FUNCTIONAL SEMANTICS AWARE BROKER BASED ARCHITECTURE FOR E-LEARNING WEB SERVICES

International Journal on Integrating Technology in Education (IJITE) Vol.1, No.1, December 2012
11
FUNCTIONAL SEMANTICS AWARE BROKER BASED
ARCHITECTURE FOR E-LEARNING WEB SERVICES
Demian Antony D’Mello
Department of Computer Science & Engineering, St. Joseph Engineering College,
Vamanjoor, Mangalore -575 028. INDIA
demian.antony@gmail.com
ABSTRACT
E-learning enables the learner to gain diverse knowledge anytime, anywhere and on any device. Learning
resources (objects) and resource providers play a very important role in e-learning applications/systems.
The increasing demand for interoperability in existing heterogeneous e-learning systems to support
accessibility and reusability is the most challenging research issue. Web services and SOA enables
interoperability between heterogeneous applications over the Web. To adopt Web services technology
towards the reusability and aggregation of e-learning services, the conceptual Web services architecture
and its building blocks need to be augmented. In this paper, a well formed functional semantics approach
is proposed to describe e-learning Web services providing variety of learning objects/resources. The
paper presents an extendible functional knowledge to map the learner’s or provider’s versions of service
descriptions into a standard form called Abstract Description. The authors propose a broker based e-
learning Web service architecture which facilitates effective e-learning service publishing and discovery
mechanisms. The paper explores a scheme to extend the WSDL 2.0 document in order to incorporate
functional semantics of e-learning Web services and their operations. The paper presents an e-learning
service knowledge called Learning Operation Tree (LOT) for the quick e-learning service discovery. The
experimentation shows that, the proposed broker based architecture for e-learning Web services
facilitates effective discovery with moderate performance in terms of recall and response.
KEYWORDS
E-Learning; Functional Semantics; Broker Architecture; Discovery; Learning Object; Learning
Operation Tree
1. INTRODUCTION
With the rapid growth of Internet technologies, knowledge can be shared, retrieved and
distributed through the Web all over the world. Currently, the educational system has been
evolved into electronic learning (or e-learning) that enriches the traditional learning system by
providing the autonomous learning for learners and allowing learner to learn anywhere and
anytime [1]. E-learning is a technology driven learning process based on Web technology. E-
learning is also defined as an acquisition, application and dissemination of knowledge facilitated
primarily by electronic means. The rapid growth of computers and access to Internet made it
possible to bring the concepts like virtual university or twenty four hour learning in reality [2].
Presently, e-learning is a major kind of knowledge and information sharing method which
allows the learner to access different kinds of learning resources such as lecture video, teaching
audio, lecture slides/handouts, e-books, downloadable application tools and software [3]. E-
learning is also a type of education that offers some interesting benefits over traditional learning
in terms of independence. The learners can work anywhere and communicate with instructor or
other learners via e-mail, electronic forums, chatting, video conferencing and other forms of
computer and Web based communication [4]. In e-learning, the reuse of learning
objects/resources and services is a key issue and standardizing e-learning technology is taking
place in order to overcome interoperability problems [5]. Reusing of existing resources and

12
infrastructures to implement an e-learning system can reduce the overall operational cost of the
system which is ideal for learners to exploit various e-learning facilities.
In e-learning domain, the learning resources (Learning Objects) and their providers play a very
important role in E-learning applications/systems [6]. Content consumed by learners and created
by authors/educators is commonly handled, stored, and exchanged in units of learning objects
(LOs). Basically, LOs are units of study, exercise or practice that can be consumed in a single
seamless session. They represent reusable granules that can be authored independently of the
delivery medium and be accessed dynamically over the network [4]. For example, a LO on the
basics of C++ can be used in classes on OOPs and Programming Languages. As the number of
resources and providers grow, meta-data on the resources becomes a critical factor. A meta-data
is needed for an appropriate description of learning objects so that plug-and-play configuration
of knowledge dissemination is enabled. To this end, several standardization efforts have been
initiated which include Learning Object Metadata (LOM) and SCORM, which is a collection of
specifications adapted from multiple learning sources to provide a comprehensive suite of e-
learning capabilities that enable interoperability, accessibility, and reusability of Web based
learning content [7].
Web services technology enables the reuse and interoperation among heterogeneous platforms.
It provides mechanisms for the description and search/lookup of computational entities over the
Internet. A Web service is an interface, which describes a collection of operations that are
network accessible through standardized XML messaging [8]. Web service discovery is the
mechanism, which facilitates the requester, to gain an access to Web service descriptions that
satisfy his functional requirements. UDDI [9] is the early initiative towards discovery, which
facilitates both keyword and category based matchmaking and discovery.
1.1. Motivation
The existing UDDI [9] and WSDL 2.0 [10] do not support e-learning service specific functional
descriptions for the discovery of e-learning resources. In order to publish e-learning service
description for the global access, the existing WSDL structure need to be augmented to
incorporate necessary information required for the discovery. As a motivating example,
consider the scenario of learner who is interested to download the study material related to
J2EE. In order to get correct e-learning resource, the learner visits and evaluates learning
content present in many Web pages using popular search engines like Google. The search
engines provide the listing of Web links of related material to the learner which makes him to
verify genuineness of the learning content. Moreover, the search engines do not distinguish Web
services and Web pages which again make him to choose only e-learning Web services for the
semi-dynamic binding requirements. Thus, the e-learning Web services need to be stored at
central repository towards consistent updating and global access. In order to initiate effective
lookup for the specific e-learning services, the learner needs to describe his learning
requirements in a specific form. Moreover, the providers of e-learning services have to publish
all e-learning services at the central repository with well accepted description formats and
procedures. This well-formed description of learning Web services and learning requirements
will improve the hit rate of the lookup/discovery mechanism for the static and dynamic binding
of e-learning resources/objects. Therefore, the existing Web service architecture and its major
building blocks (UDDI and WSDL) need to be augmented to enable effective e-learning
service/resource discovery.
1.2. Related Works
E-learning has been a topic of increasing interest in recent years. In literature, the authors of the
paper [1] propose the reference architecture for interoperating the existing e-learning system
with the help of Web services. The authors also explore a metadata-UDDI model which is
designed as a core component of the architecture. The paper [11] proposes an extensible SOA

