RGridE-LearningPresentation

"RGridE-Learning: The Role of Grid Computing in E-Learning"

Christina Braz, 2005 Page 1 of 36 Released 03/10/2005 7:18 PM RGridE-LearningV5.pdf

"RGridE-Learning:

THE ROLE OF GRID COMPUTING IN E-LEARNING"

Date: 3-Oct-05

Summary

This document presents the state-of–the-art regarding the converging field of Grid Computing technology

and e-learning. It addresses how Grid Computing has been employed in wired and mobile (wireless) E-

Learning illustrated here by a diverse spectrum of domains such as Grid Learning Services, Collective

Intelligence Sharing, Semantic Web, and Grid Clients for Mobile Devices.

Reference PROJECT I – "RGridE-LEarning: The Role of Grid Computing in E-Learning"

Course DIC9340 Knowledge-Based Learning Environments

Program of study Ph.D. in Cognitive Computing

Department Computing

Institution Université du Québec à Montréal

Professor Roger Nkambou, Ph.D.

Prepared by Christina Braz


Christina Braz 2005 Page 2 of 36 Released 03/10/2005 7:18 PM RGridE-LearningV5.pdf

TABLE OF CONTENTS

1 INTRODUCTION 2

2 ISSUES AND OPPORTUNITIES 2

3 FUNDAMENTAL CONCEPTS 2

3.1 Grid Computing 2 3.1.1 Definition 2 3.1.2 Grid Components 2 3.1.3 Reasons for Using Grid Computing 2

3.1.3.1 Taking Advantage Of Underutilized Resources 2

3.1.3.2 Parallel Central Processing Unit (Cpu) Capability 2

3.1.3.3 Applications 2

3.1.3.4 Virtual Resources And Organizations For Collaboration 2

3.1.3.5 Access To Additional Resources 2

3.1.3.6 Resource Balancing 2

3.1.3.7 Reliability 2

3.1.3.8 Enhanced Management 2 3.2 E-Learning 2

4 STATE-OF-THE-ART OF LEARNING GRIDS 2 4.1 Definition 2 4.2 A General Portal Framework for Learning Grid 2



4.3 Grid Learning Services 2

4.3.1 Dynamic Service Generation 2

4.3.2 Grid Learning Object 2

4.3.3 Learning Grid Infrastructure 2

4.3.3.1 Semantic And Ontological View Of The Grid 2

4.3.3.2 The Role of the Agents and Networking 2

4.3.3.3 Real-World Content-Rich Environments 2 4.4 Collective Intelligence Sharing 2 4.5 Semantic Grid in E-Learning 2 4.6 Grid for Mobile E-Learning (m-Learning) 2

5 CONCLUSION 2

6 REFERENCES 2



1 INTRODUCTION

The "RGridE-Learning: The Role of Grid Computing in E-Learning" presents the state-of–the-art regarding

the converging field of Grid Computing technology and e-learning. It addresses how Grid Computing has

been employed in wired and mobile (wireless) e-Learning illustrated here by a diverse spectrum of domains

such as Technological Infrastructure on the Grid, Semantic Grid, Collective Intelligence Sharing, and Grid

for Mobile E-Learning. The RGridE-Learning is based in part on seven representative papers from renowned

authors [Nkambou&al04a], [Nkambou&al04b] [Millard&al05a], [Millard&al05b], [Page&al05],

[Pankratius& Vossen99], [Jonquet&Cerri05] in the E-Learning and Grid Computing research communities.

Let us remind you the very basic idea of e-learning that is to create the conditions enabling and facilitating to

improve human knowledge. For that reason, we have been noticing among many other developments huge

efforts from the current e-Learning (including Intelligent Tutorial Systems) and Grid Computing research

communities in order to effectively develop technological infrastructures for the Learning Grid focusing for

example on real-world learning scenarios (e.g. support learner construction of theories or performance of

experiments) [enCOre05], [CoAKTing05], [CombeChem05], [Bachler&al04], [Underwood&al04], and

[Yatchou&al04] among many other developments in this area. In Section 4.3.3.3, we will be describing in

more details about content real-world learning scenarios.

Grid Computing is a set of distributed computing resources available over a Local Area Network1 (LAN) or

Wide Area Network2 (WAN) that become visible to an end user or application as one huge virtual computing

system. The objective is to create virtual dynamic organizations through secure, unlimited power,

information access, synchronized resource-sharing among users, institutions and resources.

E-Learning Grid in turn represents the amalgamation of Grid Computing and E-Learning in which of Grid

Computing functionalities are incorporated into E-Learning systems. E-Learning Grid is a collection of

computational resources on demand to match computational needs through a sort of generic service

matchmaking (e.g. a series of algebra exercises (computational resources) is presented to a learner online in

order to improve her/is mathematics capabilities (computational needs)) on the Web. In fact, we can argue

that E-Learning Grid is an expanded notion of diversified resources provision including data resources,

intelligent agent resources and even human tutorial and mentoring resources.

1 Local Area Network (LAN) is a computer network that covers a relatively small area. Most LANs cover a single building or group

of buildings. A system of LANs can be connected over any distance through telephone lines and radio waves, creating a Wide Area

Network (WAN). 2 Wide Area Network (WAN) is a network of computers connected to each other over a long distance, for example on the Internet.



2 ISSUES AND OPPORTUNITIES

The rapid growth of the Internet and Web has brought an increasing attention in Web-based distributed

computing, and several projects have been developed in the e-Learning domain which intends to take

advantage of the Web as an infrastructure for running distributed and parallel applications.

Secondly, in an era when computers systems carry out more and more of "knowledgeable" work from

individuals and when a significant number of these individuals hold more and more different roles in

different, open (e.g. new agents searching for a learning object (LO) might join and existing ones leave in a

Multi-Agent System (MAS)) and complex systems, the need of mastering a large number of systems and

subsystems of broadly conflicting natures, including specially functionalities and interactions with several

human specialists (i.e. instructional designers, teachers, experts, technicians, operators, etc.) who are

frequently distributed over physical space is critical. In the case of a supervision of wide-ranging networks

(telecommunications, transport and distribution of power, etc.) for example a sort of large-scale application,

the supervision and management of such subsystems are obviously distributed at the nodes of the network,

and require for that large number of technicians to achieve the global objective which is to make the system

work.

Thirdly, those open and complex environments mentioned above have also to provide learning resources for

people on the move that will enable end users anytime, anywhere to download courseware or any other kind

of learning resources such as learning objects on portable digital devices such as personal digital assistants

(PDA), cellular phones, laptops or even tablets PC3. The content built by authors and learners are generally

handled, stored and exchanged in units of learning objects (LOs). Generally speaking, LOs are units of study,

exercise, or practice that may be utilized in a single session, and they are characterized as reusable particles

that may be authored separately of the delivery medium itself and be accessed dynamically (e.g. over the

Web). That's all about nomadic computing4 and information environment which is a heterogeneous collection

of interconnected technological and organizational elements, which allows physical and social mobility of

computing and communication services targeted to users. Social mobility here refers to the ways in which

and the ease with which individuals can move across different social contexts and social roles, and be still

supported by the technology and services [Lytinen&Yoo02]. As society and organizations becomes more

dynamic, individuals adopt multiple social roles at an increased intensity and need their information services

adjusted in a large scale as well (e.g. as a mobile learner might move from one site to another so it is crucial

to maintain her/is current state/profile, that is the management of user mobility).

Fourthly, another important point from [Berstis02] is that “the standardization of communications between

heterogeneous systems created the Internet explosion. The emerging standardization for sharing resources,

along with the availability of higher bandwidth, are driving a possibly equally large evolutionary step in grid

computing”, and also in e-learning domain.

Finally, the use of Grid Computing in conjunction with wired and wireless e-learning will provide basically

an end-to-end high-bandwidth access and a vast range of distributed computing resources to end-users (e.g. a

learner). This integration may be in theory difficult because of the need to achieve various qualities of

3 A Tablet PC is a computer that allows you to write on its Touch screen with a stylus, as you would with a PDA. The screen is,

however, much larger - about the size of a typical notebook screen. In addition to data entry, you can also use your stylus to emulate

many of the things you would ordinarily do with a mouse. The fundamental approach is power to exceed your needs and simplicity

for unparalleled friendliness. 4 Nomadic computing is the use of portable computing devices (e.g. handhelds) in conjunction with mobile communications

technologies to enable users to access the Internet and data on their home or work computers from anywhere in the world.



service5 (QoS) which is becoming progressively more important as networks get more populated and more

refined Internet applications and services get spread out. Grid computing even if it is in its very beginning is

already being effectively employed in many scientific e-learning applications where large amounts of data

have to be handled and/or stored as we can observe in several examples in the Section 4, State of the Art of

Learning Grids. In this way, we can assume that the Web has the potential to be a platform for parallel and

collaborative work as well as a key technology to create a pervasive and ubiquitous Learning Grid-based

infrastructure.

