부산대학교 도서관

Usually high education is expensive for most of the top universities in the world. Therefore eLearning platforms provide a way of affordable education with an effective cost of investments. In the last decade, cloud computing research and adoption increased greatly due to its many advantages including economic benefits, ease of management, power saving and so on. Mainly, it provides the means to organize and deliver a wide variety of software services including eLearning environments. Along with cloud computing, file systems solutions were improved to meet requirements imposed by the distribution characteristics of clouds. The distribution level offile systems, data management, data seek methods and many other features of distributed file systems, influence performance of eLearning environments. Distributed file systems started with Andrew File System, still used today, continued with Google File System - a parallel and distributed file system suitable for Google search engine and Ceph - one of the most innovative distributed file system, developed a few years ago and considered the new dream network distributed file system. It has a very good distribution level, security, high availability and many other characteristics. Also, clouds are constructed over computer networks topologies. Basically, topologies are different types of graphs, each with a different set of advantages and disadvantages. In this article we have studied how distributed file systems over computer network clusters influence eLearning environments from an architectural point of view and how its performance may be improved. We also propose a high level architecture of cloud platforms more suitable for eLearning software by outlining several advantages that overcome issues related to distributed eLearning. In our proposed architecture, we use Ceph as a data storage environment and several domestic network topologies. In the end, we study how Software Define Network and Open Flow could model the clouds environment in a more suitable way for eLearning applications. [ABSTRACT FROM AUTHOR]