An Overview of E-Learning in Cloud Computing

A. Fern ndez1, D. Peralta2 , F. Herrera2, and J.M. Bentez2 a

1

Dept. of Computer Science. University of Ja n, e Ja n, Spain e [email protected] Dept. of Computer Science and Articial Intelligence, CITIC-UGR (Research Center on Information and Communications Technology). University of Granada, 18071 Granada, Spain {dperalta,herrera,jmbs}@decsai.ugr.es

Abstract. E-Learning is the topic related to the virtualized distance learning by means of electronic communication mechanisms, specically the Internet. They are based in the use of approaches with diverse functionality (e-mail, Web pages, forums, learning platforms, and so on) as a support of the process of teaching-learning. The Cloud Computing environment rises as a natural platform to provide support to e-Learning systems and also for the implementation of data mining techniques that allow to explore the enormous data bases generated from the former process to extract the inherent knowledge, since it can be dynamically adapted by providing a scalable system for changing necessities along time. In this contribution, we give an overview of the current state of the structure of Cloud Computing for applications on e-learning. We provide details of the most common infrastructures that have been developed for such a system, and nally we present some examples of e-learning approaches for Cloud Computing that can be found in the specialized literature.

1 Introduction
The Electronic Learning, better known as E-Learning [13], is dened as an Internetenabled learning. Components of e-Learning can include content of multiple formats, management of the learning experience, and an online community of learners, content developers and experts. The study summarized the main advantages, which include exibility, convenience, easy accessibility, consistency and its repeatability. With Information Technologies (IT), there is a growing trend regarding the research and exploitation of this kind of e-Learning platforms. There exist several initiatives at different educative levels, from which some examples are the Khan Academy1, the Virtual Learning Center of Granada University (CEVUG-UGR),
1

www.khanacademy.org

L. Uden et al. (Eds.): Workshop on LTEC 2012, AISC 173, pp. 3546. c Springer-Verlag Berlin Heidelberg 2012 springerlink.com

36

A. Fern ndez et al. a

the Open University of Catalonia, the MIT Open CourseWare, or the Free Online Course of the Standford University. The virtual courses that are supported by the e-Learning approach favors the achievement of a higher impact for the educative framework than those of the classical attendance group. As an example, in the rst edition of the Machine Learning course of Stanford2 more than 160,000 worldwide students were registered. These dimensions affects different issues; on the one hand, the infrastructure provisions that are necessary to give a concurrent service for that amount of students clearly exceed the capabilities of a conventional web server. Furthermore, the demand of the teaching resources usually vary in a dynamic and very quick way, and presents high peaks of activity. To attend requests during these periods of time without other system services to be resented, it will be necessary to prepare a quite superior infrastructure than that required for the regular working of the learning institution. An alternative would be to provide those services depending on the demand and only paying for the resources that are actually used. The answer to these necessities is the Cloud Computing environment. Cloud Computing [3, 18] is a computation paradigm in which the resources of an IT system are offered as services, available to the users through net connections, frequently the Internet. It is a model of provision of IT services offered through a catalog that answers to the necessities of the user in a exible and adaptive way, only billing for the actual usage that is made. Therefore, two of the distinctive features of this paradigm are, on the one hand, the use of resources under demand and, on the other hand, the transparent scalability in such a way that the computational resources are assigned in a dynamical and accurate manner when they are strictly necessary, without the requirement of a detailed understanding of the infrastructure from the users point of view. With these characteristics, the Cloud platforms arise as accurate alternatives to traditional computer centers. They represent a signicative alternative versus the acquisition and maintenance of the computer centers. Additionally, the e-learning platforms of the large dimensions which we mentioned above generate extensive registers of interaction among students-platformteachers. These data bases contain signicative information not dened in a precise way. Data Mining techniques must be applied to extract this information [23, 17]. Therefore Educational Data Mining3 comes up, being this a discipline whose object of interest is the development of new methodologies to explore the data that are generated in the activity of the educational systems (mainly those with a technological base) and the application of such methods to achieve a better understanding of the behaviour of the students, and how to design procedures and material that ease the learning process.
2 3

