E-Learning Distributed Cloud Built on MVC Design Patterns for Service Task Management

Abstract

Cloud computing is prevalently influencing the present day business trends and operations. Due to Cloud's enticing features, that can be measured by the ease in which the services are delivered to a range of consumer demands without changing the underlying technological implementations or the economics associated with the systems. The e-Learning arena is also not an exception to distance itself away from Cloud driven computing technology. The key point in the e-Learning environment is to process any type of diverse problem within a moderate response time across ubiquitous network channel and securely deliver the output to the authorized user on the most user-friendly interface. The bottom-line of our research is to provide a Software Development Platform enabling the clients to specify and design their task that can be hosted as a service through the Cloud which is interfaced by Web services technology. We are proposing an e-Learning enterprise Cloud based on the Model-View-Controller design patterns to expose the reusable View and Controller tiers as Distributed Cloud, each one wrapped individually in Web services. The Controller Cloud implements any task as a configurable entity, which is individually exposed as a service, and/or simultaneously uses third party Cloud services to execute the most complex problems. The loose coupling between View and Controller potentially benefits to baseline and manipulate the desired level of system's Quality of Service in terms of performance, scalability and security.

Full text

E-Learning Distributed Cloud Built on MVC Design
Patterns for Service Task Management
Ruth Cortez
University of Aizu
Aizu-wakamatsu, Fukushima
965-8580, Japan
m5132203@u-aizu.ac.jp
Sidhant Rajam
University of Aizu
Aizu-wakamatsu, Fukushima
965-8580, Japan
m5132202@u-aizu.ac.jp
Subhash Bhalla
University of Aizu
Aizu-wakamatsu, Fukushima
965-8580, Japan
bhalla@u-aizu.ac.jp
Alexander Vazhenin
University of Aizu
Aizu-wakamatsu, Fukushima
965-8580, Japan
vazhenin@u-aizu.ac.jp
ABSTRACT
Cloud computing is prevalently influencing the present day
business trends and operations. Due to Cloud's enticing features,
that can be measured by the ease in which the services are
delivered to a range of consumer demands without changing the
underlying technological implementations or the economics
associated with the systems. The e-Learning arena is also not an
exception to distance itself away from Cloud driven computing
technology. The key point in the e-Learning environment is to
process any type of diverse problem within a moderate response
time across ubiquitous network channel and securely deliver the
output to the authorized user on the most user-friendly interface.
The bottom-line of our research is to provide a Software
Development Platform enabling the clients to specify and design
their task that can be hosted as a service through the Cloud which
is interfaced by Web services technology. We are proposing an e-
Learning enterprise Cloud based on the Model-View-Controller
design patterns to expose the reusable View and Controller tiers
as Distributed Cloud, each one wrapped individually in Web
services. The Controller Cloud implements any task as a
configurable entity, which is individually exposed as a service,
and/or simultaneously uses third party Cloud services to execute
the most complex problems. The loose coupling between View
and Controller potentially benefits to baseline and manipulate the
desired level of system's Quality of Service in terms of
performance, scalability and security.
Categories and Subject Descriptors
K.3.1 [Computers and Education]: Computer Uses in Education
– Collaborative learning, Distance learning.
General Terms
Design, Human Factors and Verification.
Keywords
Task Management, e-Learning, Web services, Cloud Computing,
XML, MVC Design Patterns.
1. INTRODUCTION
Traditional classroom methodology usually consists of a set of
activities that teacher should perform to transfer knowledge to
students and develop the skills to apply this knowledge. Under
traditional scheme teachers find difficult to focus on individual
needs. Students usually play a passive role and teacher's effort is
mostly concentrated on course organization, preparation, delivery
and scoring assessments. Eventually, this might lead into
resources challenges when education demand increases in a
higher pace than the resources of education providers (e.g.
professors, teachers, facilities, etc). The use of Web technologies
such as multimedia, VoIP, chat, blogs, wikis, computational
processing and the convenience of having the information
available “anytime, anywhere”, are some of the potential
advantages of incorporating e-Learning systems in education, and
by this way overcoming traditional classroom challenges [7].
