Informatics Publishing Limited

Mahmood Zamani ¹, Mohsen Afsharchi ²

Affiliations
1 Department of Computer and Electronics, Islamic Azad University of Zanjan, IR
2 Department of Computer and Electronics, Zanjan University, Zanjan, 45371-38791, IR

Abstract

Background/Objectives: Database systems are one of the largest communicative interfaces between users and service providing organizations. Generally, all the organizations and companies which deal with sensitive financial data and online customer services, such as banks, the telephone company and their payments need to provide services with high accessibility and maximum efficiency. In order to achieve these objectives, the criterion of high accessibility is investigated in order to provide non-stop services in databases. Since the systems with a high number of users at the level of database conduct a lot of transactions, the running transactions allocate system resources to themselves; as a result, the system is faced with resource constraints and its performance is disrupted. Methods/Statistical analysis: In this research, the combination of alternative databases and clustering are used in order to facilitate the recovery and increase the accessibility. Also, job scheduling methods are stated at the time of database failure so that the running clusters can bear the workload of bad clusters. Results: The node allocation algorithm is dealt with in this research which is conducted with Weibull distribution. Compared to the random node allocation strategy, the results of simulation indicate that using the above-mentioned algorithm decreases the effect of node failure on system performance in cluster systems in large scale. Conclusion/Application: The algorithm decreases the average response time.

Keywords

Clustering, Databases, High Accessibility, Rapid Failure Recovery

Full Text