13
based platform that facilitates implementation of e-learning systems. The platform has applied a
service-oriented framework and model driven architecture into the analysis, design,
implementation and integration of e-learning applications. In literature, the researchers also
explored the application of semantic Web technologies to e-learning domain. The paper [12]
illustrates the use of semantic Web technologies such as RDF to e-learning system for helping
learners to select suitable learning course or retrieving relevant information. In literature,
annexed algorithm, called eLSDAUS, is proposed to improve the existing semantic-based
matchmaking algorithm [13]. The proposed algorithm, Introduces a new factor called “User
Satisfaction”, which is the user's level of satisfaction about the result of service discovery. This
algorithm allows users to take part in the process of e-Learning service discovery, and evaluate
the result of service discovery. The authors in [14] present an SWS architecture which is based
on Web Service Modeling Ontology (WSMO) and uses the Learning Object Metadata (LOM)
Standard to describe the learning object/resource. The paper [15] analyses three of the most
successful e-learning platforms (Blackboard, Moodle and Sakai), identifying their Web services,
and comparing their readiness for the development of a virtual campus based on these services.
The authors also provide a mechanism to facilitate the integration of these platforms in an
information technology infrastructure.
The researchers also explored the way to store e-learning resources and their descriptions into
repository. The authors of paper [6] illustrate the design and implementation of a distributed
learning resource registry system. They define Distributed Learning Resource Registry and
Discovery Model, which enable the developers and repository systems to register learning
resources into the registry system and provide a discovery mechanism to find required learning
resources. The paper [16] proposes a metadata model for indexing the learning services. The
authors propose to describe and index learning services with three dimensions: as learning
resources, as services that contribute and help researchers and as general services.
Current e-learning frameworks should take advantages of Web services and intelligent agents.
The paper [17] proposes a novel architecture for E-learning systems based on Web services and
intelligent agents. This architecture provides a flexible integration model in which all the
learning components and applications are loosely connected and can be distributed on the
Internet. In addition, through the use of agents, learning content can be intelligently customized
to fit the context and the special learning needs of particular users. The authors “Guo W. and
Chen D.” present the main features of e-learning scenario and setup an e-learning scenario
ontology with the training and application domains. The authors also present semantic querying
and semantic mapping approach for the query containing learning requirements. The paper [18]
discusses three aspects of E-learning system: theoretical framework, function components and
technology architecture. The authors provide both theoretical grounding and practical advice for
designing and implementing effective E-learning system in each of these areas. The authors of
[19] propose a Web Services based solution to exchange learner’s information among different
e-learning systems described following a Web Ontology. The proposed solution makes different
e-learning systems to cooperate with each other in order to reach a set of learner information
richer than that currently found in standard e-learning systems.
E-learning system provides a set of personalization functionalities such as personalizing
learning plans, learning materials, test and necessary instant messages etc., to online learners
[20]. The problem in the existing system is the lack of personalization due to weak-semantic
learning resources. The possibilities of personalized searching for information will be improved,
with the advance of the semantic web and available web services. The authors of the paper [20]
present an approach to e-learning personalization based on ontology and information exchange
is maintained by web services based on Service-oriented architecture. With this mechanism, the
learning process is enhanced by providing personalized learning content to the learners in an
effective and dynamic intelligent way.

14
Adopting Web 2.0 technologies and techniques in modern e-learning systems guarantees a more
interactive e-learning experience [21]. It leverages collaboration among learners and enhances
accessibility to various learning resources. The paper [21] identifies recurrent Web 2.0 and
Service-oriented architecture (SOA) design and architectural patterns that would provide
reusable building blocks for any Web 2.0 based service-oriented e-learning system. The
proposed design patterns share three elementary architecture types, client-server, peer-peer and
SOA. The paper also builds on UML4SOA techniques in modeling requirements prior
application of proposed patterns in the case study. The paper [22] tries to support a personalized
strategy customized for programming course. The authors have designed a model for
personalized learning. It included an information model and a process model which has a great
adaptability for strategy and strategy combination. The proposed service can adapt to the change
of strategies, not only to a rule's change, but also to the change of entire strategy plan. So far
there is no effort in the literature from the researchers to apply restricted natural form of
functional description to e-learning Web services/systems. Moreover, the concrete e-learning
architectures with effective mechanisms for discovery and publishing are today’s needs which
are not addressed by the e-learning research community
1.3. Contributions
The authors provide effective solutions for the key issues with respect to e-learning Web service
description and discovery. The key contributions of this paper are:
• Definition of functional semantics terminology and well-formed semantic rules for the
description of e-learning services.
• The design of domain dependent, extendible e-learning functional knowledge for the
effective e-learning service registration/publishing.
• The extension of WSDL 2.0 document structure to accommodate functional descriptions of
e-learning services in WSDL documents.
• The design of e-learning Web service knowledge called Learning Operation Tree (LOT) to
enable quick discovery.
• A broker based architecture for the e-learning services discovery and publishing.
• Matchmaking mechanism for the e-learning service discovery based on the functional
semantic descriptions of learner’s information/knowledge requirements.
1.4. Structure of the paper
The paper is structured as follows: In the next section (Section 2), the paper describes the
functional semantics terminology for e-learning services. Section 3 defines the e-learning
domain knowledge structure which stores e-learning service specific descriptions for
matchmaking. Section 4, presents e-learning service knowledge which is a tree structure to store
all published e-learning services and their operations. Section 5 presents the e-learning service
publishing using augmented WSDL 2.0 involving functional semantics. In section 6, the authors
propose the broker based architecture for effective e-learning Web services description,
publishing and discovery. Section 7 presents implementation of the broker architecture in .NET
environment, experimentation dataset followed by results. Section 8 draws the conclusions and
provides potential areas for further work.
2. FUNCTIONAL SEMANTICS FOR E-LEARNING SERVICES
Web service i.e. service is the globally accessible software whose functionality can be
embedded within another application. In order to promote e-learning service reusability, the e-
learning services need to be described in a precise way. This section of the paper presents the e-
learning service description semantics to describe learner’s request and e-learning services in a
precise way. E-learning service is a network accessible system interface having collection of
operations/functionalities that aim at providing some resources or information to the learners.