Accordingly, the activities are transitioned from being teacher
centric to student centric and increases student motivation by
playing a more active role during education process.
It is possible to distinguish several approaches to implement an e-
Learning environment. The main idea of the first approach was
essentially transition contents of traditional class to a Web
application and made available the course information to students,
but paying low attention to the pedagogical work flow [10]. The
second approach is called as a Blended Learning, which
incorporates the usage of e-Learning applications to traditional
classroom activities, using various delivery methods such as face-
to-face lecture, online material for self study, video, audio, tests,
surveys, simulations, workshops, Web seminars, discussion forum
etc. The advantage of Blended Learning is more flexibility
adapting to student’s needs due to the combination of
Synchronous physical formats, Synchronous online formats and
Self-paced Asynchronous formats, a detail description can be

found in [10]. Finally, it introduces tools for student’s
performance analysis and support. The Flex-eL e-Learning
environment [7] incorporates most of the advantages of blended
learning plus the focus on learning process work flow. In this
environment, the course content is divided into modules and
represented graphically for better comprehension. It includes the
usage of collaboration tools making possible, for example, the
ability to identify students that are working on the same activity
and promoting collaboration groups. In this specific example, the
environment is designed to allow students manage their individual
time, more tailored learning, flexible pathways in which the time
constraints depends on student time, all of this towards the idea of
its usage as Virtual University.
The last approach discussed in our review considers the e-
Learning environment as a set of Web services. It incorporates the
main benefits described on previous approaches but envisioning
the environment with the following requirements: open
architecture and interface, durability and reusability, high
interoperability for information exchange, flexibility on design to
facilitate developers adding or removing components, and
compatibility with other systems [15]. As suggested by Dagger
and O’Connor [4], next generation of Learning Management
Systems are envisioning scenarios where it is possible among
different systems to exchange not only data but tools,
functionalities, semantics, and learning workflow. Therefore,
Users will be able to customize their e-Learning platforms from a
diverse range of e-Learning services. To conclude our review, we
would like to point that in both traditional and e-Learning
environment, knowledge has to be divided in meaningful chunks
of information to develop and test a particular skill. The delivery
method triggers the interaction between the teacher and student.
That is why we are focusing on the delivery method enhancing
this interaction based on a Task Management approach.
The main goal of our research is to develop a software platform
that provides facilities to integrate a Task Management issue as a
core e-Learning component that can be applied to different
education domains. This platform will enable clients to specify
and design their Tasks hosted as a service through Cloud
interfaced by Web services technology. Other e-Learning
components can also be added to scale the environment
capabilities, and facilitate the integration with third-party
applications for high interoperability, reusability and information
exchange.
The Cloud computing enables convenient, on-demand network
access to a shared pool of configurable computing resources (such
as networks, servers, storage, applications and services) that can
be rapidly provisioned and accessed with minimal management
effort or service provider interaction. Service oriented Cloud
computing focuses on statelessness, low coupling, modularity and
semantic interoperability [8]. The main essence of Cloud
computing is to significantly virtualize the services and abstract
an implementation logic from service invoker [5]. Cloud
computing provides Quality of Service (QoS) guaranteed
infrastructures, e.g. time, cost, reliability and hardware
performance such as CPU bandwidth and memory consumption
and sustains SLA oriented resource allocation [2].
As Cloud computing is rising as a technological innovation over
the horizon of the computing science [2,8,11], we are undertaking
a research to implement e-Learning systems in the event of the
Cloud computing bolstered by a Web services technology for
interface integration. The Model-View-Controller (MVC) design
patterns are the main source of motivation to architect the domain
of modules and interconnections. The View and Controller are
represented as a Public Distributed Cloud so that both the
modules to be developed independently. In the Distributed Cloud,
every node is an independent Cloud that offers unique service like
View Cloud offers Graphical User Interface (GUI) as a Service
and Controller Cloud offers Task as a Service (TaaS). The
inception of Distributed Cloud leverages the implementation of
the most reliable, scalable and reusable services. The Model
stores the configuration definition schema used by Controller and
hence, the Model resides within a Controller Cloud. The
Controller is further normalized to host a Cloud for a set of
actions executing the desired output in order to outsource those
actions to external Clouds and/or import external Cloud services
as Application Programming Interface (API) for extending and
overloading current scope of operation. Alongside this main
course, we have other features required by e-Learning
environment like access control by user profile, session
management, rule based verification and authorization engine,
alert notification engine, grading or scoring engine, pluggable
problem processing engines like arithmetic or formula engine,
human language engine etc. included in the proposed design.