3 FUNDAMENTAL CONCEPTS

This section introduces some important concepts that are useful in understanding the landscape underpinning

the application of GRID computing for e-Learning.

3.1 Grid Computing

The research in Grid computing began in 1990 investigating the design and development of a similar

infrastructure called the “computational power Grid” [Foster&Kesselman99] for wide-area parallel and

distributed computing. The purpose for computational Grids was primarily driven by large-scale, resource

(computational and data) intensive scientific applications that demand more resource than a specific

computer (PC, workstation, supercomputer, or cluster) could supply in a single administrative domain. A

Grid enables the sharing (Figure 1), selection, and aggregation of a wide variety of geographically distributed

resources including supercomputers, storage systems, data sources, and specialized devices owned by

different organizations for solving large-scale resource intensive problems in science, engineering, commerce

and also in e-Learning (e.g. photo-realistic visualizations of a complex body model in real-time and display

the computation result on a remote screen).

“Computational Grids are widely regarded as the next logical step in computing infrastructure, following a

path from standalone systems, to tightly linked clusters, to enterprise-wide clusters, to geographically

dispersed computing environments. Generally speaking, we could consider the Grid as the new enabling

technology to transparently access computing and storage resources anywhere, anytime and with guaranteed

Quality of Service (QoS)” [Bruneo&al03]. Currently, grid computing mostly serves computationally

intensive scientific and enterprise applications and operates on cluster computers6 or supercomputers. The

main differences between grids and usual clusters are that grids connect agglomeration of computers which

do not entirely trust each other, and because of that run more like a computing utility than like a single

computer. In addition, grids usually support more heterogeneous agglomerations than are generally supported

in clusters.

5 Quality of service (QoS) refers to a broad collection of networking technologies and techniques. The goal of QoS is to provide

guarantees on the ability of a network to deliver predictable results. Elements of network performance within the scope of QoS often

include availability (uptime), bandwidth (throughput), latency (delay), and error rate. 6 A computer cluster is a group of loosely coupled computers that work together closely so that in many respects it can be viewed as

though it were a single computer. Clusters are commonly (but not always) connected through fast local area networks. Clusters are

usually deployed to improve speed and/or reliability over that provided by a single computer, while typically being much more cost-

effective than single computers of comparable speed or reliability.



At present, one of the most important implementation of grid computing is the Globus Toolkit Version 3

(GTv.3) with its Open Grid Services Architecture (OGSA). It provides a bundle of services and

specifications which can be integrated separately to form a grid middleware. The component model of GTv.3

is founded on grid services that are in fact Web services7 with particular extensions (e.g. interfaces) for use in

grids. The main idea behind OGSA is to build each of the grid middleware layers as shown in Figure 1 by

utilizing suitable grid services.

3.1.1 Definition

The term "grid" makes use of an analogy to an electrical power grid that is the access to computational

resources should be straightforward as the ordinary access to an electric power grid. It means that the grid

would let users take advantage of processing power off the Internet as without effort as electrical power can

be pulled out from the electricity grid that generates that power to our homes. Additionally, a grid user should

not have to take care of how and where this computational power s/he is presently using comes from.

Grid computing refers to a distributed, high performance computing and data management infrastructure that

integrates heterogeneous resources (e.g. storage, computing and/or communications systems, human

collaborators, etc.) and at the same time offers common interfaces for all these resources using standard and

open protocols and interfaces. It is important to highlight that we should not confound cluster computing with

Grid computing. The former generally contains a static number of processors and resources physically

contained in the same or fixed locations, which can be interconnected together. The latter refers to

heterogeneous resources, integrating storage, networking, services and resources. Resources might comprise

machines from different vendors, running various operating systems, and including the capability to control

the workload [Nkambou&al04b].

7 “A Web service is a software system identified by a Uniform Resource Identifier (URI), whose public interfaces and bindings are

defined and described using XML. Its definition can be discovered by other software systems. These systems may then interact with

the Web service in a manner prescribed by its definition, using XML based messages conveyed by Internet protocols” [Aus04].

Figure 1: Layers in a grid middleware [Pankratius&Vossen03].



Another definition of grid computing from [Foster&al01] is viewed like “Grid technologies and

infrastructures as supporting the sharing and coordinated use of diverse resources in dynamic, distributed

“virtual organizations” (VOs)8”. As noted by this author, the main problem that lays the Grid concept is

coordinated resource sharing and problem solving in dynamic, multi-institutional virtual organizations.

Figure 2: A high-level view of the Grid and interaction between

its entities in an application service context [Hoheisel&Der03], [Fraunhofer05].

As we have already stressed here, the Grid can be considered as an integrated computational and

collaborative environment. In Figure 2, we depict the high-level view of activities within the Grid. The users

interact with the Grid Resource to resolve problems, which in turn executes resource discovery, scheduling,

and the processing of application jobs on the distributed Grid resources.

From the end-user (e.g. learner) point of view, Grids may provide the following types of services:

Computational Services [Baker02], Data services [Hoschek&al00], and Application services

[Casanova&Dongarra97].

8 Two examples of VOs: the application service providers, storage service providers, cycle providers, and consultants engaged by a

car manufacturer to perform scenario evaluation during planning for a new factory, or members of an industrial consortium bidding

on a new aircraft.

Internet

Resource Grid

Job Submission

Web Server



3.1.2 Grid Components

The key Grid components of a Grid environment can be viewed in Figure 3: � Portal or User Interface for

user to launch applications that will use the resources and services supplied by the Grid; � A mechanism to

provide security (authentication, authorization, data encryption, etc.); � The Broker identifies the available

and appropriate resources to use within the grid; � Scheduler: Once the resources have been identified, the

next step is to schedule the individual jobs to run on them; � Data Management: if any data – including

application modules – have to be shifted or made accessible to the nodes where an application’s jobs will run,

so there needs to be a secure and reliable method for moving files and data to several nodes within the grid;

� Job and Resource Management: GRAM supplies the services to launch a job on a particular resource,

check its status, and retrieve its results when it is finished.

Figure 3: A high level view of the Grid components [Jacob03].

3.1.3 Reasons for Using Grid Computing

In this sub-section, we highlight briefly from [Berstis02] what grid computing is able to do independently of

e-learning domain whose will be then addressed in Sections 3.2 and 4. We consider that this sort of

detachment gives us the opportunity to better seize the variety and strength of grid computing features and its

relation to e-Learning.

3.1.3.1 Taking Advantage of Underutilized Resources

To illustrate this topic, consider to process an existing application on a distinct machine. The machine on

which the application generally is processed may be busy as a result of a peak in activity and the application

could be processed on an unoccupied machine somewhere else on the grid (e.g. a batch job that uses a

considerable amount of time processing a collection of input data to generate an output collection).

Another utility of the grid is to improve equilibrium resource utilization (e.g. distribute computations and

data transparently across all computers in a grid). In case of unforeseen peaks of activity that normally

demand more resources in an organization, the applications that are related to grid can be moved to

underutilized machines during such peaks.

Virtual Computing Resource

User

Legend: GSI: Grid Security Infrastructure. GASS: Grid Access to Secondary Data. GRAM: Grid Resource Allocation Manager.

Monitoring & Discovery Service (MDS)

�

�

�

�

�

�



3.1.3.2 Parallel Central Processing Unit (CPU)9 Capability

In case of applications that use algorithms, they can be for example split into running parts. A CPU

demanding grid application can be viewed of separately small "subjobs", each running on a distinct machine

in the grid as consequence the application turn into more "scalable" (e.g. a good scalable application would

finish 10 times faster if it uses 10 times the number of processors).

3.1.3.3 Applications

It is important to highlight that not all applications can be converted to execute in parallel on a grid and reach

scalability. There is no automatic conversion and even tools for converting these applications to take advantage of

the parallel capabilities of a grid yet. However, new applications have been already designed for parallel execution

following promising grid protocols and standards.

3.1.3.4 Virtual Resources and Organizations for Collaboration

Real or virtual organizations can share not only files but several other resources such as equipment, software,

services, licenses, and others. These resources are called "virtualized" enabling them more standardized

interoperability among diverse grid users.

3.1.3.5 Access to Additional Resources

Bigger quantities of other resources and special equipment, software, licenses, etc. can be accessed on the

grid (e.g. a user would like to raise her/is total bandwidth to the Internet to put into operation a data mining10

search engine, the job can be split among grid machines that in turn have separated connections to the

Internet). Moreover, the grid allows more sophisticated access, possibly to remote medical diagnostic and

robotic surgery tools with two-way interaction from a distance.

3.1.3.6 Resource Balancing

The grid enables a larger total virtual resource through the contribution of single machines (e.g. (i) an unforeseen

peak load can be forwarded to quite unoccupied machines in the grid; (ii) a lower priority job can be suspended

and executed again later to give room for a higher priority one).