http://www.ml-class.org/course/class/index http://www.educationaldatamining.org/

An Overview of E-Learning in Cloud Computing

37

In clear connection with this process we may nd the Intelligent Tutoring Systems4 which are computer based systems to support the teaching-learning process. Usually, they are intelligent systems able to drive the learning process of the student providing him/her feedback based on the progress of the student and the results of periodical tests. The process of Educational Data Mining interacts with an Intelligent Tutoring System by extending and rening its knowledge base. Taking into account the dimensions and growing capacity of the computational resources (stable storage, memory and CPUs) a Cloud platform is also a natural structure for the implementation of data mining techniques and their application to growing data-sets (Big Data). However, many of the data mining techniques do not have an adequate scalability. This is an aspect that grows in importance and that have attracted the interest of researchers and companies. In order to overview all these aspects, this contribution is arranged as follows. In Section 2 we introduce the main concepts on Cloud Computing, including its infrastructure and main layers. Next, Section 3 presents the features of the e-Learning approach, stressing the advantages of the migration of such a system to a Cloud Computing environment and showing some examples of real applications of this kind. Finally, the main concluding remarks are given in Section 4.

2 Basic Concepts on Cloud Computing

We may dene an SOA [15] as an integration platform based on the combination of a logical and technological architecture oriented to support and integrate all kind of services. In general, a Service in the framework of Cloud Computing is a task that has been encapsulated in a way that it can be automated and supplied to the clients in a consistent and constant way. Any component can be considered as a service, from entities closest to hardware such as the storage space or the computational time, to software components aimed at authenticating a user or to manage the mailing, the management of a data base or the monitoring of the use of the system resources. In this section we will give a brief introduction to the Cloud Computing environment, rst describing its main features, next by presenting the layers in which this platform is built of, and nally pointing out several technological difculties that should still be addressed to improve the quality of this paradigm.

2.1 Introduction to Cloud Computing

The philosophy of Cloud Computing mainly implies a change in the way of solving the problems by using computers. The design of the applications is based upon the use and combination of services. On the contrary that occurs in more traditional approaches, i.e. grid computing, the provision of the functionality relays on this use

http://aaai.org/AITopics/IntelligentTutoringSystems

38

A. Fern ndez et al. a

and combination of services rather than the concept of process or algorithm. The idea behind this is that grid computing mainly focuses on high performance computing whereas Cloud Computing offers both standard and intensive computation. Additionally, Cloud offers more services than grid computing, i.e web hosting, multiple Operating systems, DB support and much more. Finally, grids tends to be more loosely coupled, heterogeneous, and geographically dispersed compared to conventional cluster computing systems. Clearly, this brings advantages in different aspects, for example the scalability, reliability, and so on, where an application, in the presence of a peak of resources demand, because of an increase of users or an increase of the data that those provide, can still give an answer in real time since it can get more instances of a determinate service; the same occurs in the case of a fall of the demand, for which it can liberate resources, all of these actions in a transparent way to the user. The main features of this architecture are its loose coupling, high inter-operativity and to have some interfaces that isolate the service from the implementation and the platform. In an SOA, the services tend to be organized in a general way in layers or levels (not necessarily with strict divisions) where normally, some modules use the services that are provided by the lower levels to offer other services to the superior levels. Furthermore, those levels may have different organization structure, a different architecture, etc.

2.2 Cloud Computer Layers

There exists different categories in which the service oriented systems can be clustered. One of the most used criteria to group these systems is the abstraction level that offers to the system user. In this manner, three different levels are often distinguished , as we can observe in Figure 1. In the remainder of this section, we will rst describe each one of these three levels, providing the features that denes each one of them and some examples of the most known systems of each type. Next we will present some technological challenges that must be taken into account for the development of a Cloud Computing system. Infrastructure as a Service (IaaS): IaaS is the supply of hardware as a service, that is, servers, net technology, storage or computation, as well as basic characteristics such as Operating Systems and virtualization of hardware resources [8]. Making an analogy with a monocomputer system, the IaaS will correspond to the hardware of such a computer together with the Operating System that take care of the management of the hardware resources and ease the access to them. Platform as a Service (PaaS): At the PaaS level, the provider supplies more than just infrastructure, i.e. an integrated set of software with all the stuff that a developer needs to build applications, both for the developing and for the execution stages. In this manner, a PaaS provider does not provide the infrastructure directly, but making use of the services of an IaaS it presents the tools that a developer needs, having an indirect access to the IaaS services and, consequently, to the infrastructure [8].