15
Thus e-learning service operation is nothing but the execution of appropriate learning action on
specific learning object to provide information to the learner. The functionality of any e-learning
service operation can be characterized using learning action on the learning object. The
following definitions help to frame the functional semantics to describe e-learning services.
2.1. Definition of Functional Semantics for E-Learning
The functional semantics approach uses the natural way of expressing the functionality of Web
services and their operations. The functionality of an e-learning service/operation is described in
terms of learning action, qualifier, learning object(s) and noun.
Generic Action: Generic action is an action used to perform an operation on learning object or
to get information in terms of learning object. For example “check teacher availability”
description involves the generic action “check” which is commonly used across multiple
domains.
Learning Action: Learning action is an action performed on learning object in a learning
domain to render service to the learner. The examples for learning action are: Learning action
normally has a related action noun which can be used to describe the operation functionality.
For example, the learning operation description “load question set” involves an action “load”
which is learning action.
Learning Object: Learning object is an object of e-learning service for which the required
action is sought by the e-learning service/operation. For example, teacher, author, student,
question, lecture, test, book, reference are some objects found in the e-learning domain.
Learning objects are classified as Main Objects and Related Objects based on the association
among them. The main objects are the entities (objects) that constitute the e-learning service
domain. For example course, subject, topic, question, assignment, seminar, book, lecture are the
few main objects. The related object is a logical or physical part of the major object in e-
learning domain. For example set, bank, material etc. are the few related objects.
Learning Noun: Learning noun is a noun used to describe the e-learning Web service or its
operation. For example conduction, issuing, sending, loading etc. are the learning nouns used in
e-learning domain.
Learning nouns are classified as Action Noun and Simple Noun (generic noun) based on the
action represented by the learning noun. An action noun is a learning noun which has a related
learning action. A simple (generic) noun doesn’t represent any learning action on the learning
object. For example, the noun “conduction” is an action noun as it has a related specific action
“conduct” whereas the learning noun “availability” (in case of check course availability/check
text book availability”) is a simple noun, since it does not represent any learning action.
Qualifier: Qualifier is a word which adds the value to the learning object i.e. qualifier specifies
the feature or nature of the learning object. For example, the operation description “find text
book” has a qualifier “text”. Similarly, the description “check crash course availability” contains
the word “crash” which is a qualifier.
Operation Description: Operation description refers to the functionality description of an
operation which is found in the WSDL document of e-learning Web service during e-learning
service advertisement. The operation description represents one of the abstract learning
operations supported by the advertised e-learning Web service.
Abstract Operation: Abstract operation is a single, compact and complete description for the
multiple and similar descriptions of e-learning Web services or their operations. All service
operation descriptions are transformed to their corresponding abstract operation(s) during e-
learning service advertisement for the effective discovery. The important property of abstract

16
operation description is that, it does not allow generic learning action together with an action
noun to be present in its description.
2.2. Describing e-Learning Service Operations
The functional semantics approach facilitates the provider and learner of e-learning service to
use natural form of describing the e-learning service functions i.e. operation descriptions. Thus
functionality of an abstract e-learning operation can be described in the following three formats:
• Operation description = (Leaning Action) (Qualifier)*
(Learning Object)+
(Simple Noun)
• Operation description = (Learning Action) (Qualifier)*
(Learning Object)+
• Operation description = (Qualifier)*
(Learning Object)+
(Action Noun)
• Operation description = (Generic Action) (Qualifier)*
(Learning Object)+
(Action Noun)
The grammar rules to validate the abstract operation functionality are designed as follows. Let
AS be the set of terminal symbols representing various learning actions i.e. AS = {load, display
etc.}. Let AG be the set of terminal symbols representing various generic actions i.e. AG = {do,
perform, get, check etc.}. Let O be the set of terminal symbols representing various learning
objects i.e. O = {lecture, book, author, teacher, slide, handout, note, summary etc.}. Let Q be the
set of qualifier symbols representing various qualifiers i.e. Q = {text, reference, crash etc.}. Let
NA be the set of terminal symbols representing action nouns i.e. NA = {loading, conduction,
teaching etc.}. Let NS be the set of terminal symbols representing simple nouns i.e. NS =
{availability etc.}. Let AG, AL, NS and NA represent generic action, learning action, simple noun
and action noun respectively. Let S be the start symbol and X, Y and Z be the non-terminal
symbols. The production rules of the grammar to validate functional semantics of Web service
operation are presented in Figure 1(a). An example of generating (leftmost derivation) operation
description “load text book” using the grammar production rules is depicted in figure 1(b).
Figure 1. Grammar to Validate Learning Operation Description
Consider the learning scenario; the following e-learning operation descriptions follow the rules
of functional semantics.
a) check course availability
b) download reference material
c) upload course assignment
d) course assignment uploading
e) start reference material downloading
2.3. Pre-processing of E-Learning Operation Descriptions
All operation descriptions are preprocessed before being mapped into abstract e-learning
operations. The preprocessing and mapping mechanism facilitates the effective publishing and
discovery of e-learning services. The preprocessing of operation descriptions involves
elimination of co-occurrence of generic action and learning action noun present in the operation
description. The following rules guide the preprocessing of operation descriptions.