3.1.3.7 Reliability

Let us consider power supplies and cooling systems that are operated on distinctive power sources that can

fire up generators if service power is cut off. As the systems in a grid can be quite inexpensive and

geographically scattered, a power failure at one location will not affect the other parts of the grid (e.g. grid

management software can automatically resubmit jobs to other machines on a grid when a breakdown is

identified).

9 Central Processing Unit (CPU) is the brains of the computer. Also referred to simply as the processor or central processor, the CPU

is where most calculations take place. In terms of computing power, the CPU is the most important element of a computer system. 10 Data Mining, also known as knowledge-discovery in databases (KDD), is the practice of automatically searching large stores of

data for patterns. To do this, data mining uses computational techniques from statistics and pattern recognition.



3.1.3.8 Enhanced Management

Administrators of IT departments can alter for example policies to better assign resources. The grid provides

management of priorities among various projects. Previously, each project in an organization might have

been responsible for its own IT resource hardware and the costs related with it; this hardware might be

underutilized while another project encounters problems requiring more resources caused by unforeseen

events.

3.2 E-Learning

Electronic Learning (e-Learning) is the distribution of a learning, educational or training material by

electronic means (e.g. Internet, Intranet, CD-ROM, digital video disc (DVD)) using therefore a computer or

electronic device (e.g. mobile phone, PDA, etc.). It uses network technologies to create, deliver, and facilitate

learning, anytime and anywhere and delivers distinct, comprehensive, dynamic learning content in real time,

aiding the development of collaborative communities of knowledge, and also connecting different types of

users such as learners with experts, experts with instructional designers, etc. Moreover, it is a phenomenon

that delivers accountability, accessibility, and opportunity allowing people and organizations to keep up with

the rapid changes that characterize the World Wide Web.

According to [Pankratius&Vossen99], in an e-Learning system, the key players are the learners (i.e. student

or company’s employee/apprentice) and the authors (i.e. teachers or instructional designers) but also trainers

and administrators. Authors conceive content, which is put in storage in a learning management system

(LMS) and in a database as well. That content may be updated, or exchanged with other systems. A LMS is

controlled by an administrator, and it interacts with a runtime environment which is addressed by learners

who might be instructed by a trainer. It is important to highlight that these three components of an e-Learning

system might be logically and physically distributed (i.e. installed on different machines) and provided by

diverse vendors or content suppliers. Now, to make this distribution viable, standards seek to ensure plug-

and-play compatibility [IMSGLOBAL01].

E-Learning systems should provide customization of features to a specific learner’s needs

(e.g. Knowledge-Based Learning Systems that include Intelligent Tutorial Systems). A Knowledge-Based

Learning System is a program that is built to model problem solving skills of humans; it is considered as a

“learning interactive environment. They put further the accent on the simulation of the model than on its

construction. The learner “learns” by modifying the parameters and observing the consequences of her/is

actions in the simulated environment [Nkambou05]. Intelligent Tutorial Systems (ITS) are learning systems

one-to-one (tutor/learner). The goal here is to reconstitute the behaviour of an intelligent tutor in order to

provide a personalized education to the learner.

As a matter of fact, the learners can diverge considerably in several aspects such as their prerequisites,

their abilities, their goals for dealing with a learning system, their rate and way of learning, and the time and

money the learner is able to spend on learning. Hence, an e-learning system is generally able to supply

and offer content for all those groups (e.g. a student who would like to learn about database concepts or for a

company employee who would like to grasp company processes and their execution). In order to implement

this system, a learning platform needs to encounter some of the most important requirements such as

personalization, customization and adaptation, the integration of a multiplicity of materials, the

“responsiveness” of the system towards the user (e.g. a "troublemaker" agent can provide pedagogical

interventions of the system [Mengelle&Frasson96]).

As we have previously mentioned, in an e-learning context, learners and authors interact through units of

learning objects that can be accessed dynamically (e.g. over the Web). These learning objects can be stored in



a database as well as any other information pertinent to a learning system such as user profile (personal data,

learner profile), course maps, LO sequencing or presentation information. “E-learning consists of a

multiplicity of complex activities such as content authoring or learner tracking and administration which

interact with resources (including people such as learners and authors), with one another (some activities

trigger others), and with the outside world (such as existing software systems) in a predefined way”

[Pankratius&Vossen99].

We consider the e-learning standpoint of [Cerri05] an excellent and pertinent one. The author affirms that e-

learning according to own experience and also the research in the domain, that e-learning is really not an

electronic variation of traditional Education. Actually, the environment of the education is completely

different, and also the cognitive and social attitude of humans. E-Learning is in this way “not an application

of technologies to human learning, in the sense that assuming to know what to apply (the technologies) and

how (the pedagogy) one puts things together and the result will be a success (people learn). On the contrary,

each serious effort has to be considered unique in the sense that it requires specific technologies and specific

pedagogical principles to be developed and applied in a trial and error fashion within a specific context. This

is the fundamental challenge of e-Learning: services and products have to be combined differently each time,

according to each e-Learning situation”.

In the wireless field, Mobile learning delivered in electronic mobile devices, anytime, anywhere, reveals new

possibilities for technology to augment or facilitate the processes of learning and teaching (e.g. mobile

participants) and also to offer new applications that are not viable with conventional desk-top setups.

Consider the following scenario: a group of students, all equipped with a PDA, that for their Archaeology

spring assignment are working on the Field Trip project. During their activity they store information,

experience, emotion in photos, video clips, text notes, audio comments, etc. The mobile aspect of e-learning

will be detailed in the Section 4.6. Grid for Mobile E-Learning (m-Learning).

4 STATE-OF-THE-ART OF LEARNING GRIDS

In this section, we first introduce some basic concepts of Learning Grids and then we address in a very

segmental way how Grid Computing has been currently employed including aspects of infrastructure,

services and resources such as Grid learning services, semantic Grid in e-learning, collective intelligence

sharing, and Grid for mobile e-learning.

4.1 Definition

Many e-Learning platforms and systems have been developed and commercialized. In general, these

platforms are based on Client-Server, on Peer-to-Peer (P2P), or lately on Web Services architectures, with

effectively significant limitations such as scalability, availability, distribution of computing power, and

storage capabilities. Hence, e-Learning is at this time set out in fields (e.g. sciences, medicine, etc.) where

superior requirements concerning those limitations are not essential. Consider this scenario from

[Pankratius&Vossen99] where e-Learning systems can arrive at their frontier: “a medical school where

anatomy students examine the human body and prepare for practical exercise. Up to now, it is vastly

impossible to compute, say, photo-realistic visualizations of a complex body model in real-time and display

the computation result on a remote screen. With the advanced functionality of an e-Learning grid, students

could be provided with the possibility to grab, deform, and cut model elements (e.g. organs) with the click of

a mouse. Basically as before, the e-learning system could support the learner by giving advice on how to cut

or give feedback for the actions, but beyond that virtual reality at local machine would become possible and

improve the understanding of the subject considerably”.



According to [Nkambou&al05], "the "Learning Grid" refers to the promise of projects that pool together

instructional materials on distant computers. The Grid provides a wide range of available and potential

learning services and resources and does not simply refer to taking advantage of the multiplying effects of

connectivity. It supports the personalized use of the collective intelligence provided by networked computers

and supports the exchange, negotiation and dialogue within and among virtual, evolutionary and pervasive

learning communities" (i.e. collaborative learning that corresponds to human knowledge sharing).

4.2 A General Portal Framework for Learning Grid

An example of a general portal framework for Learning Grid [Yang&Ho05] can be seen in Figure 4.

Considered by the authors as the “Education Grid”, it makes use of the NMI’s Open Grid Computing

Environments (OGCE) Portal framework [OGCE&NMI05] that provides a portal architecture that supports

virtual organizations consisted of scientists and project developers, and also provides the Application

Programming Interface (API)11

for the development of reusable, modular components that may be used to

access the services being developed within the Grid organization. Grid portals enable communication

between grids and the outside world. User portals offer special services to specific members of the public and

researchers.

11 Application Programming Interface (API): is a set of definitions of the ways one piece of computer software communicates with

another. It is a method of achieving abstraction, usually (but not necessarily) between lower-level and higher-level software.

Internet Internet

Remote School A

Remote School B

OGCE Portal

Middleware Globus

Open Grid Services Architecture

Local CAI Platform

Data Web Multimedia (VOD)

Computational Grid Data Grid

Figure 4: A general Learning Grid Architecture [Yang&Ho05].

Computer-Assisted Instruction (CAI)

Video On Demand (VOD)



This architecture (Figure 4) applies Grid Computing technologies to incorporate inactive computer resources

in schools to reduce costs and make efficient use and sharing of applications and resources. Therefore,

schools with restricted budgets can also acquire better services and huge teaching resources using Grid

technology [Yang&Ho05].