An Overview of E-Learning in Cloud Computing

39

Fig. 1 Illustration of the layers for the Services Oriented Architecture

Software as a Service (SaaS): In the last level we may nd the SaaS, i.e. to offer software as a service. This was one of the rst implementations of Cloud services. It has its origins in the host operations carried out by the Application Service Providers, from which some companies offered to others the applications known as Customer Relationship Managements [5].

2.3 Technological Challenges in Cloud Computing

Cloud computing has shown to be a very effective paradigm according to its features such as on-demand self-service since the customers are able to provision computing capabilities without requiring any human interaction; broad network access from heterogeneous client platforms; resource pooling to serve multiple consumers; rapid elasticity as the capabilities appear to be unlimited from the consumers point of view; and a measured service allowing a pay-per-use business model. However, there are also some weak points that should be taken into account. Next, we present some of these issues: Security, privacy and condence: Since the data can be distributed on different servers, and out of the control of the customer, there is a necessity of managing hardware for computation with encoding data by using robust and efcient methods. Also, in order to increase the condence of the user, several audits and certications of the security must be performed. Availability, fault tolerance and recovery: to guarantee a permanent service (24x7) with the use of redundant systems and to avoid net trafc overow. Scalability: In order to adapt the necessary resources under changing demands of the user by providing an intelligent resource management, an effective monitorization can be used by identifying a priori the usage patterns and to predict the load in order to optimize the scheduling. Energy efciency: It is also important to reduce the electric charge by using microprocessors with a lower energy consumption and adaptable to their use.

40

A. Fern ndez et al. a

3 Cloud Computing for E-Learning Tasks

As we stated in the introduction of this work, with the huge growth of the number of students, education contents, services that can be offered and resources made available, e-Learning system dimensions grow at an exponential rate. The challenges regarding this topic about optimizing resource computation, storage and communication requirements, and dealing with dynamic concurrency requests highlight the necessity of the use of a platform that meets scalable demands and cost control. This environment is Cloud Computing. Along this section we will introduce the main advantages and drawbacks to be addressed for e-Learning systems (Subsection 3.1). Then, the signicance of selecting Cloud Computing for this kind of tools will be stressed (Subsection 3.2). The organization and infrastructure necessary for the virtual platform is described next (Subsection 3.3). Finally, we will review some of the e-Learning applications that have been already developed within the Cloud Computing platform (Subsection 3.4).

3.1 Current Challenges of E-Learning Systems

Among the learning technologies, web-based learning offers several benets over conventional classroom-based learning. Its biggest advantages are the reduced costs since a physical environment is no longer required and therefore it can be used at any time and place for the convenience of the student. Additionally, the learning material is easy to keep updated and the teacher may also incorporate multimedia content to provide a friendly framework and to ease the understanding of the concepts. Finally, it can be viewed as a learner-centered approach which can address the differences among teachers, so that all of them may check the condence of their material to evaluate and re-utilize common areas of knowledge [9]. However, there are some disadvantages that must be addressed prior to the full integration of e-Learning into the academic framework. Currently, e-Learning systems are still weak on scalability at the infrastructure level. Several resources can be deployed and assigned just for specic tasks so that when receiving high workloads, the system need to add and congure new resources of the same type, making the cost and resource management very expensive. This key issue is also related to the efcient utilization of these resources. For example, in a typical university scenario, PC labs and servers are under-utilized during the night and semester breaks. In addition, these resources are on high demands mainly towards the end of a semester, following a dynamic rule of use. The physical machines are hold even when they are idle, wasting its full potential. Finally, we must understand that there is a cost related to the computer (and building) maintenance, but that the educational center must pay for the site licensing, installation and technical support for the individual software packages [10].