17
Rule I. If the learning action noun is present along with the generic action, then the generic
action is replaced by the learning action which is related to the learning action noun and the
action noun is eliminated from the description.
Rule II. If the learning action noun is found in the operation description without a generic
action then the related learning action of the action noun is used, instead of the action noun.
As an illustration, consider the operation description “perform lecture downloading”. The
description contains generic action and action noun. The generic action is now replaced by
“download” which is the specific action of action noun “downloading” and the generic action is
eliminated from the description as per Rule I. This results in abstract operation description “load
lecture”. Similarly, the operation description “test conduction” is transformed into “conduct
test” by Rule II.
3. E-LEARNING DOMAIN KNOWLEDGE STRUCTURE
To perform e-learning service discovery based on the functional description of Web service
operation, we design an extendible functional knowledge which contains interdependent
knowledge structures to represent the complete functional knowledge for all categories of Web
services. The interdependent knowledge structures are: Object List, Action List, Qualifier List
and Noun List.
Object List: Object list is a sorted list with finite elements where each element contains four
fields i.e. information items. They are- object name, object identifier, object type and a pointer
to the sorted related object list having similar/related names of a specific object. The object
name refers to learning object for which learning action is to be sought, object identifier is a
unique identification string and object type refers to either main (M) or related object (R). The
object list and related object list can be implemented as dynamic array which is sorted based on
the object name.
Action List: Action list is a sorted list with finite elements each containing three fields namely
action name, action identifier and a pointer to the sorted related action list containing similar
action words for a specific learning action. The action list and related action list can be
implemented as a dynamic array.
Qualifier List: Qualifier list is a sorted list with finite elements each containing three fields
namely qualifier name, qualifier identifier and a pointer to the sorted related qualifier list
containing similar qualifier words for a specific qualifier. The qualifier list and related qualifier
list can be implemented as a dynamic array.
Noun List: Noun list is a sorted list with finite elements each containing four fields namely
noun name, noun identifier, noun type, a pointer to its corresponding action (if any) and a
pointer to the sorted related noun list containing similar noun words used to describe a specific
learning noun. The noun list and related noun list can be implemented as a dynamic array sorted
based on the noun name. The noun type refers to noun categories, learning action noun (A) and
simple noun (S).
Figure 2 depicts the partial e-learning domain knowledge structure showing interdependent
structures or lists having information of e-learning domain. The noun list contains two learning
action nouns and one simple noun. The object list has an object “set” which is categorized as a
related learning object. The unique identifiers like action identifier, noun identifier, qualifier
identifier and learning object identifiers are fixed length strings used to identify the learning
actions, nouns, qualifiers and learning objects. E-learning domain knowledge is augmented by
the e-learning service providers in order to improve the discovery rate of their advertised
services. In order to transform the description of e-learning service operation to its equivalent
abstract operation, a separate list is maintained called Abstract Operation List (AOL). The
structure of AOL is defined below.

18
Figure 2. E-Learning Domain Knowledge Structure for Discovery
Abstract Operation List: Abstract operation list is a sorted dynamic array with finite elements
each representing an abstract e-learning service operation. The element contains operation
identifier, operation pattern and e-learning service count; where operation pattern is a string of
finite length which contains fixed length identifiers of learning objects, noun qualifiers and
learning actions. The e-learning service count refers to the number of e-learning Web services
having description of operation which maps to an abstract operation.
Operation pattern is generated for each abstract operation defined in AOL. Let M be the fixed
length for identifiers of learning actions, nouns, qualifiers and learning objects. The first M
characters represent the action identifier. Next, the sets of M characters represent the qualifier
identifiers (optional), finally the sets of M characters represent the learning object identifiers
followed by the noun identifier (optional).
4. E-LEARNING SERVICE PUBLISHING WITH WSDL 2.0
The operations of e-learning services are described using the functional semantics as defined in
section 2.2. Thus e-learning Web service can be described for publishing using functional
semantics to facilitate the effective search.
4.1. Describing E-learning Web Services
Let Profile (WS) be the profile of the e-learning Web service to be published through the search
agent into augmented UDDI registry.
Profile (WS) = {service-desc, binding-desc} where, service-desc refers to e-learning service
specific descriptions like service name, provider name, operation descriptions etc and binding-
desc refers to binding details like URL for the access. Thus service-desc = {service-name,
provider-name, OPList} where, OPList is the list of e-learning operations and their descriptions
supported by the e-learning Web service.
The OPList = {opr1, opr2…oprN} where, opri is the description of an e-learning operation.