The rest of this paper is organized as follows. In Section 2, the
Task Management approach is explained along with the main
features that will be generalized to create an abstract model. It
includes an example that will be further developed as a prototype.
Section 3 explains the architecture of Distributed Cloud for e-
Learning. In Section 4, the system design issues are described
including implementation details and security measures. The
conclusion and future work are discussed in Section 5.
2. TASK MANAGEMENT E-LEARNING
ENVIRONMENT
The original idea of Task Management study initially comes from
our previous research [8] based on a comfortable task
management tool to enhance collaboration between teacher and
students, and applied to subjects that require complex formula
verification such as mathematics and physics. The Task
Management environment is implemented under client-server
architecture, where clients are workstations used by teachers and
students. An Exercise server application manages a database of
problem collection, and a set of functions to facilitate problem
preparation, distribution and automatic verification. The formula
engine facilitates the problem design and can automatically create
different versions of the same problem changing initial parameters
input by teacher. Verification engine applies different verification
strategies to evaluate not only numerical answers but also
complex formula expressions applied to subjects such as
mathematics, physics, statistics, chemistry, etc. which we will
refer in this document as Complex Formula Expression Subjects
(CFES).
2.1 Task Management Abstract Model
Scheme
The main structure of the Task Management approach can be
generalized in an Abstract Model (see Figure 1) that will be
referred in this document as Task Management Abstract Model

(TMAM). Learning support is unified under this scheme
facilitating the correspondence to extend into different domains,
and the integration to the e-Learning platform.
The main concepts and classes are defined as follows: Task
management is the process to organize a set of activities or tasks
that need to be executed to achieve a predetermined objective. An
Objective is a set of tasks strongly tied together in specific order.
The completion of all tasks defines the completion of the
objective. A Task is a set of actions that need to be completed to
reach the predetermined objective. An Action is an atomic event,
which can not be further decomposed. A Metric is an element that
collects and summarizes results to measure execution of a
learning objective. This is to be used for statistical analysis and
performance associated with the consumer behavior. An Alert is
an element used for notification about the execution of each task
and change of status task information. A Provider is the user who
owns, creates or maintains all or part of the Task Management
Elements. A Consumer is the receptor of the learning goal. It is
expected to payback with specific response according to the
actions consumed. An Evaluator is the user who does not produce
or own elements. The evaluator’s participation is limited to a
specific objective with a role of monitoring tasks execution and
with rights to score results. Interface is a protocol of
communication integration that individually wraps loosely
coupled peers e.g. Web services. Session is a time window
associated with a user to utilize the task.
2.2 Task Management Model Applied to
CFES
In framework of e-Learning, the TMAM contains the basic
description of the classes that will be transformed into a
correspondence class of the specific domain. A prototype is to be
designed applying this scheme (see Figure 2) to CFES as initial
component of the e-Learning platform. The Objective is
implemented as a project class with Course and Tournament as
the two main variants. The objective can be extended defining
new types and easily integrated under the model structure. A
Course is defined as a set of tasks to be implemented as lectures
and tests. The producer defines prerequisites for each task to
trigger access to next one. It is possible to set more than one task
with the same prerequisite. In this case the order of execution is
optional to the consumer. The Tournament is the second variant
of learning objective. The tasks are implemented as Exams with
different levels of difficulty and a specific order of execution. A
Test can be implemented with three different variants: An
Exercise is an informal evaluation of the acquired knowledge. A
Quiz is a scored test, time bounded and generated with unique
values for each student. An Exam has the Quiz characteristics
with the constraint that the time duration is less than 24 hours. A
Dashboard class corresponds to the performance monitor
implementation and presents real time scores for tournament
participants, as well as specific metrics set by producer. The
Judges are evaluator users with a lifetime in the system strictly
related with the learning objectives in execution.
