International Research Journal of Engineering and Technology (IRJET) Volume: 04 Issue: 02 | Feb -2017
www.irjet.net
e-ISSN: 2395 -0056 p-ISSN:2395-0072
Partitioning of Query Processing in Distributed Database System to Improve Throughput. Ms. Pratibha B. Patil1 , Mr. Rahul P. Mirajkar2, Ms.Neeta B. Patil3 1Student,
Department of Computer Science and Engineering, Bharati Vidyapeeth’s College of Engineering, Kolhapur, Maharashtra, India
2Assistant
Professor, Department of Computer Science and Engineering, Bharati Vidyapeeth’s College of Engineering, Kolhapur, Maharashtra, India
3Student,
Department of Computer Science and Engineering, Bharati Vidyapeeth’s College of Engineering, Kolhapur, Maharashtra, India
---------------------------------------------------------------------***--------------------------------------------------------------------------
Abstract - Query processing in distributed system calls for
the transmission of records among computers in community. The arrangement of statistics transmission and local information processing is known as distribution strategy for a query. Two cost measures reaction time and total time are used to judge the great of distribution method. Numerous algorithms are used that derive distribution strategies that have minimal reaction time and minimum overall time. The optimal algorithms are used as a foundation to increase a general query processing algorithms. In this article, we introduce using graph partitioning to partition the task of query to optimize the throughput on distributed structures. New heuristic set of rules we've advanced, the congestion avoidance partitioning (CA) algorithm to optimize the throughput and query execution (parallel) algorithm developed for concurrent execution of query in distributed database system. We evaluate the query processing device with and without partitioning algorithm to analyze throughput end result. Key Words: Computer Network, Distributed database, query processing, graph partitioning, concurrent execution.
1. INTRODUCTION In recent years, with the development of computer network and database technology, dispensed database is more and more extensively used; with the increasing application, facts queries are increasingly more complicated, the performance requests are an increasing
© 2017, IRJET
|
Impact Factor value: 5.181
|
number of excessive, so query processing is a key trouble of the dispensed database device. In a distributed database surroundings, data stored at exclusive sites linked through community. A distributed database management system (DDBMS) aid advent and maintenance of disbursed database. The studies literature proposes a huge form of query optimization algorithms and overviews on diverse query optimization techniques for distributed database management system[5,6,7]. However, these overviews do no longer try and increase a model of query optimization that explains and gives the algorithms in a uniform manner. This knowledge in case we want to exchange or enlarge current algorithms to conform them to new necessities. In this research we take into account query processing algorithms for an allotted Database machine. To gain good performance for query processing on parallel structures with shared reminiscence, above mentioned initiatives have deployed sever scheduling strategies. The primary trend is to maximize the parallelism among all cores by way of mapping computational nodes of the queries into individual cores. Special graph partitioning strategies have been taken on to provide load stability so far both regardless the communication value or bear in mind it simplest with a low priority[1,2,3]. On distributed structures in which communication cost has an large effect on performance, those strategies are now not appropriate. In this paper, we advocate a
ISO 9001:2008 Certified Journal
|
Page 1058