4.3 Grid Learning Services

4.3.1 Dynamic Service Generation

The STROBE Model [Jonquet&Cerri05] introduces a social oriented model that is based on interaction

centred of agent representation and agent communication. It represents how an AA (Artificial Agent) might

“learn” dynamically (at run time) at the Data, Control and Interpreter level, especially focusing on

"learning by being told" mode (i.e. use of AAs that learn (by being told) during conversations with other

AAs, therefore that demonstrate a dynamic behaviour that adapts to the context.

The model depicts how agents might execute the interactive, dynamic generation of services on the Grid.

Services here are constructed interactively between agents depending on a conversation. The approach

consists of integrating selected features from MASs and agent communication, language interpretation in

applicative/functional programming and e-learning/human-learning into a simple view that benefits

interactions, including control. “The main characteristic of STROBE agents is that they develop a language

(environment + interpreter) for each of their interlocutors. The model is inscribed within a global approach,

defending a shift from the classical algorithmic (control based) view to problem solving in computing to an

interaction-based view of Social Informatics” [Jonquet&Cerri05].

The kind of MAS employed by STROBE model is as a matter of fact a Multi Artificial and Human Agents

System (MAHAS), a system where AAs and HAs might interact and exchange information and knowledge

effortlessly, where computers might make suggestions to humans and humans to computers, where

collaboration and cooperation is infinite, where an agent might ask to another one to do a task or help it, a

system which might progress dynamically in time and with a nondeterministic behaviour and finally a system

where queries (i.e. problems to solve) and their solutions might come into view through interactions.

Dynamic Service Generation refers to services constructed on the fly by the provider according to the

conversation it has with the user and implies learning, as we mentioned above, interaction. It is a

nondeterministic process depending on the conversation, interaction between two agents. Dynamically

generated services in fact represent a new concept of service involving a collaborative generation of

knowledge (i.e. learning).

We agree with [Jonquet&Cerri05] that when taking into account grid computing, we bring fundamentally to

mind agents and that the grid, they stress, is an evolution of both Web and agent research. According to their

example, we are not able to shift from a Client/Server model based network (e.g. Web) to a distributed

resource sharing system (e.g. grid) without taking into consideration societies of autonomous interacting

agents providing dynamically delivered services (i.e. dynamic services generation) by means of interaction

among AAs and Human Agents (HAs) existing in the society. As the authors emphasize, Learning Grids (i.e.

societies of learning agents) in fact turn into societies of agents (i.e. HAs and AAs) supporting human

learning.



A Meta-Level Learning: "Teacher-Student" Dialogue

According to the authors, the goal of education is to change the interlocutor’s state. Actually this shift is

realized after evaluating new elements carried by the communication. The example in Figure 5 demonstrates

that a STROBE agent can alter its way of perceiving things (i.e. of evaluating messages) by "changing" its

dedicated interpreter while communicating.

Let us consider the following scenario:

The goal of education is to change the interlocutor’s state. This change is completed after evaluating new

elements carried by the communication. The example in Figure 5 shows that a “STROBE agent can change

its way of seeing things (i.e. of evaluating messages) by "changing" its dedicated interpreter while

communicating” [Jonquet&Cerri05]. Actually it is a typical "teacher-student" dialogue. An agent

teacher requests to another agent student to broadcast a message to all its correspondents.

Nevertheless, student does not initially know the performative12

used by teacher. As a result,

teacher conveys two messages (assertion and order) explaining to the student the way of

processing this performative by changing the function which interprets the messages (evaluate-

kqmlmsg). In the end, teacher expresses again its query to student and gets then satisfaction. The

dialogue occurring in the experimentation is described Figure 5. After the last message procedure, the

student function devoted to the evaluation of message (evaluate-kqmlmsg) is modified. The

corresponding code in its environment dedicated to this conversation is changed. Then student agent can

process broadcast messages sent by the teacher.

12 A performative states an agent intention, for example, broadcast, assertion, order, etc.).

Figure 5: Learning of the performative broadcast learning teacher-student dialogue.



The authors emphasize the importance of seeing the learning process as co-construction of knowledge. “An

interaction between two entities is a process that implies an action to occur on the interacting entities. That

means that interactions have some consequence on these entities, i.e. changes of state. However, each entity

may only change state if the change is performed by the entity itself. For HAs, these changes can be

“learning”, the definite purpose of ITS. In e-learning scenarios, it is quite unlikely that human learning occurs

on the simple basis of interacting with a static system. Real interactions modify both entities, including the

artificial one. The learning process should be seen as a co-construction of knowledge (social constructivism).

For this process to be cumulative (i.e. unlimited) in knowledge production, the two entities have to learn from

each other during this process and re-inject what they learn into the loop [Jonquet&Cerri05].

4.3.2 Grid Learning Object

In [Pankratius&Vossen03], e-learning grid architecture is proposed including a Learning Management

System (LMS) and a Core Grid Middleware (CGM). However, what really differentiates it from others e-

learning grid architectures is a new concept of a “Grid Learning Object” (GLOB) for using the grid in e-

learning applications

The LMS and the CGM which are based on Web Services and Grid services are depicted in Figure 6. The

LMS interacts transparently with the CGM hence a learner is not conscious of the grid, and all s/he needs is a

Java-enabled Internet browser to use both the LMS and the CGM.

Core Grid Middleware (CGM)

The CGM in fact implements several layers according to below:

1. The Fabric layer: it is implemented as a Java applet which offers the same interfaces to all

resources in the grid. The user accesses a Web page with her/is Web browser to authenticate in the

grid.

2. The Connectivity layer: it refers to the Grid Login service that performs all access

control operations to the grid middleware.

3. The Resource layer includes an Information service which is aware of the status and type

of all resources in the grid.

4. The Collective layer contains a Broker which implements a grid scheduling algorithm and

also is in charge of distributing computations and data across the grid.



Figure 6: Architecture of an E-Learning Grid including the Core Grid

Middleware and the Learning Management System [Pankratius&Vossen03].

Link CGM/LMS Link CGM/LMS Link CGM/LMS

GLOB



Learning Management System (LMS)

The LMS in general manages all learning activities. A learner who uses a PC for a learning session interacts

only with the LMS and as we have already mentioned is not conscious of a grid in the background. The LMS

provides grid-content or not grid-content functionality related. It is important to mention that all Web

Services of the LMS are accessed via Web pages, in this way the learner only needs a Web browser to use the

LMS.

Since the learner is logged in and authenticated, s/he can access a Web page for course management (i.e.

functionality implemented in a Course Mgmt service). An Ontology service supports the semantic

search for courses. After that the authoring service offers an environment to create, edit, and publish e-

learning content in the ContentRegistry, in this way they can be discovered by the LMS.

The Web services that provide e-learning content is comprised of three essential components: the

learning object (e.g. usually a lesson) or a course comprising of many learning objects; the

assessment element which defines online tests and finally the metadata for search engines that details

the content in a standardized mode.

Other services also are provided by the LMS such as Discussion boards, Chat rooms where learners can

interact with instructors or other learners and ask questions, a Progress Monitor and an Accounting

service.

Integration of Grid Middleware and LMS

The LMS Login Service allows the e-learning PC to become a resource in the grid. As soon as the

learner authenticates her/him on the Web page which in turn is connected with the Login Service of the

LMS, the Fabric layer applet of the grid can be transported as mobile code and be set off locally on the

e-learning PC. So this makes possible the communication with the grid.

Grid Learning Object (GLOB)

A Grid Learning Object (GLOB) is an advanced version of the conventional learning object with grid

functionalities. In Figure 7, we can examine the structure of a GLOB which was designed by the authors in

order to include both traditional e-learning content and content that makes use of grid functionalities. The

GLOB is wrapped by a Web Service which enables it be effortlessly integrated into the LMS (Figure 6).

Moreover, the Web Service offers procedures to access specific parts of the GLOB, to convert content (e.g.

from XML to HTML), or to produce online tests. The GLOB is compounded of several parts: An

Overview (a lesson), metadata (used to find GLOBs), many reusable information objects

(RIOs), and a summary. The User Interface may be implemented as a Java Applet that coverts user

input (e.g. mouse clicks) into tasks for the grid service in the application layer (e.g. a query to recalculate a

3D model).



Figure 7: The structure of a Grid Learning Object (GLOB) [Pankratius&Vossen03].

According to the authors, e-learning management systems and Grid computing can effectively work together

especially for applications or learning scenarios where superior computational power is demanded or even

the tool sets on which learning should be carried out are too high-priced to be granted to each and every

learner. Future issues to be investigated consist of for example transactional guarantees for service executions

over a grid (e.g. atomicity or recovery protocols that assist re-establish an operational state after a grid

breakdown).

4.3.3 Learning Grid Infrastructure

This sub-section presents the Learning grid services and also the functional requirements of a Learning Grid

infrastructure.

Learning Grid Services

In the Proceedings of the First Workshop on Grid Learning Services [GLS04], several approaches were

presented in order to develop a technological infrastructure for the Learning Grid. Contributions realized by

participants were grouped into three categories which describe the participants’ standpoints:

4.3.3.1 Semantic and Ontological View of the Grid

The first standpoint into this category refers to a Service oriented model which requires the use of semantic

tagging for the recognition of and service to individual users (personalization) [Allison&al04].