Fi
g
ure 1. Task Mana
g
ement
–
Abstract Class Model
3. DISTRIBUTED CLOUD TO DELIVER
TASK AS A SERVICE
3.1 Architecture of Distributed Cloud
Figure 3 envisages the architectural view of the proposed
Distributed Cloud based on MVC design patterns concept. The
architecture concentrates on the core task processing work and
therefore the MVC design patterns [6] are used in order to reuse
the loosely coupled components and objects.
Figure 2. Example of Class Model applied to CFES

Figure 3. Architecture of Distributed Cloud for e-Learning
The design has identified the View and Controller as independent
entities in the form of a Cloud which will communicate through
Web services interface with Extensible Markup Language (XML)
protocol over Simple Object Access Protocol (SOAP). The Cloud
implementation is done by using Virtual Machine (VM) Cluster
where a pool of infrastructure resources are load balanced by
hypervisors in vertical as well as horizontal scalability. The
Utility Computing is incorporated in the Controller Cloud as
utility billing is charged with respect to the task's valuation as
decided by the task owner and configuration of which is modeled
in the database in the form of rules or templates. The end user
client is going to be thin as the GUI layer is detached from the
task controller in order to View as an independent Cloud which is
responsible for only request and response message mapping in the
Web Service Definition Language (WSDL) XML interface.
There are two modes of communication: Direct and Discovery.
The Direct mode enables two Clouds to communicate through
pre-shared XML Web services interface over SOAP call. In the
Discovery mode, Universal Description, Discovery and
Integration (UDDI), is used as a service manager wherein
individual Cloud publishes their WSDL interfaces which is
invoked by the client Clouds. In case of direct mode the
individual Cloud has to make sure that the current updated copy
of the XML interface is multicast to the associated third party
Cloud. This overhead is resolved by using UDDI as interface
definition publication is centralized. In the next subsection we
discuss the View and Controller Clouds with respect to the task
management in order to deliver TaaS.
3.2 Controller Cloud
Controller Cloud is the central point in our e-Learning task
management environment that eventually delivers TaaS.
Controller Cloud has two controllers viz. Session Controller and
Task Controller. Task Controller is exclusively identified as a
Cloud within the main Controller Cloud to expose TaaS. The
Task Controller Cloud communicates in XML messages. The
XML response rendered from the Task Controller Cloud can be
represented in any format at the user end. Consequently, the
services of the Task Controller Cloud are exposed in both
Business to Consumer (B2C) and Business to Business (B2B)
modes of interaction. A Model is locally used by Task controller
Cloud to configure the schema definition of the data and task
processing logic rules as well as templates.
3.2.1 Session Controller
The Session controller evaluates and verifies the session of every
task request with respect to the user session configuration as well
as binds the request parameters from View to the corresponding
task service definition's Data Transfer Object (DTO). The Session
controller is a first step whenever user and task authentication is
required prior to the service invocation. Session controller is not
mandatory in case only TaaS is needed by an External Cloud. The
session controller performs following tasks:
User ID Authentication: As e-Learning applications are
vulnerable for re-purposing and other network attacks, the session
controller authenticates the user by using a two level
authentication scheme. If first level authentication (e.g. user id
and user specific credential like date of birth) fails, the user
account is soft-locked (i.e. user account is locked only for short
period of time) and if second level authentication (e.g. user's pin
or password) fails, the user account is hard-locked (i.e. user
account is permanently locked).
User Session Authorization: If the valid user successfully logs
into the e-Learning application, a session ID is assigned to the

user by Session Controller. This session ID is verified (by session
controller) for all subsequent communication between View and
Controller for the given user.
Task Request Binding or Mapping: If the above first two
authentications are successful, Task Request Binder maps the
user’s request parameters (e.g. input parameters in HTTP request)
to the DTO parameters of Task Controller interface, as given in
Figure 4. This mapping is called as View and Controller Task
Request Binding (VCTRB). Hence, this is a message passing
from View to Controller. This binding isolates the Task Controller
from the GUI modifications done at View and therefore, View as
well as Controller becomes the loosely coupled modules.