An Overview of E-Learning in Cloud Computing

41

3.2 On the Suitability of Cloud Computing for E-Learning

E-Learning in the Cloud can be viewed as Education Software-as-a-Service. Its deployment can be performed very quickly since the hardware requirements of the user are very low. Furthermore, as we stated previously, it lessens the burden of maintenance and support from the educational institution to the vendor, allowing them to focus on their core business, also obtaining the latest updates of the system without charges and sharing key resources using Web 2.0 technology. In what follows, we summarize the consequences and implications regarding the development of e-Learning services within the Cloud Computing environment, as pointed out by Masud and Huang in [12]: Accessed via Web: It implies an ease of access since anywhere, any time and any one can access the application, greater demand for Web Development skills. No client-side software needed: Therefore, it has reduced costs for subscriber, as no installation, software maintenance, deployment and server administration costs, and a lower total cost of ownership, reduced time-to-value, fewer IT staff is needed by the institution. Pay by subscription based on usage: Which is suitable for Software Model Education market, and can gain access to more sophisticated applications. SaaS server may support many educational institutions: Since the application is running on a server farm, the scalability in inherent to the system. As student usage grows, the software performance will not degrade. All subscriber data held on SaaS server: Very high level of security is needed by SaaS provider in order to gain trust of subscribers and sophisticated multitenanted software architecture. The subscriber data is distributed between many providers and it must be integrated in order to gain overview of business, higher demand for system and data integrators. Finally, several potential values of Cloud Computing for education as stressed by Ouf et al. in [14] include the following: No need for backing up everything to a thumb drive and transferring it from one device to another. It also means students can create a repository of information that stays with them and keeps growing as long as he wants them. Crash recovery is nearly unneeded. If the client computer crashes, there are almost no data lost because everything is stored in the cloud [16]. Allow students to work from multiple Places (home, work, library ... etc), nd their les and edit them through the cloud and browser-based applications can also be accessed through various devices (mobile, laptop and desk top computers, provided internet access is available) [2]. Flexibility: Scale infrastructure to maximize investments. Cloud computing allows user to dynamically scale as demands uctuate [6]. Improved improbability : it is almost impossible for any interested person (thief) to determine where is located the machine that stores some wanted data (tests, exam questions, results) or to nd out which is the physical component he needs to steal in order to get a digital asset [16].

42

A. Fern ndez et al. a

Virtualization: makes possible the rapid replacement of a compromised cloud located server without major costs or damages. It is very easy to create a clone of a virtual machine so the cloud downtime is expected to be reduced substantially. Centralized data storage: losing a cloud client is no longer a major incident while the main part of the applications and data is stored into the cloud so a new client can be connected very fast. Imagine what is happening today if a laptop that stores the examination questions is stolen. Monitoring of data access becomes easier in view of the fact that only one place should be supervised, not thousands of computers scattered over an extensive geographical area, for example. Also, the security changes can be easily tested and implemented since the cloud represents a unique entry point for all the clients [22].

3.3 Organization of the Cloud Computing Environment

The architecture of a Cloud Computing platform as depicted in Figure 2 is usually common to most e-Learning approaches on the Cloud. In the rst layer we can observe the interface with the Cloud environment, which consists in several management subsystems for determining the current necessities of the user in terms of computational resources, being these the planner for the storage services, the management for distribution of the execution load among the virtual machines, a system administrator to monitor and to initiate activities of each layer, and a security component to ensure the privacy, recovery, integrity and security of user data and transactions, among others. The second layer represents the virtual machines implemented within the system. Finally, the third layer includes all the physical architecture of the system.

Fig. 2 Overview of a cloud architecture for e-Learning

An Overview of E-Learning in Cloud Computing

43

Additionally, Liang and Yang describe in [11] the functions used in the Cloud IaaS and Saas which must be expected for developing such a system. These features can be observe in Figure 2 and are enumerated below: From the IaaS perspective: 1. 2. 3. 4. Storage management for the learning system and the users. Load Balance for all learning systems. Scaling management for virtual machines. Backup and Restore for the learning applications.