The actual content of a unity of study.

Generate online exercises for the learners.

Generate online tests for final exams.



The second one refers to the vision of the Grid to Semantic Web technologies which presented a human

centered approach to e-learning. One of the strategies of this approach is the use of hypertext and knowledge-

based tools to augment the capacities of collaborative mediated spaces such as (e-Science) [CoAKTing05]: i)

Ontologies to enhance group discussions; ii) Knowledge-based planning and task support to enhance

process/activity discussions; iii) Scholarly discourse and argumentation to enhance collaborative meeting

activities; iv) Presence and visualization to enhance group peripheral at a distance.

The third standpoint propose a Grid Learning which can be personalized13

to an individual learner using

Semantic Web techniques applied to resources, learners’ characteristics and content categorization that are

indispensable to learners and teachers [Razmerita&Gouardères04]; this approach defines Semantic Web as “a

mesh of instructional resources linked in such a way as to be easily computable by machines on a global

scale” [Woolf&Eliot04).The semantic Grid is considered as a means to assist user-centered, personalized,

contextualized and experiential approaches as well [Gouardères&al02].

4.3.3.2 The Role of the Agents and Networking

Intelligent and autonomous agents are a kind of agents which might perform complex tasks for the learners

such as identify errors and misconceptions, recommend a diverse range of learning objects with also a

diverse spectrum of features, support learners to obtain new concepts, accomplish different goals (e.g. as in a

MAS [Woolf&Eliot04], [Roda&al03]), and respond to dynamic aspects of the environment.

The first standpoint refers to a MAS carrying out training and cognitive supervision through a network

distributed training system ASIMIL [Gouardères&al00]. Another MAS called Actor Specification for

Intelligent Tutoring Systems (ASITS) makes use of agents interacting individually with actors (e.g. human,

intelligent agents, etc.) through a common flow of messages (i.e. agents offering diagnoses, advice and

support to users such as learners, instructors, etc.).

Another standpoint concerns an agent representation and communication model derived from a social

approach to accomplish the dynamic generation of services through the interaction of Artificial Agents (AAs)

and Human Agents (HAs) (STROBE Model [Jonquet&Cerri04]). The objective is in fact to enhance HA-

learning (e-learning) by using AA-learning.

Finally, the third standpoint refers to a set of agents that handle computer-grid communication through

devices (Grid-e-Card) [Gouardères&al04].

4.3.3.3 Real-World Content-Rich Environments

The focus here is on a real world content-rich environment where services should be created according to

teachers and learners’ needs providing if possible (and strongly recommended!) real-world content-rich

environments [Allison&al04]. A diverse spectrum of Grid learning projects have been developed such as: i)

e-Qualification process (dynamic classification of users who enter the grid according to their need in their

activity domain [Yatchou&al04]); ii) Project:EnCOrE (building and using an Encyclopaedia of Organic

Chemistry by virtual communities communicating on the Web); iii) Collaborative Advanced Knowledge

Technologies in the Grid project (CoAKTinG Project) that seeks to advance the state of the art in

collaborative mediated spaces for distributed e-Science; iv) CombeChem project [Bachler&al04); v) Live

communication with remote scientists using mobile sensing equipment in the Antarctic (Antarctic Remote

Sensing Project and the Urban CO Monitoring project) [Underwood&al04].

13 On one hand, Personalization refers to make a system suitable for what a particular user needs (teacher or system demand). On the

other hand, customization refers to changing something to make it just right for you (learner).



Functional Requirements for Grid Learning Resources Services

The goals here represent that learning services enhance classical classroom activities and make changes in

teacher practices. The functional requirements are comprised of two basic functionalities that need to be

provided: pedagogical considerations and identification of services. Pedagogical considerations refer to a set

of characteristics that should accomplish those goals according to below:

• Focus on the learner, not the teacher (student-centered methods).

• Computer-mediated teaching is well suited to support and promote constructivism teaching.

• Personalization and customization in order to improve the efficiency of computer interaction with

users and make complex systems more usable.

• Rich-environments to provide closed interactions between entities, humans and computers for

knowledge construction during e-learning sessions.

• Promote more constructivist and learner-centred learning (simulations, multimedia, virtual reality).

Identification of Services

As we have noticed above, with a diverse range of pedagogical factors involved, Grid services must be

provided in the Learning Grid such as Collaboration services (members of a community sharing and

executing tasks to reach a common goal), Communication services (services offered by the OGSA),

Customization services (pertinent curriculum for each learner), Personalization services, Support services,

Learning styles services, Searching services, and finally Qualification services (qualify a resource for a

curriculum, assess the quality of resources (e.g. user comment and rating) and identify learner capabilities

[Vassileva&al99)].

4.4 Collective Intelligence Sharing

The extensive augmentation of information nodes, the diversity of computers in complex networks, the

cognitive overload, and the transactional distance14

[Moore73,93] demand for an appropriate set of learning

services and devices for the Grid. Consequently, find the Virtual Learning Community (VLC) that shares

learner’s centers of interests is not at all an easy task (Figure 8). In [Gouardères&al04], an approach to reduce

cognitive overload and transactional distance for VLC on the grid trough a computer-grid communication

device called “Grid-e-Card” is proposed.

14 Transactional distance refers to the psychological space created between the teacher and the learner. It is a function of two

variables: dialogue and structure. Dialogue refers to the nature and the quality of communications between the teacher and the learner

while structure relates to the rigidity of the course, the organization of the instruction and the teachings' strategies [Gouardères&al04].



The Grid-e-Card, defined by the authors as a VLC auto-organizer device for collective intelligence sharing

on the grid, intends to bring together users according to their signature for collective intelligence sharing in a

social context: knowledge they have acquired, objectives they wish to attain or learning services

corresponding to their requirements [Gouardères&al04]. The methodology employed is based on P2P-agents

that handle user’s electronic portfolio (e-Portfolio) as “knowledge prosthesis” and exploit e-Learning

qualification (e-Qualification) processes as aggregation methods to dynamically assemble people in pertinent

VLCs.

Figure 8: How to find the VLC coping with my interests? [Gouardères&al04].

Generally speaking, e-Qualification refers to a context where an individual or global assessment of human

actors, of architectures or devices takes place. In [Gouardères&al04], that notion is expanded to the Grid to

indicate the following: i) the exploration that is realized to find the best VLC for the user; ii) the iterative

construction of knowledge from an early state of knowledge to an expert knowledge state; iii) the assessment

of the trainee when progressing inside her/is community during debriefing. “The e-Qualification process

helps the self-organization of nodes by the dynamic classification of people who enter the grid

according to their need in their activity domain” [Gouardères&al04].

Figure 9: e-Qualification loops to a user to integrate a VLC [Gouardères&al04].

The Figure 9 shows some of the e-Qualification process that takes place when the Grid-e-Card is plugged

(See Figure 9 for a general overview of the system). From the trainee e-Portfolio, the learner human agent (L)

finds the corresponding virtual learning community (V1, V2, V3 or V4).

LEGEND L Learner human agent V Virtual Learning Community

???



Grid-E-Card and P2P Communication

In the learning process in Figure 10, every user that necessarily belongs to a community, is connected to the

learning grid through her/is Grid-e-Card. Each trainee is characterized by an agent in her/is community. It is

important to highlight that all the members of a community have a piece of their e-Portfolio that is

comparable to other’s one. In fact, that mutual piece is the signature of the VLC. In order to consent to a

trainee integrate a specific community, rules are activated by agents where those rules are settled on the

virtual learning community signature and the trainee e-Portfolio features. As described by the authors, in this

process, several agents are activated according to the basic processes below:

• A user agent is associated to every user Grid-e-Card. It communicates with her/is community and

automatically triggers the e-Qualification process. It will answer to identification and authentication

requirements of the system through a matchmaking dialog which at the end accept a member in the

community or not.

• The process called “matchmaking” evaluates the content of the messages with the user agent and

enables the categorization of agents in VLCs according to the pertinence of their knowledge in the e-

Portfolio or the goal of the new member in the loop. As the agents adopt a social behaviour so that

they are able to reason on the knowledge states of the other members to which her/is is linked, and

take into consideration her/is own knowledge that other agents will would like to share. As the

authors mention, “the basic loop of the e-Qualification process: mapping of peers into a common

VLC tacitly qualify each one in a shared competence group which is the Virtual Learning Group

Communities (VLGC). From a technical point of view agents need to processes in P2P mode in order

to be organized in groups and interact in pairs, and should be mobile due to the exigency of the grid

environment” [Gouardères&al04].

Figure 10: Grid-e-Card: Basic Dynamic View of the system [Gouardères&al04].

LEGEND LA Learning Agent GeC Grid-e-Card S Service K Knowledge B Broadcast

User

Mirroring: Display the state of the newcomers in relation to others.