3.2.2 Task Controller Cloud
In order to observe TaaS, the logic associated with it is wrapped
inside a Web services XML interface in order to reuse export and
import of the task objects from External Clouds. As this Cloud is
identified inside the main Controller Cloud, it is invoked by the
Session controller in the Direct mode. The Task Controller Cloud
manages the task objects in a container across the VM cluster.
Task object management is done by using many components like
Transaction Manager (to control the 2-phase commit XA
connection across the distributed system needed by the Task
Controller), Task Object Factory (to control the object generation
during coordination among actions), Task Object Pool (to control
object store for significant performance boosting), Task Object
Observer (to control the change in object state and notification)
and Task Object Monitor (to control concurrent usage of the same
object for mutual exclusion) [6]. The main components associated
with the Task Controller Cloud are as following:
Task Request and Response Parser verifies and parses the
request received from the VCTRB for the operation of the task
controller. The data value is unmarshalled or deserialized
(converting back to DTO) for the specific task processor [3]. This
parser also serializes or marshals the Controller’s response which
is returned to View after task execution.
Verification Engine performs the answer verification (discussed
in more details in section 4.1) and validation of the given task
request with respect to the request fields, parameter's data type
and width, request specific validations etc. The verification
processor is a configurable engine which is defined by developing
certain rules and using those rules as a template (see Table 1). The
rule or template designed is stored in the Model.
Table 1. Verification Engine: Rule based sample Schema
Authorization Engine is responsible for authorization of a request
associated with the usage of the task by task consumer and task
owner e.g. If a participant (task consumer) of the tournament is
submitting his answer, this request is submitted for the
authorization of a teacher (task owner) or the group of teachers.
The authorization processor is a configurable engine which is
configured by developing rules and templates specific to the user
and/or task. The rules can also be developed according to the
hierarchy of the users (authorizer). These rules and templates are
stored in the Model or task database (see Table 2).
Table 2. Authorization Engine: Rule based sample Schema
Alert Engine triggers the notification services whenever the state
of the task is changed. The task owner as well as consumer is
notified for the task state changes by Short Message Service
(SMS) or E-Mail channels. This is a background asynchronous
process running in the task object container in the form of a timer
thread e.g. after the grade registration by the teacher, the task
status is changed to complete and corresponding notification is
sent to student.
Grading or Scoring Engine keeps track of the student’s progress
report by evaluating the transcript as per the task rule definition
devised by the task owner like teacher. While defining or
publishing a task like Quiz, the teacher has to configure the
grading rule and policy for that task. The student subscribes or
registers for the task like Quiz and student's performance in terms
of grade or score is computed by the Grading engine. If the
authorization rule is enabled for this task, the Grading engine
waits for the final authorization otherwise by default Verification
Figure 4. VCTRB Sample XML protocol structure for
message passing from View to Controller

engine validates the task and according to the result of validation,
the task gets authorized for either failure or success.
3.2.3 Supporting Engines
The Task Controller Cloud implements supporting engines to
abstract certain task processing related functionalities. Such
abstraction enables the reusability and scalability of the task
operation like mathematical query processing task, language
related query processing task etc. The supporting engines are
plugged as add-ons and therefore, any type of task processing can
be hooked to the main controller. As these engines are loosely
coupled with the main controller, any third party Cloud utility can
be interfaced as an add-on to the main controller. The task
controller can use following engines for data processing:
Arithmetic and Formula Engine provides certain helper
operators to perform mathematical operations like integration,
derivation, limits, permutation, addition, subtraction etc.
Human Language Engine processes the grammar of a specific
human language and thereby solves language related problems,
for example, in English language, problems related to the
active/passive voice, direct/indirect sentences, vocabulary etc.
Logical Reasoning Engine processes logic related tasks of
deduction, induction, abduction e.g. finding an odd number from
a given series of numbers etc.
3.2.4 Model
The Model is a local database schema used by the task controller.
