Title: Performance evaluation of speculative semantics-based algorithm for read operations in distributed file system

Authors: T. Lakshmi Siva Rama Krishna; T. Ragunathan

Addresses: Department of Computer Science and Engineering, Koneru Lakshmaiah Education Foundation, Vaddeswaram, Guntur, Andhra Pradesh, 522502, India ' Department of Computer Science and Engineering, SRM University AP, Amaravati, Andhra Pradesh, India

Abstract: File sharing semantics is proposed in the literature for sharing the files among the concurrent applications executed in the system. Many file sharing semantics are proposed for the distributed file system (DFS). Existing file sharing semantics in the distributed file system are unable to provide the most recent data to the read client applications while concurrent write client applications are modifying that data. We have proposed a new file sharing semantics called speculative semantics which can provide the most recent data to read client applications without affecting data consistency. We have proposed a new metric called data currency to compare the performance of the file sharing semantics. In this paper, we have developed an analytical model to measure the performance of proposed speculative semantics-based algorithm with session and linearisability semantics-based approaches. The results obtained through analytical model indicate that the proposed speculative semantics-based algorithm performs better than the session and linearisability-based approaches by considering the metric data currency.

Keywords: distributed system; distributed file system; DFS; concurrency; file sharing semantics; consistency; speculation; performance; big data.

DOI: 10.1504/IJCNDS.2019.098870

International Journal of Communication Networks and Distributed Systems, 2019 Vol.22 No.3, pp.275 - 293

Received: 14 Dec 2017
Accepted: 18 Feb 2018

Published online: 09 Apr 2019 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article