A Learning Grid scenario can be seen in ASIMIL, an e-Qualification process (Figure 11) application from

the aerospace domain that stands for “A Network Distributed Simulator Training System” [ASIMIL05]. The

experimental framework is simulation-based intelligent. A P2P review process which is executed by

autonomous agents (i.e. knowledge, ergonomic, psychological). Each agent scans separately a common

stream of messages coming from other actors (Human, intelligent agents, physical disposals) (see Table 5).

They perform coalitions to supply a given community of users (instructors, learners, moderators, etc.) with

diagnoses, advice and help among actors in the community. A dedicated P2P Agents architecture for

perception and qualification of erratic user’s behaviours has been constructed which consists of a cognitive

monitoring based on intelligent agents.

Figure 11: The e-Qualification process in ASIMIL [ASIMIL05].



4.5 Semantic Grid in E-Learning

The Semantic Grid refers to applying Semantic Web technologies to the Grid therefore setting off new

avenues for automation. Semantic Web is “a mesh of instructional resources linked in such a way as to be

easily computable by machines on a global scale” [Gouardères&al02]. Let us consider the proposed E-

Learning Grid Infrastructure [ELeGI05] for example in a widespread Grid environment, so that it would be

viable to automatically integrate new services into a local learning environment.

Independently by offering a distributed system of superior performance of computational resources, the Grid

must also enable structured access to the generated data and also an environment within the collaboration can

take place (e.g. meetings between researchers, shared access to experiments, etc.). The Grid nowadays can be

seen as a “composite of computational grid, data grid and collaborative grid functionalities” [Page&al05].

The CoAKTinG (Collaborative Advanced Knowledge Technologies in the Grid) Project [CoAKTing05]

seeks to enhance the state of the art in collaborative mediated environments for distributed e-Science. It

encompasses four tools: instant messaging and presence notification (BuddySpace), graphical meeting and

group memory capture (Compendium), intelligent “to-do” lists (Process Panels) and meeting capture and

replay. These tools in fact are incorporated into existing collaborative environments and via shared ontology

in order to exchange structure, promote improved process tracking and navigation of resources before, after,

and while a meeting occurs.

BuddySpace

BuddySpace [Eisenstadt&al03; Vogiazou&al05] is an Instant Messaging environment with both client and

server functionality lengthened to improve presence awareness. It presents automatic list construction and

intelligent service discovery on the server, and also the graphical visualization of users and their presence

states in an image, geographical or conceptual map according to Figure 12.

Figure 12: BuddySpace showing a virtual organisation and presence indicators, (a) with live/clickable

presence dots superimposed on geographical and office locations, and (b) with the office dots superimposed

on a conceptual map depicting KMi’s research themes as generated from an underlying ontology

[Eisenstadt&al03; Vogiazou&al05].



Compendium

Compendium [CoAKTing05] is a hypermedia software tool for publishing Dialogue Maps (Figure 13) which

are concept networks which structure Issues, Ideas and Arguments in a dialogue, linked to background

multimedia documents and internet resources. It is better defined as a knowledge management environment

for supporting personal/group discussions and memory, merging hypermedia, modelling and mapping skills

[Conklin&al01].

Issue-maps can be used in learning perspectives to for example sum up: background information about a

difficult issue to be tackled, evidence as it is collected and how it is appropriate to issues under debate,

contributions to online discussions forums.

Compendium is usually used as a means of assembling together diverse resources into a common place for

organization and analysis. For example, students, teachers and researchers can make use of Compendium’s

maps to drag and drop multimedia resources onto a map. Also, Open University PhD students are making use

of Compendium as a visual database for managing their literature reviews, as a manner to improve their

research questions, and to support virtual supervision of e-PhD students as well.

Figure 13: Example use of Compendium by an instructional designer to organise issues, ideas and resources

from diverse sources: (1) The key problem to be addressed is framed as a question; (2) open courseware

resources are dropped from a web browser onto the map; (3) an existing course Unit 3 is added in response to

the issue about one of the web resources; (4) a catalogue of resources is created; (5) a relevant email is linked

to as a response to two different questions [CoAKTing05].



I-X Process Panels for Task-oriented e-Learning

The purpose of I-X research project [Tate&al02] is to create a propitious environment for mixed-initiative

(i.e. engaging human and computer agents) synthesis tasks. In the point of view of a user, the main interface

to the I-X is the Process Panel. A Panel introduces to users the present state of the collaboration from their

individual standpoints, and enables them to dissociate activities, improve elements of the plan, delegate

issues, and invoke the automated agents, etc. all these characteristics supporting to shift the whole task

towards a finishing point.

All features of I-X seek to give confidence novice users to develop their own expertise whereas executing

tasks within the context of a distributed virtual environment of shared resources, aid agents and other users of

several levels of expertise.

Meeting Replay

The meeting replay tool [CoAKTing05] enables individuals to rethink the ideas and topics discussed after a

meeting has occurred. Some features implemented are the meeting time, location, attendees, audio/video

recordings, any presentations given (and related Web versions), and argumentation annotation from

Compendium (Figure 14).

Figure 14: The meeting replay tool [CoAKTing05].



Finally, we describe below where the CoAKTing tools can be used in learning grid scenarios:

• BuddySpace (enhanced presence/communication): to create a Virtual Community consisting the

individual learners & teachers and to provide the “social affordance scaffolding”.

• Compendium: to capture collective thinking within a group who are physically distributed and used

to plan, structure, and access other learning resources providing significant interactive and reflective apparatus for the learner.

• I-X Process Panels: to plan and structure learning tasks, goals, and experiments and to provide

mechanism for tracking issues and tasks when part of collaboration in this manner providing critical

task-level support.

• The Shared semantic ontology -> structured metadata from the various tools can be merged with new

material to generate additional services.

4.6 Grid for Mobile E-Learning (m-Learning)

Grid services can be suitable in a mobile context as mobile electronic devices (e.g. PDAs, cellular phones,

laptops, tablets PC, etc.) change their networks (i.e. instable due to discontinuous connectivity and poor

bandwidth) more regularly than desktop installations. In this way, these mobile devices can take advantage

from being able to discover and utilize services that are local to the device (e.g. to use a projector screen to

show information that couldn’t be displayed onto a palmtop screen). Another factor is that mobile devices

also typically possess currently limited resources (i.e. less computing power) than wired devices, and can

take advantage also from the Grid’s characteristic that is to shift computation to a more powerful system.

Moreover, there is a need to deal with the transparency of the service, and the mobility of users demands

huge efforts in the design of a proper middleware [Bruneo&al03].

In the same way mobility can be very useful to e-learning in order to facilitate and enhance the processes of

learning. The former, it would enable learning resources more straightforwardly available to learners and also

teachers. The latter, it would augment learning experience through the use of mobile devices in, for example,

laboratory work and field trips [Braz04]. In this context, mobile devices can be used in place of classical

paper and pen to organize these experiences or gather learner information [Millard&al05b].

Mobile Grid computing in turn consists of providing Grid services anytime, anywhere from mobile devices.

As we mentioned above, mobile devices have low processing power, in addition its battery life is very short

and its screen is very limited in size and quality. All these restrictions bring as a matter of fact significant

advantages of using mobile Grid technology such as “mobile-to-mobile and mobile-to-desktop collaboration

for resource sharing, improving user experience, convenience and contextual relevance and novel application

scenarios. A grid-based mobile environment would allow mobile devices to become more efficient by off-

loading resource-demanding work to more powerful devices or computers [Millard&al05a]”.

In [Millard&al05a], a mobile e-learning client is proposed, “Finesse e-Learning System”, using Grid

technologies, that is, a mobile learning Grid [ELeGI05]. Finesse (Finance Education in a Scalable Software

Environment) is a Web-based collaborative learning and teaching environment for the finance domain.

Learners are able to manage on-line portfolios and buy and sell shares making use of real-time market data.

The objectives of this project were to conceive a set of Grid services that reproduced the functionality of the

initial Finesse, that is, the Finesse Grid Services (FIGS), and also to create a mobile interface for FIGS which

would enable the portfolios accessible via a PDA.



Unfortunately, conclusions from the authors related to the examination of Grid technologies stated that none

of the technologies presently available (white color) or those still in development or may only have partial

releases (grey color) shown in Figure 15 can effectively support Grid clients on a mobile device. According

to the authors the reason why is that “…they make assumptions about the capabilities of their host

environment, for example OGSI.Net will not work with .NET CF and Globus Toolkit 3 assumes too

high a level of Java support”.

Figure 15: Grid Technologies [Millard&al05b].

The authors argue that lightweight implementations such as Mobile OGSI.NET can provide s subset of the

enterprise services, rather than a new lightweight view; even though these lightweight implementations may

be possible to run on a PDA, it is a wholly different interface onto OGSA (the “?” layer in Figure 15), that is

required to support the type of e-learning mobile Grid applications that was just described.