From the SaaS perspective: 1. Application Registry management for the commercial provides to register their applications. 2. Application Server for managing and deploying the subscribed learning contents to the users. 3. Account manage system for the authorized users. 4. Virtual Desktop Deployment for providing the personalized desktop including the subscribed learning contents. 5. Session Management for ensuring the Virtual Desktop used by the authorized user. 6. Personalized management for managing the subscription of the favorite learning contents.

Fig. 3 The Architecture of the Virtual Personalized Learning Environment [11]

3.4 Applications of Cloud Computing for E-Learning

We must emphasize the necessity on setting the basis for a educational information infrastructure to alleviate the issues enumerated on the previous section. As we pointed out along this contribution, Cloud Computing may promote a new era of learning taking the advantage of hosting the e-Learning applications on a cloud and following its virtualization features of the hardware, it reduces the construction and maintenance cost of the learning resources.

44

A. Fern ndez et al. a

At the present, the combination of cloud technologies and e-learning has been scarcely explored. Some relevant efforts to use IaaS cloud technologies in education focuses on the reservation of Virtual Machines to students for an specic time frame [21]. Another example of application that can be found in the specialized literature is BlueSky [4], whose architecture has several components aimed at the efcient provision and management of the e-Learning services, being able to pre-schedule resources for the hot contents and applications before they are actually needed, to safeguard the performance in concurrent access, although no details have been found with regard to how this is achieved. On the other hand, CloudIA [19] is a framework which provides on-demand creation and conguring of VM images so that the students are able to have their own Java servlet environment for experimentation, containing MySQL, Tomcat, PHP, and Apache web server. With this approach, students can focus more on developing, deploying and testing their applications in a servlet container. In [11], the authors present a new service model that enhances the efciency within a virtual personalized learning environment. This system is intended for subscribing the selected learning resources as well as creating a personalized virtual classroom, and allows the learning content providers to registry their applications in the server and the learners integrate other internet learning resources to their learning application pools. Other proposals for personal and virtual learning interact with services that rely on the cloud, such as YouTube or GoogleDocs [1]. Finally, we may nd some cloud-related works for performing a comparison on the efciency of online models versus traditional models [7]. The most representative work is this area is developed in [20], where the authors focused on the impact of supporting technologies or the perceived ease of use and acceleration of the learning process. Furthermore, they analyze the appropriate level of abstraction (i.e., IaaS or PaaS) that should be delivered to students to enable them to focus on the course topics.

4 Concluding Remarks
In this work we have exposed the main components of e-Learning, focusing on the exibility, convenience, easy accessibility, consistency and repeatability of this kind of systems. In this manner, an E-learning system is facing challenges of optimizing large-scale resource management and provisioning, according to the huge growth of users, services, education contents and media resources. We have settle the goodness of a Cloud Computing solution. The features of the Cloud Computing platform are quite appropriate for the migration of this learning system, so that we can fully exploit the possibilities offered by the creation of an efcient learning environment that offers personalized contents and easy adaptation to the current education model. Specically, the benets considering the integration of an e-Learning system into the cloud can be highlighted as good exibility and scalability for the resources, including storage, computational

An Overview of E-Learning in Cloud Computing

45

requirements and network access; together with a lower cost considering the payper-use billing format and the save in new hardware and machines and software licences for educational programs. Finally, we have enumerated several approaches that have been already proposed for addressing e-Learning on Cloud Computing, describing these models and how they take advantage of this environment to enhance the features of the educational system. However, we must stress that these are just initial steps towards an open line for research and exploitation of e-learning and cloud computing platforms.

