Building Cost-Efficient Computer-Aided Learning Environments via Virtualization and Service-Based Software Shueh-Cheng Hu1, I-Ching Chen2, and Yaw-Ling Lin3 1
Dept. of Computer Science and Comm. Eng., Providence University, Taichung, Taiwan, ROC 2 Dept. of Information Management, ChungChou Institute of Technology, Changhua, Taiwan, ROC 3 Dept. of Computer Science and Information Eng., Providence University, Taichung, Taiwan, ROC
[email protected],
[email protected],
[email protected]
Abstract. With increasing and various pedagogical activities are conducted within them, computer-aided learning environments play significant roles in educational institutes, enterprises, and organizations. The cost of operating computer-aided learning environments inevitably goes up along with increasing demands from instructors and learners. Consequently, it is a critical issue to figure out how to provision these pivotal environments more cost-efficiently without compromising service level and users experience. This article elaborates how the two key enabling techniques of cloud computing: system virtualization and web service could be applied to create an agile computeraided learning environments with improved cost-efficiency. Keywords: computer-aided learning, cloud computing, virtualization, Web service, SaaS.
1 Introduction During the last two decades, computer-aided learning (CAL) platforms and applications prevail over schools and organizations around the globe [1]. Basically, CAL environments include Web-based learning environments which are also known as virtual learning environments (VLEs) [2] and classrooms that are equipped with information and computer technologies (ICT). Not only offering flexibility for learners and instructors, various information and computer technologies also enriched pedagogical resources as well as innovated instructional methodologies [3]. From the perspectives of main users in CAL environments: instructors and learners, learning effectiveness is their top concern. Thus, they tend to keep asking more offering and support to facilitate their teaching and learning activities within CAL environments. Generally speaking, the operation of CAL environments relies on acquiring and consuming two types of resources: pedagogical software and supportive S. Lin and X. Huang (Eds.): CSEE 2011, Part IV, CCIS 217, pp. 525–529, 2011. © Springer-Verlag Berlin Heidelberg 2011
526
S.-C. Hu, I.-C. Chen, and Y.-L. Lin
resources. The former one delivers learners instructional materials and aims to fill the gap between learners' current level of knowledge and the desired level. The latter one include learning management system, operating systems, hardware, utilities, and maintenance staff. In contrast with instructors and learners who overlook cost issue, administrators must be cost-conscious while dealing with operation of CAL environments. The increasing prevalence and significance of CAL environments makes administrators must find not only a learning-effective, but also a cost-efficient way to build and operate them. In other words, they need to find a good balance between the quality of pedagogical services being delivered through CAL environments and the corresponding cost for provisioning them. The subsequent section analyzes the cost of provisioning CAL environments, the conceivable waste and the corresponding reasons. Followed by a section describing promising solutions for reducing the waste, and a conclusion remarks.
2 Where the Cost Was Spent and How They Was Wasted To provision VLEs and ICT-equipped classrooms, learning management system, operating systems, hardware, utilities, and maintenance staff are required. Thus, the cost for acquiring and maintaining them is a major part of the overall cost for provisioning CAL environments. In addition, license fee for acquiring software that deliver pedagogical value to learners taking various courses accounts for another significant portion of the overall cost. In contrast to conventional educational environments, the occurrence and duration of learning activities in VLEs are unpredictable due to temporal flexibility and selfdirected aspect of asynchronous Web-based learning environments [4], which currently dominate the various types of VLEs. Consequently, it is very difficult to forecast and allocate exact amount of resources to provision VLEs that aim to support pedagogical activities with timing irregularity. To offer acceptable quality of pedagogical services and avoid complaints from unsatisfied users, most organizations tend to take the over-provisioning approach to operate their VLEs; i.e., they need to constantly and excessively supplying resources including servers, peripherals, power, cooling facilities, and staff, to meet the peak workload of their VLEs' operation. Taking this excessive approach to provision VLEs will satisfy demanding users, but definitely will result in that equipments and all other supportive resources are under-utilized most often, which will lower down the overall cost efficiency of operating CAL environments. Furthermore, operating under-utilized computer systems wastes volume of ineffectual electricity power [5] and cooling utilities. Regarding software usage, if a particular software is helpful to facilitate instruction or improve learners' skills, before acquiring and installing them to augment learners learning activities, organizations usually need to pay multiple copies' full year license, no matter how intensively the licensed software will be used within the organization. Obviously, waste will arise from under-utilized software, which might have been seldom used by relative few users even though multiple copies were licensed for full year.
Building Cost-Efficient Computer-Aided Learning Environments
527
To contract the aforementioned waste, administrators need to find new schemes for provisioning Web-based learning environments and acquiring pedagogical software, instead of adhering to the over-provisioning approach.