Open Grid Service Infrastructure

Open Grid Service Architecture

[4] [1] [5] [2] [3] [6] [3]

Web Service Resource Framework

Detailed Legend:

[1] Web Services Resource Framework (a PERL implementation of the current WSRF

definition).

[2] Web Services Resource Framework (an initial implementation of WSRF on .NET1).

[3] Globus Toolkit (a collection of services, written in Java, which can be used to

deploy and discover other services).

[4] Open Middleware Infrastructure Institute (an open source and secure Web services

platform for building Grid applications).

[5] Open Grid Service Infrastructure .NET (an implementation of the OGSI.NET

implementation for mobile devices).

[6] OGSI.NET (a container framework that allows .Net applications to access Grid

Services).



Consequently, the only solution found by the authors was to implement the mobile Grid client using

a proxy (i.e. a browser on the PDA) that accessed Grid services running on a Web server as shown

in Figure 16.

Figure 16: Finesse proxy architecture for mobile clients [Millard&al05a].

The authors used Java Server Pages (JSPs) to invoke Java code on a remote Web server. They implement

their mobile client using JSPs. The Finesse Grid Services are deployed on the Grid, and clients are

implemented as a set of Java Beans talking to the Grid and providing suitable responses as JSP pages to

mobile devices. Requests from mobile users are input via the mobile device’s Web interface, these requests

are first handled on the proxy by client beans which then issue proper Grid Services requests to the Grid.

Once the proxy receives responses from the Grid, it generates and serves appropriate pages to the mobile

client [Millard&al05a].

The authors conclude that “...the current set of Grid technologies does not fit well with the loosely coupled

requirements of mobile e-learning and are often too heavy-weight to fit on a mobile device. Unless this is

addressed it will make the emerging e-learning Grid infrastructures inaccessible to mobile devices, and stunt

the development of novel mobile e-learning applications [Millard&al05b]”.

5 CONCLUSION

In this report, we presented the state-of–the-art regarding the converging field of Grid Computing technology

and e-learning. It addressed how Grid Computing has been employed in wired and mobile (wireless)

E-Learning illustrated here by a diverse spectrum of domains such as Grid Learning Services, Collective

Intelligence Sharing, Semantic Web, and Grid Clients for Mobile Devices.

Recently, there have been some very important developments in the Grid and e-learning coming from

research communities such as an increase of a multitude of collaborative e-learning environments and

components, the amalgamation of different technologies and learning theories (e.g. leading toward fusion of

the Grid with P2P networks and at the same time providing a co-construction of knowledge (social

constructivism), the augmentation of the capacities of collaborative learning through Semantic Web

technologies, the arrival of numerous e-learning objects (e.g. grid-enabled learning objects, learning objects,

etc.), and finally huge efforts has been developed to realize effective Mobile Grid Learning.

Deployed Finesse Grid Services (FIGS): Interest, Notebook, Portfolio,

Sharedata, Userdata.

GRID



In our opinion, Grid computing, semantic Grid, Web services technologies and mobility are crucial factors to

be considered when providing learning resources to learners and teachers in an e-learning environment due to

the following:

• Modularity: where services are dynamically coupled at runtime.

• Interoperability: where we notice the standardization of the service interfaces.

• Extensibility: where services can be automatically and easily integrated into a local learning

environment.

• Distributed Knowledge Management: where further functionalities further than a regular SOA can

be provided such as security and state awareness.

• Suitable Trust Services: where the “intelligence of the services” [Nkambou&al04b] in relation to

the Semantics in the description, the discovery, the selection and the composition of services.

• Communication: where new possibilities are offered to users in order to effectively communicate

more easily without obstacles.

• Automatism: where conversational agents dynamically generated Learning Grid Services.

• Compatibility and expandability: where we can build learning virtual organizations for collective

intelligence sharing through the use of an e-Portfolio as an entry point for e-Qualification of a grid

learning service.

• Accessibility: where users can access from simple to complex e-learning resources anytime,

anywhere.

In a nutshell, the use of Grid Computing in conjunction with wired and wireless e-learning will provide

basically an end-to-end high-bandwidth access and a vast range of distributed computing resources to users

such as learners, teachers, instructional designers, etc.



6 REFERENCES

It is important to highlight that several references cited in this report, “RGridE-Learning”, are included in the

papers themselves mentioned in Section 1, while any others references can be found in this section.

Item

Author (s)

Description

1 [ASIMIL05] ASIMIL (2005) ASIMIL is promoted by PROACTe (Promoting European

Education & Training Research & Development) is a service to

communicate work funded by the European Union under the Education

Area of the Information Society (IST) Programme. Institute universitaire

de technologie, Bayonne, Pays Basque (France). Retrieved August 28,

2005 from <http://www.iutbayonne.univ-pau.fr/article246.html>

2 [Aus04] Austin, J. (2004) “Web Services Architecture Requirements”, W3C

Working Group Note 20040211, World Wide Web Consortium.

3 [Baker02] Mark Baker1, Rajkumar Buyya

2 & Laforenza, D.

3 (2002) “Grids and Grid

Technologies for Wide-Area Distributed Computing”. 1School of

Computer Science, University of Portsmouth, Mercantile House,

Portsmouth (U.K.); 2Grid Computing and Distributed Systems Laboratory,

Department of Computer Science and Software Engineering, The

University of Melbourne (Australia); 3Centro Nazionale Universitario di

Calcolo Elettronico (CNUCE), Instituto del Consiglio Nazionale delle

Ricerche (CNR), Area della Ricerca CNR, Pisa (Italy).

4 [Berstis02] Berstis, V. (2002) Fundamentals of Grid Computing, Redbooks Paper,

IBM Corp. Retrieved July 28, 2005 from

<http://www.redbooks.ibm.com/redpapers/pdfs/redp3613.pdf>

5 [Bingham&al05] Bingham, A., Coles, S., Danos, L., Frey, J., Fu, H., Humfrey, N., Lewis,

S., Luck, M., Mansson, R., Meacham, K., Mills,H., Peppe, S., Smith, G. &

Woods,D (2005) Combechem Project. Retrieved September 1, 2005 from

<http://www.combechem.org/>

6 [Braz04] Braz, C. (2004) Academic work of the COMP6751 Human Computer

Interface Design course, based on the article "Musex: A System for

Supporting Children's Collaborative Learning in a Museum with PDAs"

from Yatani, K. Sugimoto, M. & Kusunoki, F. Proceedings of the Second

IEEE Workshop on Wireless and Mobile Technology in Education (WMTE

2004) pp. 109-113. Concordia University, Montreal, Quebec (Canada).

7 [Brown&Burton78] Brown, J.S. & Burton, R.R. (1978) "Diagnostic Models for Procedural

Bugs in Basis Mathematical Skills", Cognitive Science, 2, 155-191.

8 [Bruneo&al03] Bruneo, D., Scarpa, M., Zaia, A. & Puliafito, A. (2003) "Communication

Paradigms for Mobile Grid Users". In Proceedings of the 3rd IEEE/ACM

International Symposium on Cluster Computing and the Grid

(CCGRID.03), Messina University, Department of Mathematics, Messina

(Italy).



9 [Casanova&Dongarra97] Casanova, H. & Dongarra, J. (1997) “NetSolve: A network Server for

Ssolving Computational Science Problems. International Journal of

Supercomputing Applications and High Performance Computing.

10 [Cerri05] Cerri, S. (2005) “An Integrated View of Grid Services, Agents and Human

Learning”, LIRMM: CNRS & University Montpellier II (France).

11 [CoAKTing05] “CoAKTinG (Collaborative Advanced Knowledge Technologies in the

Grid) Project” (2005). Retrieved July 2, 2005 from <

http://www.aktors.org/coakting/>

12 [enCOre05] Keady, M., Shields, C., Szondi, J., Swanson, L., Handley, G., Doig, R.,

Owen-McGee, D., Hellman, E., Keehan, M., Rimmer, S (2005) “Project:

enCOre”, University of Derby (UK) Retrieved September 1, 2005 from

<http://lib.derby.ac.uk/encore/projectteam.html>

13 [ELeGI05] ELeGI (2005) European Learning Grid Infrastructure Project. Retrieved

July 9, 2005 from <http://www.elegi.org>

14 [Foster&Kesselman99] Foster, I. & Kesselman, C. (1999) “The Grid: Blueprint for a Future

Computing Infrastructure”, Morgan Kaufmann Publishers (USA).

15 [Foster&al01] Foster, I., Kesselman, C. & Tuecke, S. (2001) “The Anatomy of the Grid:

Enabling Scalable Virtual Organizations”, International Journal of High

Performance Computing Applications, 15 (3). Retrieved September 1,

2005 from <www.globus.org/research/papers/anatomy.pdf>

16 [Fraunhofer05] Fraunhofer Resource Grid (2005) I-Lab Research Project. Retrieved at

September 11, 2005 from <http://www.fhrg.fhg.de/>

17 [GlobusAlliance05] The Globus Alliance (2005) Open Grid Services Architecture (OGSA)

Retrieved August 8, 2005 from < http://www.globus.org/ogsa/>

18 [GLS04] GLS’04 (2004) Technological Infrastructure for the Learning Grid,

Proceedings of the 1st Workshop on Grid Learning Service (GLS’04),

Maceio (Brazil).