19
The description of each operation is: opri = {opr-name, desc-list, info-list}; where desc-list is the
functional semantics description of operation as defined in section 2.2 and info-list is additional
information (optional) to update the extendible e-learning domain knowledge.
desc-list={learning action, qualifier(s), learning object(s) , noun} where, qualifiers and noun are
optional.
info-list = {action-set, qualifier-set, learning object-set, noun-set} where, action-set contains
similar learning action words, qualifier-set contains similar qualifier names for a given qualifier
and object-set contains similar learning object names and noun-set contain similar noun names.
As an illustration, let us consider the description of “summer course registration” with an
abstract operation “register course”.
Profile (Learning Service) = {service-desc, binding-desc}.
service-desc={“summer course registration”, “VTU”, OPList}.
OPList={opr1} and opr1={register course, desc_list1, info-list}.
desc_list1={action, qualifier, object, noun} and action={register}, object={course},
qualifier={summer} and noun={Φ}.
Info-list={action-set, object-set}.
action-set={enroll} and object-set={object1} and object1= {subject, topic}.
qualifier-set={qualifier1} and qualifier1={vacation}.
noun-set = {Φ}.
4.2. WSDL 2.0 Document Structure
WSDL 2.0 [10] separates the description of a Web service's abstract functionality from the
concrete details of how and where that functionality is offered. This separation facilitates
different levels of reusability and distribution of work in the lifecycle of a Web service and the
WSDL 2.0 document that describes it. Figure 3 shows the abstract structure of WSDL 2.0
document.
Figure 3. Structure of WSDL 2.0 Document

20
Every WSDL 2.0 document has a description element as its top-most element. This merely acts
as a container for the rest of WSDL 2.0 document, and is used to declare namespaces that will
be used throughout the document. The documentation element is used to present the brief
information about the document and the service. The documentation element allows the WSDL
2.0 author to include some human-readable documentation inside a WSDL 2.0 document. It can
appear in number of places within the description element.
WSDL 2.0 allows message types to be defined directly inside the types element, which is a child
of the description element. A WSDL 2.0 interface defines the abstract interface of a Web
service as a set of abstract operations, each operation representing a simple interaction between
the client and the service. Each operation specifies the types of messages that the service can
send or receive as part of that operation. Each operation also specifies a message exchange
pattern that indicates the sequence in which the associated messages are to be transmitted
between the parties. The binding name specifies the concrete message format and transmission
protocol details for an interface. A WSDL 2.0 service element specifies a single interface that
the service will support and a list of endpoint locations where that service can be accessed.
4.3. Extension of WSDL 2.0 for E-Learning Web Services
WSDL 2.0 structure is extended to publish the e-learning Web services with functional
semantics as follows. Table 1 provides the newly defined XML elements and their description.
Table 1. XML Elements for WSDL 2.0 Structure
Element Name Purpose
<operationDesc> Container for functional description of learning operation
<operationList> Container for all operation descriptions of learning Web
service
<operation> Container for descriptions of learning operation
<operationName> To represent learning operation name
<semantics> Container for functional description
<action> To represent learning action
<object> To represent learning object
<qualifier> To represent qualifier
<noun> To represent noun
<information> Container for the information to augment functional
knowledge
<related> Represents related words for action, noun, qualifier and
objects
The documentation element is chosen to insert the information which is necessary for the
effective service discovery into WSDL. A new tag called operationDesc is defined to insert the
functional semantics of all abstract operations present in the e-learning service. The new
elements operationList, operation, action, qualifier, object and noun are found within the
element operationDesc. The new elements are defined in the XML schema which governs the
structure of extended documentation element.
The functional semantics of an operation is defined within the element semantics and this
element is placed within the element operation. The elements like action, qualifier, object and
noun are used within semantics element which provides the functionality description of an
abstract e-learning operation. The functional semantics for the illustration presented in section
4.1 is depicted in Figure 4.

21
5. E-LEARNING SERVICE KNOWLEDGE STRUCTURE
E-learning service can advertise multiple learning operations. To store such redundant e-
learning service operations in an efficient way for the discovery, we define two data structures
called Web service list (ESL) and Service Operation tree (LOT).
Figure 4. Extended WSDL for “Summer School” E-learning Service
E-learning Service List (ESL): E-learning service list is a sorted dynamic array having four
fields namely, e-learning service key (es-key or ws-key), es-id (unique identifier generated by
the broker), es-link and lot-link where es-id is a e-learning service identifier, es-link is a pointer
to the e-learning service entry in ESL having same set of operations and lot-link is the pointer to
the leaf node of Learning Operation Table (LOT) which corresponds to a link to its operations
in the LOT or a pointer to the predecessor e-learning service in ESL having same operations.
Learning Operation Tree (LOT): A learning operation tree is a binary tree with each node
consisting five fields. They are operation identifier (opr-id) which specifies the operation
identifier of e-learning service operation (abstract operation); child pointer (child-link) which is