The Model stores the schema definition for the task including user
profile, access rights, user relationship like student-teacher
association, problem definition, session control, verification
templates, authorization templates, task rules for grading, alert
notification etc. The Model is preferably a relational database
because the various entities associated with Controller are defined
in a normalized relationship model. The database connection
between Task Controller and database Model is implemented by
either Open Database Connectivity (ODBC) bridge in the form of
a database driver or Object-relational Mapping (ORM).
3.3 View Cloud
The View is a GUI Cloud to construct the presentation look and
feel design and to exchange input and output messages along with
the end client user. The GUI is outsourced from the core task
processing activity and therefore, the View is built autonomously
as per the client requirements. Hence, an identical response
message from Controller Cloud is realized in different views at
the user end. Likewise, the View can support a wide range of
clients including personal computer, mobile phone, Personal
Digital Assistant (PDA), Tablet etc by hosting a website or portal.
The client can access this interface over Ubiquitous channels like
the Broadband Internet, Wireless standards like Bluetooth,
ZigBee, Infrared, Wi-Fi etc. Such loose coupling with the task
processing logic significantly decreases the development,
debugging and troubleshooting efforts as well as time.
Web container is clustered to host the static/dynamic HyperText
Markup Language (HTML) pages, Servlet, Cascaded Style Sheets
(CSS), JavaScript code files, images, movies, flash objects, and
applets etc. to construct the presentation at client end. The
Extensible Stylesheet Language (XSL) transformation technique
is used to generate the HTML output by imposing the XML
response message as received from Controller Cloud onto the
XSL files and applying the XML Schema Definition (XSD) in
order to validate the patterns of the XML response data.
3.4 External e-Learning Cloud
The proposed TaaS Cloud model is scalable in terms of
processing the logic. Task Controller Cloud has Web services
interface to enhance the task execution intelligence by importing
third party Cloud services and at the same time by hosting the
task for third party Cloud for exporting TaaS. Therefore, the
external e-Learning Cloud includes Moodle Course Management
System (CMS), External Task Processing Engines or Processors,
Programming Contest Control Project, External user profile
model hosted as Data as a Service (DaaS) etc. The External e-
Learning Cloud can only communicate with the Task Controller
Cloud either in the Direct mode or in the UDDI mode.
4. DESIGN ISSUES
Complexities associated with the e-Learning environment for
rendering knowledge based online training services, needs to
chalk out rationally. Accordingly, let us discuss few common
issues related to the e-Learning environment with respect to the
proposed research work.
4.1 Answer Verification Strategies
There are three main strategies applied for automatic verification.
This section contains a summary and more detail information can
be found in [13].
Numerical Answer Verification. To evaluate numerical answers,
the verification engine uses the initial values and a precision rate
that teacher defines when the problem is created. It simply
compares the numerical answers of the student and teacher. In
case result is different, it calculates the absolute value of the
difference and performs a comparison with the precision rate,
storing the result for teacher review.
Multiple Answer Verification. This strategy is applied when the
evaluation is performed for different variants of the same
problem. In this case teacher’s answer is stored as a formula. The
engine calculates teacher’s initial parameters using the stored
formula and applies numerical answer verification to evaluate
student’s answer comparing with values generated by teacher.
Formula Answer Verification. This strategy is applied when
student’s answer is an equation system (see Figure 5). Precision
rate and formula are stored in teacher’s Test Data. In this case the
formula verifier composes the student and teacher formula, a
symbolic transformer is applied to simplify the composed
expression. The absolute value of the difference is compared with
the precision rate. If the difference is within the precision rate the
answer is scored as correct, otherwise is scored incorrect. In both
cases the value is stored for teacher review.
4.2 Task Management GUI
The GUI for Task Management environment is a Web interface
that can be supported under any browser. It also includes a simple
interface for PDA that has no embedded Java component.

Teacher Interface. Workstation Teacher Interface provides tools
for problem design (see Figure 6). Teacher can use a formula
editor to customize the equation variables and operators, and
hence offers Multilanguage support. During problem preparation,
the teacher can verify the input data and correct it if necessary.