19 [Gouardères&al04] Gouardères, G., Yatchou, R., Nkambou, R., Saber, M. (2004) "The Grid-e-

Card: An Architecture for Collective Intelligence Sharing on the Grid",

Département d'informatique, IUT de Bayonne Université de Pau (France)

and Département d'informatique Université du Québec à Montréal

(Canada).

21 [Hoheisel&Der03] Hoheisel, A. & Der, U. (2003) “An XML-based Framework for Loosely

Coupled Applications on Grid Environments”. In P.M.A. Sloot et al.

(Eds.): ICCS 2003, LNCS 2657, pp. 245–254, 2003.c Springer-Verlag

Berlin (Germany).



22 [Hoschek&al00] Hoschek, W., Jaen-Martinez, J., Samar, A., Stockinger, H. & Stockinger,

K. “Data Management in an International Data Grid Project”. In

Proceedings of the 1st IEEE/ACM International Workshop on Grid

Computing (Grid’2000), Bangalore (India).

23 [IMSGLOBAL01] IMS Global Learning Consortium, Inc. (2001) “IMS Content Packaging

Best Practice Guide”, Version 1.1.2.

24 [Jacob03] Jacob, B. (2003) “Grid computing: What are the Key Components? Taking

Advantage of Grid Computing for Application Enablement”, ITSO

Redbooks Project Leader, IBM Corp. Retrieved July 2, 2005 from <

http://www-128.ibm.com/developerworks/grid/library/gr-overview/>

25 [Jonquet&Cerri05] Jonquet, C. and Stefano A. Cerri (2005) The STROBE model: Dynamic

Service Generation on the Grid LIRMM, CNRS & University Montpellier

II. France Draft paper submitted to AAIJ special issue on Learning Grid

Services.

26 [Lévy05] Levy. P. (2005) "Collective Intelligence" Université d'Ottawa

Chaire de Recherche du Canada en Intelligence Collective. Retrieved May

30, 2005 from <http://www.collectiveintelligence.info/>

27 [Lytinen&Yoo02] Lyytinen, K. & Yoo, Y. (2002) "The Next Wave of Nomadic Computing:

A Research Agenda for Information Systems" Research Department of

Information Systems Weatherhead School of Management, Case Western

Reserve University of Cleveland, OH (USA). Retrieved March 24, 2005

from <http://weatherhead.cwru.edu/pervasive/2001/content/kalle.pdf>

28 [Mengelle&Frasson96] Mengelle, T. & Frasson, C. (1996) "A Multi-Agent Architecture for an ITS

with Multiple Strategies", Université de Montréal, Département

d'informatique et de recherche Opérationnelle, Montréal, Québec

(Canada).

29 [Millard&al05a] David E. Millard, Arouna Woukeu, Feng Tao, Hugh C. Davis (2005)

“Experiences with Writing Grid Clients for Mobile Devices”, School of

Electronics and Computer Science, University of Southampton (UK).

30 [Millard&al05b] David E. Millard, Arouna Woukeu, Feng Tao, Hugh C. Davis (2005) “The

Potential of Grid for Mobile e-Learning”, School of Electronics and

Computer Science, University of Southampton, (UK).

31 [Nkambou&al04a] Nkambou, R., Gouardères, G., Yatchou, R., & Saber, M. (2004) "The Grid

e-Card: An Architecture for Collective Intelligence Sharing on the Grid",

Department of Computing, Université du Québec à Montréal, Quebec

(Canada) and Department of Computing, IUT de Bayonne, Université de

Pau (France).



32 [Nkambou&al04b] Nkambou, R.1, Gouardères, G.

2 & Woolf, B.

3 (2004) “Toward Learning

Grid Infrastructures: Report on the Grid Learning”, Services Workshop

(GLS’04) held during ITS 2004, the Seventh International Conference on

Intelligent Tutoring Systems, Maceio, (Brazil);1Département

d’Informatique, Université du Québec à Montréal (Canada) ; 2Equipe

ISIHM-LIUPPA, IUT de Bayonne (France); 3Department of Computer

Science, University of Massachusetts (USA).

33 [Nkambou05] Nkambou, R. (2005) "Course Notes-DIC9380 Knowledge-Based Learning

Systems", Department of Computer Sciences, Université du Québec à

Montréal (Canada).

34 [OGCE&NMI05] Open Grid Computing Environments (OGCE) Collaboratory and NSF

Middleware Initiative (NMI) (2005) University of Michigan (USA).

Retrieved August 28, 2005 from < http://www.ogce.org/index.php>

35 [Page&al05] Page, K.3, Danius T. Michaelides

3, Simon J. Buckingham Shum

1, Yun-

Heh Chen-Burger2, Jeff Dalton

2, David C. De Roure

3, Marc Eisenstadt

1,

Stephen Potter2, Nigel R. Shadbolt

3, Austin Tate

2, Michelle Bachler

1, and

Jiri Komzak1:

1 KMI, The Open University, Milton Keynes (UK),

2AIAI,

University of Edinburgh, Edinburgh (UK)3 ECS, University of

Southampton, Southampton (UK) (2005) “Collaboration in the Semantic

Grid: a Basis for e-Learning”. Retrieved August 28, 2005 from

<http://www.aktors.org/coakting/>

36 [Pankratius&Vossen99]

Pankratius, V. & Vossen, G.(2003)Towards E-learning Grids: Using Grid

Computing In Electronic Learning in Proceedings of IEEE Workshop on

Knowledge Grid and Grid Inteliigence (in conjunction with 2003

IEEE/WIC International Conference on Web Intelligence & Intelligent

Agent Technology), AIFB Institute – University of Karlsrube and

Department of Information Systems - University of Münster (Germany).

37 [Pankratius&Vossen03] Pankratius, V., Vossen, G. (2003) Towards E-Learning Grids: Using Grid

Computing in Electronic Learning. In Proceedings IEEE Workshop on

Knowledge Grid and Grid Intelligence (in conjunction with 2003

IEEE/WIC International Conference on Web Intelligence/Intelligent Agent

Technology), pages:4-15, Santa Mary's University, Halifax, Nova Scotia

(Canada).

38 [Piaget72] Piaget, J. (1972) "Development and Learning". In Lavatelly, C. S. &

Stendler, F. "Reading in Child Behavior and Development", Hartcourt

Brace Janovich, New York, NY (USA).

39 [SunMicrosystems05]

Sun Microsystems, Inc. (2005) The Sun™ Grid Solution: Deploying Grid

Computing for Competitive Advantage, Executive Brief, Santa Clara, CA

(USA).



40 [Tate&al03]

Tate, A. (2003) <I-N-C-A>: an Ontology for Mixed-initiative Synthesis

Tasks. In Proceedings of the Workshop on Mixed-Initiative Intelligent

Systems (MIIS) at the International Joint Conference on Artificial

Intelligence (IJCAI-03), Acapulco (Mexico).

41 [Tian03] Hongfei Tian (2003) “Grid Computing as an Integrating Force in Virtual

Enterprises”, Department of Civil and Environmental Engineering,

Massachusetts Institute of Technology (MIT) (USA).

42 [Vygotsky86] Vygotsky, L. (1986) Socio-Cultural Theory, University of Colorado at

Denver, School of Education. Retrieved August 18, 2005 from

<http://web.archive.org/web/20010604093446/carbon.cudenver.edu/~mry

der/itc_d ata/soc_cult.html

43 [Weiss99] Weiß, G. (1999) "Multiagent Systems – A Modern Approach to

Distributed Artificial Intelligence" The M.I.T. Press Cambridge,

Massachusets (USA) and London, England (UK).

44 [Wooldridge99] Wooldridge, M. (1999) "Intelligent Agents" The M.I.T. Press Cambridge,

Massachusets (USA).

45 [Yang&Ho05] Yang, C.T. & HO, H.C. (2005) “An e-Learning Platform Based on Grid

Architecture” at Journal of Information Science and Engineering, 21, 911-

928 High-Performance Computing Laboratory, Department of Computer

Science and Information Engineering, Tunghai University, Taichung

(Taiwan).

46 [Yatchou&al04] Yatchou R., Gouardères G. and Nkambou, R. (2004) Ubiquitous

Knowledge Prosthesis for Grid Learning Services. In: Proceedings of the

1st Workshop on Grid Learning Service (GLS’04), Maceio (Brazil).

RGridE-LearningPresentation

Documents

learning grid infrastructure

grid learning services

grid learning object

mobile elearning

role of grid computing

elearning date

semantic grid

elearning christina