22
a pointer to the remaining operations of a e-learning service; sibling pointer (sibling-link) is a
pointer to the list operations which shares a common operation prefix; parent pointer (parent-
link) is a pointer to its predecessor node; e-learning service link (esl-link) is a pointer to the ESL
entry to which opr-id is the last learning operation in the sorted advertised operation list. The
root node of LOT is labeled with T and has only child-link which points to Web various service
operation sequences. The property of LOT is that at any node X of LOT, the opr-id at X will not
be repeated at the child or sibling branch which is linked to X.
As an illustration consider six e-learning services having a total of nine learning operations to be
advertised into repository. Let Opr1 to Opr9 be the operation identifiers of abstract operations
obtained after mapping them into abstract operations. ES1 = {Opr1, Opr2, Opr3, Opr4}, ES2 =
{Opr1, Opr2, Opr3}, ES3 = {Opr5, Opr6}, ES4 = {Opr7, Opr8, Opr9}, ES5 = {Opr5, Opr6} and ES6
= {Opr7, Opr8, Opr9}. Figure 5 shows the LOT and ESL after insertion of learning operations of
e-learning services. The numbers within circles (node) indicate the operation identifiers (Opr1 to
Opr9) in an ascending order. In ESL, ES1 to ES6 represents the e-learning service key which is
same as Web service key.
Figure 5. E-Learning Service Knowledge Structure
6. THE BROKER BASED ARCHITECTURE FOR E-LEARNING
SERVICES DESCRIPTION AND DISCOVERY
The broker based architecture facilitates an effective discovery and publishing of e-learning
Web services. Figure 6 depicts different roles and operations supported by the broker based e-
learning service architecture.
Figure 6. Learning Broker Based Architecture for E-learning Services

23
The architecture assumes that the e-learning service requesters and LO providers use functional
semantics to describe learning operations and learning needs. An additional role is introduced
by the authors to the conceptual Web service architecture [8] named Learning Broker (broker)
and a new operations namely Register Learning Service and Find Learning Service. The broker
is defined between Web service registry and learner (and provider) which facilitates the learner
and provider to specify the needs and learning services in terms of functional semantics. The
find learning service operation is defined between the broker and learner, which effectively
explore the learning services from LOT. The register learning service operation is defined
between the learning resource provider and broker for the e-learning service publishing.
6.1. Components of Learning Broker
The learning broker is designed with four internal components namely Learning Service
Publisher, Learning Service Finder, Learning Domain Knowledge and Learning Service
Knowledge (Service Knowledge). Figure 7 depicts different components of a broker and the
interactions among the components within the architectural boundary. The service publisher
component facilitates the registration, updating and deletion of business and e-learning service
related information. The main functionality of service finder is to discover the e-learning
services which satisfy the learner’s demands. The domain knowledge is an interlinked data
structure which represents learning actions, qualifiers, learning objects and nouns of e-learning
service domain. The service knowledge is an abstract representation of all published e-learning
Web services and their learning operations.
Figure 7. Components of E-Learning Broker and Component Interactions
The sequence of interactions among various architectural roles for e-learning service
registration (publishing) is presented below.
1. The service publisher supplies the WSDL (with functional semantics information of
learning operations) document to the broker along with other information pertaining to the
e-learning service provider.
2. The broker publishes the service information into service registry and obtains the service
key (es-key).

24
3. The broker now extracts functional description of all learning operations and updates the
learning domain knowledge and service knowledge accordingly.
4. On successful service registration, the service key is returned to the e-learning service
provider as an acknowledgment.
The sequence of architectural component interactions for learning service discovery is given
below.
1. The learner sends the learning request as per the functional semantics format to the broker.
2. The learning service finder of the broker finds an abstract learning operation for the request
by traversing the learning domain knowledge.
3. The broker now traverses the LOT to find possible learning services.
4. The broker now sends the service keys (and descriptions) of all discovered services to the
learner.
Now the learner initiates a session with the provider towards execution of requested learning
activity.
6.2. Publishing of E-learning Web Services
The provider of the e-learning service publishes the extended WSDL 2.0 into the UDDI (e-
learning service registry) through the broker. The steps involved in the e-learning service
publishing are presented below.
1. The provider registers the e-learning service by sending the WSDL and other necessary
details.
2. The WSDL is processed by the broker to obtain the service name, binding details and the
operation details like operation name and functional semantics.
3. The broker publishes the Web service into UDDI registry and obtains the service key.
4. The operation descriptions are preprocessed according to Rule 1 & 2 as defined in section
2.3.
5. The operation pattern is generated for the preprocessed operation description by obtaining
the appropriate action, object, qualifier and noun identifiers from the e-learning domain
knowledge. If the action/noun/qualifier and object is not present in the functional
knowledge, then these are inserted by generating appropriate identifiers.
6. Search the operation pattern in ALOL. If found, return the operation identifier of the
operation having the generated operation pattern otherwise insert the operation as a new
abstract operation along with its pattern into ALOL.
The operation identifiers of all e-learning Web service operations along with service key are
stored LOT to enable quick discovery.
6.3. E-Learning Web Service Discovery
E-learning service discovery for the learner’s request and the matchmaking process is
summarized below.
The learner sends the request, enriched with functional semantics to the broker for discovery.
1. The service finder of broker validates the functional semantics of learner’s request (task to
be carried out) using the rules as defined in section 2.2.
2. The learner’s request is preprocessed according to Rule 1 & 2 (section 2.3) to retrieve the
learning requirement.
3. The action list, qualifier list, object list and noun list of the e-learning domain knowledge
are searched to get the corresponding identifiers. The unavailability of any identifier results
in discovery failure.
4. The operation pattern for the request is formed using action, qualifier, object and noun
identifiers.
5. After building the operation pattern, the pattern is searched in ALOL. If the pattern is
found then the corresponding operation identifier is retrieved from the ALOL otherwise
discovery failure is reported.