Student Interface. The student’s GUI interface (see Figure 7) is
generated in correspondence with the problem definition. Formula
Construction Palette shows variables and operators related to the
particular problem is being resolved. If defined by teacher, each
problem might include hint information to show more explanation
to resolve the problem. The result information is automatically
displayed after answer has been submitted by student.
The prototype integrates the formula and verification engines,
which are hosted under the Task Controller Cloud. The GUI is
designed based on the View Cloud to support a wide range of
clients including workstations, mobile phones, PDA, Tablet PCs
etc.
4.3 Security in Cloud
The analysis of the latest market data published, such as the IDC
[14] shows that the security is a prime hindrance to the
advancement and adoption of the Cloud computing driven
economy due to it’s vulnerability to various attacks caused by
unauthorized intrusion, malicious computing, human errors,
privacy, natural disaster etc. The risk associated with security
fulfillment should be jointly borne by both service provider and
service consumer for speedy and hassle free Cloud
implementation because the public channels of communication
like the Internet is not under the control of a single autonomous
body and therefore it is unsafe. This proposes a security technique
which is defined by the consumer of service and implemented by
the provider of service. On the other hand, such security
technique should be flexible enough for customizations as per the
client’s desired level of security on a case by case basis.
Therefore, we propose to devise a security kernel wrapper [1] to
abstract the core Cloud processing engine, for example a
hypervisor in VM cluster, by forming a protective coating around
Cloud engine. The concept of security kernel is inherited from the
operating system kernel. The security kernel is designed for the
basic standard network security measures; it is provisioned to
monitor the flow of data and events of various security operations
according to the user defined customizations and rules. The kernel
acts as a pluggable security view to the core Cloud processing
engine and provides a shield against the network based attacks.
This idea induces a high esteem into the accreditation of the e-
Learning environment by educational institutions over non-trusted
internet channel.
Figure 5. Task Management Formula Answer
Verification Scheme
5. CONCLUSION AND FUTURE WORK
In a nutshell, the proposed research has the potential to be
realized as a Software Development Platform of Distributed
Cloud Computing for implementing Task Management inside the
e-Learning environment with a great ease of development and
maintenance. The common issues related to the Quality of Service
(QoS) such as scalability, performance, and security, can be
handled depending on the desired level of QoS. This alleviates the
bottleneck of systems adoption by various kinds of vendors as the
risk associated with the QoS is born by both service provider and
consumer.
Figure 6. Task Management Teacher Workstation Interface
Figure 7. Task Management Student Workstation
Interface
In the conventional MVC Design Patterns, the View and Model
can have a direct association [6]. By proposing the architecture of
an e-Learning Distributed Cloud, we have re-defined the MVC
Design Patterns by completely decoupling View from Model.
Accordingly, the View is loosely coupled with the Controller

Cloud. Thereby, the presentation logic has been outsourced, and
the developers can be concentrated on computational services
related to e-Learning subject itself. The B2B and B2C interaction
modes leverage an endpoint usability of the e-Learning
computational services that are deployed in the Task Controller
Cloud.
In the proposed model, the Task Controller Cloud delivers the
TaaS. The controller can be enhanced or scaled in order to tackle
any kind of tasks associated with any field of study like human
language exercise, mathematical problems, logical and critical
reasoning, statistical problems, etc. The same is also true for the
View Cloud because it delivers the GUI as a Service. The
comparison with Web services implementation approaches
specified in [15], our approach adds significant improvement in
open architecture, durability, reusability, high interoperability for
information exchange, flexibility and compatibility. This is
reached because of the loosely coupled View and Controller. The
proposed Cloud Security Kernel is provisioned for
implementation of the customized security solutions requested by
the client. The risk associated with the unsafe communication
channel is jointly borne by both provider and consumer of the
Cloud services and thereby easing the sound formulation of the
Service Level Agreement (SLA) to accelerate the Cloud driver e-
Learning business.
Going forward, we are planning to introduce the JavaScript
Object Notation (JSON) for representing internal data model of
the task rather than using XML [12]. The performance economy
associated with the XML data model is inversely proportional to
the size of the XML data and hence, the parsing of the bigger size
XML becomes cumbersome. The JSON standard is a most
lightweight standard to model the data representation.
