DAIS: dynamic access and integration services framework for cloud-oriented storage systems
- PDF / 2,886,609 Bytes
- 20 Pages / 595.276 x 790.866 pts Page_size
- 43 Downloads / 178 Views
(0123456789().,-volV)(0123456789(). ,- volV)
DAIS: dynamic access and integration services framework for cloudoriented storage systems Ajay Kumar1 • Seema Bawa1 Received: 26 May 2019 / Revised: 24 November 2019 / Accepted: 9 March 2020 Ó Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract This paper proposes a framework named Dynamic Access and Integration Services (DAIS) framework. The framework provides on-demand and large scale integrated data access in cloud computing environment. The sole motivation of this research paper is to enhance the performance of Cloud-Oriented Storage (COS) systems. DAIS framework consists of two basic modules namely (1) Adjacency COS Overlay Topology (ACOT) and (2) Cloud Storage Resource Discovery (CSRD). ACOT introduces a self-organizing well balanced m-way tree which removes the necessity of centralized approaches and adopts dynamic and scalable data access in cloud computing environment. CSRD module provides data integration mechanism and efficient resource discovery from geographically distributes COS systems. In this paper, Hadoop test-bed setup has been deployed for experiment and results analysis. All input and output data have been collected in the Hadoop Distributed File System (HDFS). In the experiments, performance, bandwidth and energy efficiency of the proposed DAIS framework have been demonstrated and compared with three different approaches like, Static DAIS, P2P and DSSM. To make it energy efficient, measurement unit and its mathematical model have been defined. Energy efficiency results have been compared with two popular approaches HBase and Hadoop-MR, and found better in most of the cases. Keywords Cloud computing Cloud storage cluster Self-organizing Dynamicity Scalability Storage services Integration services
1 Introduction The distributed systems have been used widely for business-oriented applications and high-performance scientific computations for a significant volume of data calculation in last three decades. In this regards, various distributed computing concepts have been evolved to solve the computational and storage problems like cluster computing [16], grid computing [83], P2P computing [11], marketoriented computing [17], service-oriented computing [78], etc. The cluster computing uses the processing capability within the organization itself. The grid computing optimizes and shares the resources among geographically & Ajay Kumar [email protected] Seema Bawa [email protected] 1
Thapar Institute of Engineering and Technology, Patiala, Punjab, India
distributed systems [5, 50]. P2P computing directly shares the capability to peer node without restriction [31]. Marketoriented computing uses both cluster and grid computing concepts to provide services via Internet on pay basis. Service-oriented computing offers various business solutions in the form of web services. The concepts of cluster, grid, market-oriented and service-oriented computing combined together and working through Internet is known as the cloud comp
Data Loading...