25
The LOT is traversed to search the requested learning operation identifier and all e-learning
services with requested learning operation are returned to the learner as the suitable e-learning
services/resources.
7. IMPLEMENTATION AND EXPERIMENTS
The proposed broker based e-learning service discovery mechanism is implemented on the
Windows 7 platform using Microsoft Visual Studio .NET 2005 development environment and
Microsoft visual C# as a programming language. The broker is designed and implemented as a
Web service which is referenced in a standalone visual program. This visual program interacts
with the e-learning resource/service provider and learner through different interface forms. The
service repository is implemented as a Web service which in turn communicates with the SQL
server 2005 database. The database table is created to store the information about the published
e-learning Web services.
The e-learning service is published by activating the publish menu which pops out a window
form where, the provider supplies the Web service information and attaches augmented WSDL
of the e-learning Web service. The WSDL is processed by the agent program which stores and
updates the necessary information in it. The broker also publishes e-learning Web service
information into e-learning service repository. The service request processing is done by
activating relevant interface form from the menu. The learner is expected to follow the
functional semantic format while supplying learning request. The request is validated by the
agent and the suitable Web service keys are obtained from its local store as the discovery result
after successful matchmaking. Several experiments have been conducted to verify the
effectiveness of functional semantics based matchmaking concept. We have also performed
preliminary experiments to evaluate the system performance in terms of precision and recall.
Precision = (Relevant ∩Retrieved) / Retrieved
Recall = (Relevant ∩ Retrieved) / Relevant
The collection of 40 e-learning services having total of 65 distinct operations is used for
experimentation. Thirty Three e-learning requests based on their short natural language
descriptions were framed. From the experimentation it is observed that, the recall is less than
100% as the learner sometimes may not follow the functional semantics in a precise way. The
mechanism also exhibit low precision if the published and requested e-learning operations are
described with incorrect functional semantics. Figure 8 show the average recall values obtained
for the different experiments.
Several experiments have been conducted to verify the effectiveness of functional semantics
based matchmaking concept. The collection of 30 e-learning services having total of 45 distinct
operations were used for experimentation. Fifty Three e-learning requests based on their short
natural language descriptions were framed. From the experimentation it is observed that, the
recall is less than 100% as the learner sometimes may not follow the functional semantics in a
precise way. Figure 8 show the average recall values obtained for the different experiments.
The same set of e-learning services is represented using LOT at the broker. The LOT
representation yields a compression ratio of 30%. The abstract learning operation of each
learning service is stored in the main memory of the broker; the discovery mechanism need not
load the entire page having the information of advertised services. This will substantially reduce
the secondary memory access time which in turn improves the response time of the discovery
mechanism. This is because the LOT acts as an index to the e-learning service descriptions
present in the UDDI registry.

26
Figure 8. Performance Evaluation (Average Recall)
The empirical data reveals that, the proposed system has high Recall and 100% precision
provided the learning requests are formed as per functional semantics rules. If the published or
requested operation description does not follow the functional semantics (description is ill-
formed) correctly, then the precision and recall of the proposed system suffers. In order to
improve the recall of the e-learning discovery system, both the e-learning resource/service
provider and the learner have to describe the published/requested e-learning operations as per
the functional semantics rules.
8. CONCLUSION
E-learning service discovery is an important activity which explores multiple e-learning service
services for the given learner’s knowledge requirements. The authors propose a well-defined
functional semantics to describe the e-learning services for publishing and lookup. The e-
learning domain knowledge is designed which facilitates easy and effective service lookup and
publishing. A compact index for service registry called Learning Operation Tree (LOT) is
proposed which enables quick service lookup. The authors also propose broker based e-learning
Web service architecture for discovery mechanism which finds the suitable e-learning
services/resources for the given learning requirements. The broker architecture is implemented
for experimentation and several experiments were carried out to deduce observations. The
experimentation reveals that, the use of functional semantics in describing e-learning services
and the use of compact service knowledge at the broker will improve the effectiveness (Recall,
Precision and response) of e-learning service discovery.
ACKNOWLEDGEMENTS
This research work is sponsored by “VTU Research Grants Scheme” of Visvesvaraya
Technological University (VTU), Belgaum, Karnataka, INDIA (No. VTU/Aca./2009-10/A-
9/11619, dated: 04/01/2010).