The Cloud service prototype was built on the top of the VM
cluster computing, and we are planning to implement the grid
computing to get the maximum possible utilization of the
underlying resources by harnessing together many servers running
in parallel. A change from VM cluster to grid is only related to
the performance enhancement and does not specifically influences
the task management process.
6. ACKNOWLEDGMENTS
This work is supported by the Japanese Society for the Promotion
of Science, Grant-in-Aid for Scientific Research: 21500948,
2009-2011.
7. REFERENCES
[1] Ames,S., and Gasser, M. 1983. Security Kernel Design and
Implementation: An Introduction. Tutorial Computer and
Network Security (October 1986). IEEE Computer Society
Press, Washington, D.C., United States. 142-150.
[2] Buyya, R., Yeo, C., Venugopal, S., Broberg, J., and Brandic,
I. 2008. Cloud computing and emerging IT platforms: vision,
hype, and reality for delivering computing as the 5
th
utility.
In journal of Future Generation Computer Systems. 25(6),
599–616.
[3] Comer, D. 2004. Computer Networks and Internets with
Internet Applications. 4
th
Edition 2004. Pearson Prentice
Hall, New Jersey. 591-605.
[4] Dagger, D., O’Connor, A., Lawless, S., Walsh, E., and
Wade, V.P. 2007. Service-Oriented E-Learning Platforms
from Monolithic to Flexible Services. In IEEE Internet
Computing, vol. 11. No. 3. May/June. 28-35.
[5] Dong, Bo., Zheng, Q., Yang, Jie., Li, H., and Qiao, M. 2009.
An E-learning Ecosystem Based on Cloud Computing
Infrastructure. In 9th IEEE International Conference on
Advanced Learning Technologies, ICALT. Riga, Latvia July
15-17. 125-127.
[6] Gamma, E., Helm, R., Johnson,R., and Vlissides, J. 1994.
Desing Patterns Elements of Reusable Object Oriented
Software. Addison-Wesley, New York.
[7] Lin,J., Ho,C., Sadiq,W., and Orlowska, M. 2001. On
Workflow Enabled e-Learning Services. In Proceedings of
the IEEE international Conference on Advance Learning
Technologies (August 06- 08, 2001). ICALT. IEEE
Computer Society, Washington, DC. 349.
[8] Mell, P., and Grance, T. 2009. Definition of Cloud
Computing. The National Institute of Standards and
Technology (NIST). United States Department of Commerce
Version 15. URL: http://csrc.nist.gov/groups/SNS/cloud-
computing/
[9] Oka, T., Vazhenin, D., and Vazhenin, A. 2008. In
proceedings of the International Conference on
Computational Technologies in Electrical and Electronics
Engineering Sibircon (July 21-25, 2008). IEEE Region 8
Novosibirsk, Russia. 100-105.
[10] Singh, H. 2003.Building Effective Blended Learning
Programs. Issue of Educational Technology, Vol. 43, 6, 51-
54.
[11] Sun Microsystems. 2009. Introduction to Cloud Computing
Architecture. White Paper (June 2009).
http://www.sun.com/featured-articles/CloudComputing.pdf
[12] Sun Microsystems. 2008. Using JavaScript Object Notation
(JSON) in Java ME for Data Interchange.
http://java.sun.com/developer/technicalArticles/javame/json-
me/
[13] Vazhenin, D., Cortez, R., and Vazhenin, A.2009. Task
Verification Strategies in e-Learning Environment. In
Proceedings of the 12th International Conference on Humans
and Computers (December 7-10, 2009). Japan. 157-162.
[14] Yachin, D. 2008. Cloud Computing. It Is Time For Stormy
Weather. IDC SaaS Summit Spring 2009.
http://www.grid.org.il/_Uploads/dbsAttachedFiles/IDC_Clou
d_Computing_IGT_final.ppt
[15] Zhu, Z. 2009. Design and Implementation of Web-Services
Based E-Learning System. In 1st International Workshop on
Education Technology and Computer Science. Vol 3, 233-
237.