ED-MEDIA 2005 - Page 1421
Peer-to-Peer versus Federated Search:
towards more Interoperable Learning Object Repositories


Stefaan Ternier
Katholieke Universiteit Leuven
Leuven, Belgium
stefaan.ternier@cs.kuleuven.ac.be

Daniel Olmedilla
L3S Research Center and Hanover University
Hanover, Germany
olmedilla@l3s.de

Erik Duval
Katholieke Universiteit Leuven
Leuven, Belgium
erik.duval@cs.kuleuven.ac.be



Abstract: This paper reports on our experiences in bridging the world of learning object repositories
and peer-to-peer learning networks. More specifically, we have been developing interoperability
“bridges” between the ARIADNE Knowledge Pool System, a distributed client-server based learning
object repository, and the Edutella peer-to-peer learning network. In our developments, we rely
heavily on the rapidly maturing Simple Query Interface (SQI) standard. Our work not only
demonstrates that it is possible to interconnect more centralized repositories and more distributed
peer-to-peer approaches. We also clarify how these two approaches are complementary. Unifying the
two paradigms for learning object management and access will help to accelerate the evolution
towards a critical mass of easily available, relevant learning objects of high quality.

1 Introduction
The World Wide Web has become a common medium for communication among people for private, academic and
business affairs. As a consequence, the amount of digital material that is sent along and stored in the network increases
rapidly. Obviously, learning is not indifferent to this trend, and the amount of Learning Objects (LO’s henceforth) in
schools, academy and business continues to grow rapidly. As a consequence of this evolution, the focus shifts to new
questions, like for example “Where shall the LO’s be stored?”, “Who manages them?” or “Are they easily findable?”.

In the past, due to the lack of storage capacity and network bandwidth, especially in most desktop computers, dynamic
sharing of information from end user machines was prohibitively costly. As a consequence, networks of computers were
mostly reduced to set of powerful connected servers. In this configuration, it is relatively simple to know which servers
are available and which information is available where to whom. This is also the typical architecture in business
coalitions where several companies share their assets within a network of e.g. partners. Searching for information in
such a network typically consists of querying each of the systems (known in advance) and of gathering and joining the
results received from each of them. This approach is known as Federated Search. The advantages are that search is fast,
the more so as this approach also allows specific optimization techniques based on the structure of the network. A
drawback is that, when a new entity wants to join the network, it must agree with an existing member of the network on
its incorporation into the network.

On the other hand, with the boom of Web-based file-sharing services (e.g., Napster, Gnutella, Morpheus), peer-to-peer
(P2P for brevity) networks have become more relevant. The advantages of the P2P approach include: high flexibility for
peers to join or leave the network dynamically, scalability (recently it was shown that for really large networks, a hybrid
solution with super-peers scales better [14]), autonomy as peers do not relinquish control over their resources and high

ED-MEDIA 2005 - Page 1422
resilience against peer failures. The main disadvantage is that the P2P network requires constant management, as peers
join and leave continuously. This produces an extra load on the network and may slow response times during search.

Therefore the answer to the questions stated at the beginning of this section is context dependent. For example, on the
one hand, Ariadne [1] is an association that enables share and reuse of LOs through a distributed repository that relies
on a client server architecture. On the other hand, Edutella [13] is a schema-based P2P network for an open world
scenario in which LO’s are freely offered (at no charge) and everybody is able to join (no agreement with an existing
member of the network is required).

In this paper, we build on the Simple Query Interface [7, 17] initiative (SQI for brevity), a rapidly maturing standard for
federated search across learning object repositories, to report on our experience in bringing the P2P and client-server
architectures together. We demonstrate that they are not only compatible , but also complementary.

The paper is organized as follows: Section 2 explains the basic motivation for this work. Section 3 briefly explains how
the Simple Query Interface (SQI) works. Finally, section 4 draws how both the ARIADNE & edutella architectures
build on top of SQI.
2 Interoperability of LO Repositories
LO’s can be found on the web, as well as in LO repositories. However, as these repositories are often not
interconnected, users cannot easily get access to all of the relevant material. Rather, users must know which repositories
to search, how to locate them, and then search them individually. This increases user search time and decreases
satisfaction. The European IST project ELENA [3] and the European Network of Excellence PROLEARN [6] try to
remedy this situation. On a more organizational level, the recently launched GLOBE consortium addresses the same
problem: the promise there is to enable federated search across some of the most known LO repositories in the world
(ARIADNE, MERLOT, EDNA, EDUSOURCE and NIME) and thus to trigger global interconnection of all such
repositories.

In this section, we start with a brief description of our experience in connecting two learning object systems and we
generalize from our experience to analyze requirements for interoperability in this context .
2.1 Ariadne and Edutella
In its aim to facilitate both academic education and corporate training, the ARIADNE Foundation supplies its members
with tools and methodologies for producing and reusing learning objects. The core of these services is a distributed
network of repositories that replicate content and metadata. Doing so, each node contains a copy of all metadata
instances . The LO’s however can only be replicated to other servers if no download restrictions apply to them. This
infrastructure, also known as the Knowledge Pool System, enables the ARIADNE user community to transparently
manage learning objects.

Often, learning object providers do not want to abandon control over their resources to a common server, even among
the members of a coalition. In ARIADNE, this concern is addressed through the distributed nature of the Knowledge
Pool System, which enables every organization to set up and maintain its own local server. However, the same concern
about abandoning control also often applies to individuals , who may not want to give away their content to any
centralized repository. More distributed environments have shown to be a feasible solution for interconnection,
integration and access to large amounts of information that deal with this issue. Peer-to-peer networks are an example of
the impact this distribution of information might have in the sharing of information. In such networks, peers can offer
various services to the user that range from search and delivery to personalization and security services. In addition,
they present a solution to the information growth where every learning resource provider offers its information but does
not loose the control over it.

The Edutella P2P network [13] was developed with these principles as main design requirements. It is a schema-based
P2P network with various service facilities implemented like for example query or publishing/subscription. Schema -
based means that peers interchange RDF meta-data (data about data) among each other but not the resources
themselves, that is, they interchange information about e.g. title, description, language and authors of a resource. This
information can be queried using the QEL query language [15] (a Datalog based query language). Metadata interchange
and search services provide the basic infrastructure needed to retrieve information about resources and services.