27
REFERENCES
[1] Ngamnij Arch-int, Chidchanok Lursinsup and Peraphon Sophatsathit, “A Reference
Architecture for interoperating existing e-Learning Systems using Metadata and Web Services
Model”, In Proceedings of the International Conference on Intelligent Agents, Web
Technologies and Internet Commerce (CIMCA-IAWTIC’05), IEEE 2005.
[2] Afroza Sultana and Ishrat Sultana, “E-School: A Web-Service Oriented Resource Based E-
learning System”, In Proceedings of the International Conference on Networking and
Information Technology, IEEE 2010.
[3] Jianhua Jiang, Shuxin Guo and Jie Zhang, “Architecture Design of Campus Information
Convergence System for E-learning Based on Web Service Technology”, In Proceedings of the
2nd International Conference on Future Computer and Communication, IEEE 2010.
[4] ZHU Zhen, “Design and Implementation of Web-services based E-Learning System”, In
Proceedings of the First International Workshop on Education Technology and Computer
Science, IEEE 2009.
[5] Judith Rodríguez, Luis Anido and Manuel J. Fernández, “How can the Web Services Paradigm
improve the E-learning?”, In Proceedings of the 3rd IEEE International Conference on
Advanced Learning Technologies (ICALT’03), IEEE 2003.
[6] Zhao Gang, Yang ZongKai and Yan Kun, “Design and Implementation of a Distributed
Learning Resource Registry System”, In Proceedings of the Fourth International Conference on
Computer and Information Technology (CIT’04), IEEE 2004.
[7] Gottfried Vossen and Peter Westerkamp, “E-Learning as a Web Service (Extended Abstract)”,
In Proceedings of the Seventh International Database Engineering and Applications Symposium
(IDEAS’03), IEEE 2003.
[8] H. Kreger, “Web Services Conceptual Architecture (WSCA 1.0)”, Published May 2001, [online]
Available: www.ibm.com/software/ solutions/webservices/pdf/wsca.pdf, [visit: April 2007].
[9] Riegen, C.V. (Ed), 2002. UDDI Version 2.03 Data Structure Reference [online]. OASIS Open
2002-2003. Available from: http://uddi.org/pubs/DataStructure-V2.03-Published-20020719.htm.
[10] David Booth and Canyang Kevin Liu, “Web Services Description Language (WSDL)” Version
2.0 Part 0: Primer, W3C Recommendation, Published: 26 June 2007, [online] Available:
http://www.w3.org/TR/2007/REC-wsdl20-primer-20070626.
[11] Zhifen Cheng, Tinglei Huang and Jia Nong, “An Extensible Development Platform for SOA-
Based E-Learning System”, In Proceedings of the 2008 International Conference on Computer
Science and Software Engineering, pp. 901-904, IEEE 2008.
[12] Wenying Guo and Deren Chen, “Semantic Approach for e-learning System”, In Proceedings of
the First International Multi-Symposiums on Computer and Computational Sciences
(IMSCCS'06), IEEE 2006.
[13] ZHU Zheng-zhou, WU Zhong-fu and ZHOU Shang-bo, “An e-Learning Services Discovery
Algorithm Based on User Satisfaction”, In Proceedings of the International Conference on
Networking, Architecture, and Storage, pp. 342-348, IEEE 2008.
[14] Mohamed Hanafy and Medhat Fakhry, “A proposed Architecture for Learning Object Selection
and Discovery based on WSMO”, In Proceedings of the 4th International Conference on Next
Generation Web Services Practices, pp. 10-14, IEEE 2008.
[15] Huertas F. and Navarro A., “Web services availability in e-learning platforms”, In Proceedings
of the 7th International Conference on Next Generation Web Services Practices (NWeSP),
Page(s): 170 – 175, IEEE 2011.
[16] Oussama Kassem Zein, Yvon Kermarrec, Serge Garlatti, Jean-louis Tetchueng and Sylvain
Laube, “A Metadata Model for Web Services Applied to Index and Discover E-Learning
Services”, In Proceedings of the International Conference on Information and Communication
Technologies (ICTTA '06), IEEE 2006.

28
[17] Xu Wei and Jun Yan, “An E-learning System Architecture Based on Web Services and
Intelligent Agents”, In Proceedings of the Ninth International Conference on Hybrid Intelligent
Systems (HIS '09), Page(s): 173 – 177, IEEE 2009.
[18] Zhijun Rong, Kuisheng Chen, Min Zhou and Jinsong Xiao, “Designing and Implementing the E-
learning system”, In Proceedings of the 2008 International Conference on Computer Science and
Software Engineering, pp. 797-800, IEEE 2008.
[19] Daniela Leal Musa, LydiaSilva Muñoz and Moreira de Oliveira J P, “Sharing Learner Profile
through an Ontology and Web Services”, In Proceedings of the 15th International Workshop on
Database and Expert Systems Applications (DEXA’04), IEEE 2004.
[20] Rani, S.J., Ashok, M.S. and Palanivel, K., “Adaptive content for personalized E-learning using
web service and semantic Web”, In Proceedings of the International Conference on Intelligent
Agent & Multi-Agent Systems (IAMA 2009), Pp. 1-4, IEEE 2009.
[21] Hussein D., Alaa G. and Hamad A, “Web 2.0 Based Service-Oriented E-Learning Systems:
Recurrent Design and Architectural Patterns”, In Proceedings of the 11th ACIS International
Conference Software Engineering Artificial Intelligence Networking and Parallel/Distributed
Computing (SNPD), Page(s): 227 – 234, IEEE 2010.
[22] Zhang jinghua, “A semantic web based personalized learning service for programming course in
e-learning”, In Proceedings of the International Conference on Mechatronic Science, Electric
Engineering and Computer (MEC), pp. 1563 – 1566, IEEE 2011
Authors
Demian Antony D’Mello received his Bachelor degree in Computer Engineering in
1999 from Mangalore University, India and his Master degree in Computer
Science and Engineering in 2003 from National Institute of Technology Karnataka
(NITK), Surathkal, India. He completed his PhD from the Department of
Information Technology, NITK in the year 2010. Presently, he is working as a
professor in the Department of Computer Science and Engineering, St. Joseph
Engineering College, Mangalore, India. His research interests are in the areas of
Web technologies, Web services, Distributed Computing and Cloud Computing.

FUNCTIONAL SEMANTICS AWARE BROKER BASED ARCHITECTURE FOR E-LEARNING WEB SERVICES

More Related Content

What's hot (18)

Similar to FUNCTIONAL SEMANTICS AWARE BROKER BASED ARCHITECTURE FOR E-LEARNING WEB SERVICES (20)

Recently uploaded (20)

FUNCTIONAL SEMANTICS AWARE BROKER BASED ARCHITECTURE FOR E-LEARNING WEB SERVICES