References
1. Al-Zoube, M.: E-learning on the cloud. Intl. Arab Journal of e-Technology 1(2), 5864 (2009) 2. Al-Zoube, M., El-Seoud, S.A., Wyne, M.F.: Cloud computing based e-learning system. Intl. Arab Journal of e-Technology 8(2), 5871 (2010) 3. Buyya, R., Broberg, J., Goscinsky, A.: Cloud Computing: Principles and Paradigms. John Wiley and Sons (2011) 4. Dong, B., Zheng, Q., Qiao, M., Shu, J., Yang, J.: BlueSky Cloud Framework: An ELearning Framework Embracing Cloud Computing. In: Jaatun, M.G., Zhao, G., Rong, C. (eds.) Cloud Computing. LNCS, vol. 5931, pp. 577582. Springer, Heidelberg (2009) 5. Duer, W.: CRM, Customer Relationship Management. MP editions (2003) 6. Ercan, T.: Effective use of cloud computing in educational institutions. Procedia - Social and Behavioral Sciences 2(2), 938942 (2010) 7. Hu, Z., Zhang, S.: Blended/hybrid course design in active learning cloud at south dakota state university. In: 2nd ICETC, vol. 1, pp. V1-63V1-67 (2010) 8. Hurwitz, J., Bloor, R., Kaufman, M., Halper, F.: Cloud Computing for Dummies. Wiley (2010) 9. Jolliffe, A., Ritter, J., Stevens, D.: The online learning handbook: Developing and using Web-based learning. Kogan Page, London (2001) 10. Kwan, R., Fox, R., Chan, F., Tsang, P.: Enhancing Learning Through Technology: Research on Emerging Technologies and Pedagogies. World Scientic (2008) 11. Liang, P.-H., Yang, J.-M.: Virtual Personalized Learning Environment (VPLE) on the Cloud. In: Gong, Z., Luo, X., Chen, J., Lei, J., Wang, F.L. (eds.) WISM 2011, Part II. LNCS, vol. 6988, pp. 403411. Springer, Heidelberg (2011) 12. Masud, A.H., Huang, X.: ESaaS: A New Education Software Model in E-learning Systems. In: Zhu, M. (ed.) ICCIC 2011, Part V. CCIS, vol. 235, pp. 468475. Springer, Heidelberg (2011) 13. Mayer, R., Clark, R.: E-Learning and the Science of Instruction: Proven Guidelines for Consumers and Designers of Multimedia Learning, 3rd edn. Pfeiffer (2011) 14. Ouf, S., Nasr, M.: Business intelligence in the cloud. In: IEEE 3rd International Conference on Communication Software and Networks (ICCSN 2011), pp. 650655 (2011) 15. Papazoglou, M., Van Den Heuvel, W.J.: Service oriented architectures: Approaches, technologies and research issues. VLDB Journal 16(3), 389415 (2007) 16. Pocatilu, P., Alecu, F., Vetrici, M.: Measuring the efciency of cloud computing for elearning systems. W. Trans. on Comp. 9, 4251 (2010) 17. Romero, C., Ventura, S.: Educational data mining: A review of the state of the art. IEEE Transactions on Systems, Man, and CyberneticsPart C: Applications and Reviews 40(6), 601618 (2010)

46

A. Fern ndez et al. a

18. Sosinksy, B.: Cloud Computing Bible. John Wiley and Sons (2011) 19. Sulistio, A., Reich, C., Doelitzscher, F.: Cloud Infrastructure & Applications CloudIA. In: Jaatun, M.G., Zhao, G., Rong, C. (eds.) Cloud Computing. LNCS, vol. 5931, pp. 583588. Springer, Heidelberg (2009) 20. Vaquero, L.M.: Educloud: Paas versus iaas cloud usage for an advanced computer science course. IEEE Transactions on Education 54(4), 590598 (2011) 21. Vouk, M., Averitt, S., Bugaev, M., Kurth, A., Peeler, A., Shaffer, H., Sills, E., Stein, S., Thompson, J.: Powered by VCL - using virtual computing laboratory (VCL) technology to power cloud computing. In: 2nd Intl. Conference on the Virtual Computing Initiative (ICVCI), Research Triangle Park, North Carolina, USA (2008) 22. Wheeler, B., Waggener, S.: Above-campus services: Shaping the promise of cloud computing for higher education. EDUCAUSE Review 44(6), 5267 (2009) 23. Witten, I., Frank, E., Hall, M.: Data Mining. Practical Machine Learning Tools and Techniques, 3rd edn. Morgan Kaufmann (2011)