between systems. In such environments, if a sequence of procedures to control resource allocation is not possible to create a deadlock exists. Deadlock problem for a distributed database system that uses locking as a concurrency control algorithm, as there are inherent. Grid systems and Peer to Peer networks, the most common way for the purpose of sharing resources, services in heterogeneous environments, is dynamic and distributed. Data Grid is a distributed computing architecture is a wealth of information and computing resources into a single system to manage the virtual data integration. Grid data management is no easy way, some problems that should be considered such as reliability and availability of data, user access, coverage, network failures or malicious attacks during execution, etc. This problem can be solved using replication technology. In data replication architecture, data is replicated at multiple sites. If one site fails, the problem is independent of the other nodes. Thus, data replication can lead to reliability and increase accessibility. Replication is one of the most widely used techniques in distributed systems. Replication in distributed environment to provide efficient access to data, fault tolerance and enhance the performance of the system is taken into consideration. Iteration, concurrency control and deadlock detection is the most important problem in distributed systems and when data should be shared by the management. This paper is a comprehensive study on deadlock detection in distributed database systems. Deadlock detection of neighbor replication on grid has been used. After deadlock detection by the proposed method can choice optimal of aborting the transaction as a deadlock victim. The following new rule for the modeling of the proposed method using colored Petri nets is presented. In the model proposed the new rules for mapping TWFG with colored Petri nets for modeling the deadlocks detection and resolve. Colored Petri net is considered one of the most widely used formal methods capable of modeling a wide variety of distributed systems are concurrent. A lot of work being done to define the concurrency execution of transactions in Petri nets is that none of these methods of communication with how mapping TWFG with colored Petri nets for modeling the deadlocks detection and resolve. The proposed method is compared with previous work raises a good idea to choose a victim.

Select Volume / Issues:

• Volume 5 Issue 1, Jan. 2014

Year:

2014

Type of Publication:

Article

Keywords:

Resolution Deadlock Cycle; Colored Petri Net; Mapping TWFG; Neighbor Replication on Grid; Distributed Database

Journal:

IJECCE

Volume:

5

Number:

1

Pages:

220-228

Month:

January