3 Related Techniques and Their Roles in Making CAL Environments More Elastic Two key components that enable cloud computing [6]: system virtualization [7] and Web service enabled model: software as a service (SaaS) [8], are promising solutions for improving the cost-efficiency of CAL environments' operation. Both techniques deliver cost-efficiency through elasticity of provision. Virtualization has been adopted by increasing number of enterprises and organizations seeking agile computing environments, in which computing capacity could be dynamically expanded or contracted according to actual workload. Basically, virtualization and the technology associated with it enable multiple virtual machines (VM) co-exist within a single machine, and each VM can function with different operating systems. Through virtualization, under-utilized software systems such as VLEs running on different physical machines could be migrated and subsequently consolidated into fewer physical machines. This kind of consolidation means reduced waste on provisioning under-utilized hardware equipments and supportive resources. On the other side, the overall capacity of a system could be expanded if workload excess its current capacity. Regarding how to set up and elastically provision VLEs, Chieu et. al. presented a practical approach about configuring a set of Web applications that could be elastically provisioned according to real demand [9]. That work is noteworthy to those who plan to deploy their elastic VLEs due to most VLEs including the open-sourced Moodle and the commercial BlackBoard are no other than Web applications. Software as a Service (SaaS) reshaped the paradigm through which software deliver functions to users. Unlike traditional software that need to pre-acquired and installed on particular host(s), SaaS deliver user functions that were wrapped in Web-based interfaces through the Internet, or more specifically, the Web service techniques [10, 11]. Users of SaaS can access the latest updated functions of a software without installing and refreshing the required software by themselves, the software are maintained and controlled by software providers on remote sites. The most valuable advantage of SaaS toward building an elastic CAL environment is that SaaS create a new payment scheme: pay-per-use in software industry. Pay-peruse means that users request functionality, in form of service, from software provider whenever they actually need to use the function. After requesting and receiving the service, users need to pay for the particular usage only, and the one-time payment usually is only a fraction of the full-year license fee. With wide adoption of SaaS, operators of CAL environments do not need to license various software with pedagogical value, instead, they pay what their students actually used. Doubtless, this flexibility of payment scheme will reduce waste that are
528
S.-C. Hu, I.-C. Chen, and Y.-L. Lin
caused by pre-licensed but under-utilized software. Moreover, the pay-per-use scheme enables a finer-grained budget allocation. With SaaS, bulky and full-year license for acquiring a pre-determined set of software is no longer necessary. In contrast, a fixed pool of budget could be allocated and used much more efficiently and agilely. For example, licensing an expensive software for instructing only a minor portion of a course that only few students take is almost impossible when taking cost and impact into account, even that software is very helpful to learning the subject. However, if the function of that software could be delivered through the SaaS approach, requesting usage of that software by few users from time to time will become more affordable and rational.
4 Conclusions With increasing prevalence and wider influence, computer-aided learning environments are not only vital to virtual educational institutes, but also to most traditional schools and organizations. However, the current approaches for provisioning Web-based virtual learning environments and pedagogical software waste valuable resources on operating under-utilized CAL environments. Fortunately, there are two techniques: virtualization and Web service that enabling SaaS , have successfully reshaped usage and payment patterns in enterprise computing applications. They make the provision of computing resource more flexible, in addition, make allocation of resource more fine-grained. Thus, they are ideal for reducing the waste on provisioning CAL environments that are seldom fully-utilized by instructors and learners. Acknowledgment. This research work has being funded by the grant from the National Science Council, Taiwan, ROC, under Grant No. NSC 99-2632-E-126-001MY3. We deeply appreciate their financial support and encouragement.
References 1. Schreurs, B., Bacsich, P., Bastiaens, T., et al.: The Rise of Large Scale E-learning Initiatives Worldwide. In: World Conference on E-Learning in Corporate, Government, Healthcare, and Higher Education 2010, Orlando, Florida, USA (2010) 2. BECTA, Virtual and managed learning environments (2003) 3. Mimirinis, M., Bhattacharya, M.: Design of Virtual Learning Environments for Deep Learning. Journal of Interactive Learning Research 18(1), 55–64 (2007) 4. Mike, M., Dafoulas, G.: Patterns of use of Virtual Learning Environments and Students’ Approaches to Learning: a Case Study of Undergraduate Students. In: World Conference on Educational Multimedia, Hypermedia and Telecommunications 2008, Vienna, Austria (2008) 5. Panda, P.R., Silpa, B.V.N., Shrivastava, A., et al.: Power Issues in Servers and Data Centers. In: Power-Efficient System Design, pp. 183–205. Springer, US (2010) 6. Dikaiakos, M.D., Katsaros, D., Mehra, P., et al.: Cloud Computing: Distributed Internet Computing for IT and Scientific Research. IEEE Internet Computing 13(5), 10–13 (2009)
Building Cost-Efficient Computer-Aided Learning Environments
529
7. Dong, H., Hao, Q., Zhang, T., et al.: Formal Discussion on Relationship between Virtualization and Cloud Computing. In: 2010 International Conference on Parallel and Distributed Computing, Applications and Technologies (PDCAT), pp. 448–453 (2010) 8. Goth, G.: Software-as-a-Service: The Spark That Will Change Software Engineering? IEEE Distributed Systems Online 9(7), 3–3 (2008) 9. Chieu, T.C., Mohindra, A., Karve, A.A., et al.: Dynamic Scaling of Web Applications in a Virtualized Cloud Computing Environment. In: IEEE International Conference on eBusiness Engineering, ICEBE 2009 (2009) 10. Petrie, C.: Practical Web Services. IEEE Internet Computing 13(6), 93–96 (2009) 11. Yi, W., Blake, M.B.: Service-Oriented Computing and Cloud Computing: Challenges and Opportunities. IEEE Internet Computing 14(6), 72–75 (2010)
